Search is not available for this dataset
article
stringlengths 4.36k
149k
| summary
stringlengths 32
3.35k
| section_headings
listlengths 1
91
| keywords
listlengths 0
141
| year
stringclasses 13
values | title
stringlengths 20
281
|
|---|---|---|---|---|---|
Stimulation with rhythmic light flicker ( photic driving ) plays an important role in the diagnosis of schizophrenia , mood disorder , migraine , and epilepsy . In particular , the adjustment of spontaneous brain rhythms to the stimulus frequency ( entrainment ) is used to assess the functional flexibility of the brain . We aim to gain deeper understanding of the mechanisms underlying this technique and to predict the effects of stimulus frequency and intensity . For this purpose , a modified Jansen and Rit neural mass model ( NMM ) of a cortical circuit is used . This mean field model has been designed to strike a balance between mathematical simplicity and biological plausibility . We reproduced the entrainment phenomenon observed in EEG during a photic driving experiment . More generally , we demonstrate that such a single area model can already yield very complex dynamics , including chaos , for biologically plausible parameter ranges . We chart the entire parameter space by means of characteristic Lyapunov spectra and Kaplan-Yorke dimension as well as time series and power spectra . Rhythmic and chaotic brain states were found virtually next to each other , such that small parameter changes can give rise to switching from one to another . Strikingly , this characteristic pattern of unpredictability generated by the model was matched to the experimental data with reasonable accuracy . These findings confirm that the NMM is a useful model of brain dynamics during photic driving . In this context , it can be used to study the mechanisms of , for example , perception and epileptic seizure generation . In particular , it enabled us to make predictions regarding the stimulus amplitude in further experiments for improving the entrainment effect .
Electrophysiological measurements such as magneto- and electroencephalography ( M/EEG ) , local field potentials ( LFP ) or single unit recordings contain rich information on brain function , which may be related to specific cognitive processes , to general brain states , or to certain pathological conditions . For example , it is known that stimulation by repetitive light flashes entrains the intrinsic alpha EEG rhythm ( i . e . , frequency entrainment ) . Neurons in the human visual cortex synchronize their firing to the frequency of flickering light ( at rates of about 5 to 30 Hz ) , causing the EEG alpha frequency to change toward the stimulation frequency [1] , [2] , [3] . Clinically , this resonance effect is called photic driving . The effect of photic stimulation of the human EEG was first studied in the 1930s and 40s [4] . As early as 1947 , photic driving was reported in three cases as a potential cause for epileptic activity in patients [5] . A review of the clinical routine can be found in Niedermeyer et al . [4] . The occurrence of this effect is often interpreted as an indicator for the functional flexibility of the cortex and thus as a sign of healthiness . Today , photic driving is widely used as an activation method in clinical practice , for instance , in epilepsy , migraine , schizophrenia or depression [6] , [7] , [8] . Note , however , that only 50 to 80% of healthy volunteers show a response in the alpha range of EEG [9] , . Basic properties of the alpha rhythm during photic driving have been investigated by electroencephalographic methods [11] , [12] , [13] , [14] . A closer examination of electroencephalographic photic driving effects was given by Herrmann [2] . In that investigation , a flicker stimulus from 1 to 100 Hz in 1-Hz steps was presented . Miranda de Sa and Infantosi [10] stimulated at 4 , 5 , 6 , 8 , 10 , and 12 Hz and showed that stimulation close to the alpha peak was much more effective . The quantification of photic driving from EEG as well as MEG recordings was carried out for the first time by Kalitzin and Parra [15] , [16] . They estimated the phase clustering index of harmonically related frequency components in the EEG and MEG of normal controls and epileptic patients during light stimulations with 10 , 15 and 20 Hz . Topographic effects of encephalographic photic driving in the case of children and adolescents were described by Lazarev et al . [9] , [17] , for patients with migraine by de Tommaso et al . [18] , and for patients with schizophrenia by Jin et al . [19] . In order to gain further insight into mechanisms underlying such brain resonance effects and their relevance to brain function and pathology , as well as to make predictions concerning the stimulation parameters , generative models can be used . Such models are called biologically plausible if their state variables and parameters are biophysically meaningful . By fitting the model parameters to measurements , one can test hypotheses on the implementation of brain function . To ensure that this inversion is mathematically tractable and at the same time physically meaningful , the model must strike a balance between mathematical simplicity and biological realism . One class of models designed to meet these criteria is referred to as neural mass models ( NMMs ) ( e . g . , [20] , [21] , [22] , [23] , [24] , [25] , [26] ) . NMMs describe neural function at a mesoscopic level [27] , [28] , in contrast to single neuron models such as simple integrate-and-fire models [29] and the more elaborate Hodgkin and Huxley type of models ( e . g . , [30] , [31] ) . NMMs quantify the mean firing rates and mean postsynaptic potentials ( PSPs ) of neuronal populations , the neural masses ( NMs ) . Although , at the microscopic level , single neurons are considered the primary computational units of the brain's architecture [32] , [33] , it is also widely accepted that relevant information processing underlying brain function in both healthy and diseased states can be carried out by ensembles of interacting neurons at the mesoscopic level ( e . g . , [27] , [28] , [34] , [35] , [36] , [37] , [38] ) . In other words , NMMs describe brain activity on a scale that is highly relevant to brain function [39] , [40] , [41] . Moreover , when EEG or MEG data are used , NMMs have also the advantage that they predict exactly what is measured by these modalities , namely coherent activity in entire populations of neurons . However , this type of modeling also involves a number of simplifications that may lead to limitations . First of all , it is based on a simplified notion of the function of a neuron , namely the firing rate model: The neuron convolves the rate of incoming spikes with an alpha-shaped function and thereby generates a change in membrane potential ( PSP ) , and produces an output spike rate that is a non-linear ( e . g . , sigmoid ) function of the membrane potential . These are the most important aspects of neuronal function . However , in the brain , things are usually more complicated . For examples , , modeling is made difficult due to feedback influence of action potentials on the dendritic membrane potentials ( back propagation ) [42] , specific intrinsic firing patterns ( e . g . , bursting ) [43] and dendritic hierarchies [44] . It remains to be investigated if and to what extent such physiological details affect the properties of NMs at the mesoscopic level . A second simplification that leads to limitations is that spike time dependent effects will be missed since the model relies on firing rates rather than on actual spikes . Third , as the distributions of the neural states are simply described by their means , the impact of higher statistical moments is ignored . In order to capture the variability within a NM , one may use the Fokker-Planck formalism [45] , [46] , [47] . Finally , NMMs approximate the spatial scale of neuronal populations to be point-like [22] , [23] , [24] , [28] , ignoring the domain of spatial dynamics . In that line , the approach can be generalized , leading to neural field models [20] , [25] , [26] , [27] , [28] , [48] , [49] , which take into account the spatial extent of neural circuitry by dealing with aggregated activities in the vicinity of a given location . This puts field theories somewhere between neural mass theories and discrete neuronal networks , allowing them to address , for instance , distance-dependent delays . A quantitative analysis of neural field models can be found in Atay and Hutt [50] , [51] , for example . In this work , we use a particular local network of NMs first described by Jansen and Rit [23] , [52] , based on earlier works of Lopes da Silva et al . [24] , [53] and Zetterberg et al . [54] . This NMM comprises an elementary circuit of three interconnected NMs ( i . e . , pyramidal cells and excitatory and inhibitory interneurons ) meant to account for a cortical area , such as the primary visual cortex in our photic driving experiment . Although local neuronal circuits can be very complex [55] and may be modeled using more than three NMs ( e . g . , [56] ) , the circuit used here is the most reduced representation of the features that are relevant for the temporal dynamics , that is , positive and negative feedback loops . The Jansen and Rit structure has been shown to account for both oscillatory [57] and seizure-like EEG recordings [58] , [59] . Its dynamic behavior , in terms of stabilities and bifurcations , was first characterized by Grimbert and Faugeras [60] and , more generally , by Touboul [61] and Spiegler et al . [62] . Several such NMMs can be combined to describe networks of coupled cortical areas and account for more complex transient and oscillatory behaviors [23] , [57] , [59] , [63] , [64] , [65] . The Bayesian inversion of such network NMMs given M/EEG data ( referred to as dynamic causal modeling ( DCM ) [64] , [66] ) has been successfully used for the analysis of event-related [64] , [67] , [68] and steady-state responses [69] . To date , the dynamics of this system has been systematically investigated only under the assumption of constant extrinsic input levels , thereby allowing the system to settle in a stable state ( e . g . , fixed point or limit cycle ) [60] , [61] , [62] . However , in a photic driving experiment , one has to consider rhythmic input . Moreover , the model's response to such input is also of great importance in many other settings , since , in the brain , such local neural circuits are embedded in global brain networks and may experience high amplitude time-varying input from other parts of the brain . Because neuronal ensembles tend to oscillate intrinsically , such input is very often periodic , as evidenced by the widespread occurrence of rhythmic activity in both extracranial and intracranial recordings [70] . In this paper , we use a continuous-time periodic function as model input approximating a periodic train of pulses . In this continuous function , each single pulse is similar ( but not equal ) to the single event used by Jansen and Rit for eliciting visual evoked potentials [23] , [52] , or used in dynamic causal modeling ( e . g . , [63] , [66] ) . We systematically vary both amplitude ( intensity ) and frequency of the stimulation within the effective ranges provided by Spiegler et al . [62] . We find the frequency entrainment effect spreading over broader stimulus frequencies for higher stimulus intensities , while away from the entrainment ranges , we find complex behavior , including periodic , quasi-periodic , and chaotic dynamics . The latter behavior , in particular , provides continuous spectra . Networks of such chaotic NMMs ( incorporating network variability , for example , by different characteristic constants of time and potential ) can be used to describe colored noise sources that produce continuous portions in the spectra , such as 1/f-characteristics , that are commonly observed in M/EEG or LFP data [71] . Finally , we fit the output of the periodically forced NMM to data from the photic driving experiment in terms of the largest Lyapunov exponent and frequency detuning . The largest Lyapunov exponent measures the exponential separation or convergence of nearby trajectories . It thereby quantifies the predictability or , at the other extreme , the chaoticity of the behavior of the system and has been demonstrated to be an important marker for pathologically altered brain dynamics , especially in epilepsy [72] , [73] , [74] . In this way , we show that the NMM is a suitable model for the dynamics of brain resonance phenomena at the cortical level and demonstrate that useful predictions concerning the parameter choice of entrainment experiments can be derived . To our knowledge , this is the first study to investigate a photic driving experiment using a NMM . We demonstrate that with this NMM , one can explain effects of complex behavior in such an experiment . The results also indicate that a relatively simple model of a local neural circuit is capable of producing surprisingly complex and diverse phenomena , which are observable in brain data and relevant to the explanation of brain function .
In our previous work on the extended Jansen and Rit neural mass model ( NMM ) for a cortical area [62] , we found a self-sustained oscillation due to a stable limit cycle with a certain intrinsic frequency for constant input . Forcing such a limit cycle with periodic input to the NMM causes accelerations and/or decelerations of the oscillation , depending on the timing . If their cumulative effect is non-zero , entrainment occurs . For more details on the precise mechanism of entrainment effects , see [75] , [76] . Indeed , we observe frequency entrainment , that is , the cortical area responds with the stimulus frequency instead of the intrinsic frequency , thus forming a plateau in the frequency-detuning curves ( see colored ranges in Figure 1 ) . The detuning curve shows intrinsic behavior that is characterized by a typical repetitive s-shape . With increasing frequency , this s-shape becomes more pronounced . For the model with the stimulus amplitude that fits the experimental data best , this s-shape pattern of the detuning-curve is frequently interrupted near the stimulus hitting the intrinsic frequency 0 . 5≤η/ηint≤1 . 5 by complex behavior such as chaos ( see below , as well as Figure 1 and Figure 2 ) . Apart from the interruption of the repetitive s-shape pattern by irregularities around the intrinsic frequency ηint , the general trend of frequency detuning seems to be shifted towards the intrinsic frequency ηint by the response frequencies ηresp , which explains the experimental data ( see B in Figure 2 ) . Moreover , stimulating near the intrinsic frequency the response matches the stimulation frequency and entrainment occurs ( see Figure 1 and Figure 2 ) . Outside the entrainment ranges , more complex dynamics , including periodic , quasi-periodic and chaotic behavior , are observed ( see Figure 3 ) . Periodic and quasi-periodic behavior are associated with discrete power spectra with frequency peaks ηi that are commensurable ( i . e . , ∑ki ηi = 0 for some non-zero integers ki ) for the periodic state and incommensurable ( i . e . , ∑ki ηi≠0 for any set of non-zero integers ki ) for the quasi-periodic case . Chaotic behavior is indicated by non-closed bounded trajectories in state space , broadband continuous spectra and positive Lyapunov exponents ( see Figure 4 ) . Here , chaotic regimes arise by traversing a homoclinic Shil'nikov bifurcation ( see Figure 13 in [62] ) for non-rational ratios between the frequencies of the stimulation and the intrinsic model kinetics . This route to chaos [77] has also been identified in more theoretical neural models ( e . g . , [78] , [79] ) . Arnol'd tongues or mode-locking structures ( i . e . , entrainment regions in the parameter space [76] ) are apparent in Figure 4 as a result of negative largest Lyapunov exponents . At low amplitudes , we observe several distinct ranges of such mode locking , which seem to merge or overlap at higher amplitudes . Note that chaotic “islands” occur at incommensurable ratios between stimulation frequencies and intrinsic limit cycles and interrupt frequency locking . For example , at a stimulus amplitude of ζ = 0 . 8 , entrainment occurs for stimulus frequencies between 0<η≤0 . 06075 , 0 . 06831≤η≤0 . 07403 and 0 . 09474≤η≤0 . 1206 interrupted by chaotic regimes between 0 . 004835≤η≤0 . 03464 . At a stimulus amplitude of ζ = 2 . 4 , entrainment occurs for stimulus frequencies between 0<η≤0 . 1447 , 0 . 1545≤η≤0 . 1608 and η>0 . 1749 interrupted by chaotic regimes between 0 . 0365≤η≤0 . 04397 , 0 . 05237≤η≤0 . 06818 and 0 . 08906≤η≤0 . 09547 . Note that these entrainment ranges are rough estimates due to the finite sampling of the parameter space and due to the occurring “islands” of chaos . The chaotic regimes that are present in the parameter space feature a single positive largest Lyapunov exponent that is equal to the entropy of the attracting set ( see Figure 4 ) . By studying the Lyapunov spectra , configurations are discovered where the system has two zero Lyapunov exponents and evolves on a two-dimensional invariant torus , indicating quasi- and bi-periodicity ( see Figure 5 ) . In general , the model is dissipative ( i . e . , the sum of Lyapunov exponents is negative ) and does not exhibit hyperchaos , which is a higher form of chaos with at least two directions of hyperbolic instability on the attractor [80] ( see Materials and Methods for further explanation ) , as seen from the observation that the second largest Lyapunov exponent is non-positive and the Kaplan-Yorke dimension ( see Materials and Methods ) never reaches or exceeds the value of two ( see Figure 6 ) . This means that the dynamics are low dimensional , not only for the periodic , but also for the chaotic regimes , as compared to the dimensionality of the system ( which is six plus one dimension for the force ) . In general , the maximum Kaplan-Yorke dimension is a non-integer because of the complex geometry of the attractor . The periodic forcing seems to work mostly in the direction of entrainment , and although there are occasional “islands” of chaotic regimes , the regular forcing does not let the dynamics become exceedingly chaotic . Furthermore , we find that the model is indeed able to explain frequency entrainment that is observable during a photic driving experiment ( see also [1] , [2] , [81] ) . Note that Figure 5 in Schwab et al . [81] contains an error in the labeling of the y-axes . Each graph in this figure correctly plots the ratio of stimulus to response frequency ( y-axis ) against the ratio of stimulus to alpha frequency ( x-axis ) . In this case , a horizontal line indicates an entrainment effect , while absence of entrainment would result in a diagonal line . We estimated the largest Lyapunov exponents from the data ( see Methods section ) . In order to probe the stability of this estimate , we repeated it with the same data after adding various levels of Gaussian noise . The pattern of the Lyapunov exponents as function of stimulus frequency appears to be quite stable except for very low signal-to-noise-ratios SNR≤3 dB ( see supplementary Figure S1 ) . We compare our model outcome with these experimental Lyapunov exponents ( see the section Experimental data in the Materials and Methods section and Figure 7 for the experimental paradigm ) and find a particular stimulus amplitude for which , for all ten subjects , the model predicts Lyaponov exponents that are in close agreement with those estimated from the experimental data ( Figure 8 and Table 1 ) , with the amplitude being close to ζ = 3 . 6692 for all ten subjects . In seven of the subjects , the correlation between model prediction and measurement over stimulus frequencies was significant ( p<0 . 05 , corrected ) . A bootstrap test yielded a probability of error ( significance ) for the mean over subjects of 6 . 2% ( see also Figure 8 ( B ) and Table 1 ) . For three subjects ( numbers 3 , 6 and 7 ) , the individual fit was not significant ( see Table 1 ) . Interestingly , this is reflected in the means and the standard deviations of the shift-and-scale parameters u and v ( see the section Comparison in the Materials and Methods section and supplementary Figure S2 ) . For the corresponding model configuration , we present a compact representation in Figure 9 and describe the system states qualitatively in Table 2 . In the range of stimulus frequency η between 0 and 0 . 0534 , the system performs limit cycles and appears to undergo a cascade of period-adding bifurcations [82] with descending stimulus frequency . This local bifurcation consists of saddle-node bifurcations in which a ( n+1 ) -periodic orbit arises out of a n-periodic orbit for n ∈ N1 [83] , [84] . For stimulus frequencies η above 0 . 0534 , ranges of chaotic , periodic and quasi-periodic behavior occur . Due to the high dimensionality of the system , an instructive presentation in the form of a video is available , comprising orbits ( PSPs ) , time series , and power spectra ( see Video S1 ) .
It should be pointed out that many aspects of our results are in close agreement with previous studies of other types of periodically driven oscillators ( see , for example , [75] , [76] , [85] ) . Frequency entrainment effects have been described in , for example , the Rössler system [86] , the Oregonator model [87] describing chemical oscillators such as the Belousov-Zhabotinsky reaction ( e . g . , [88] ) , the Duffing oscillator describing mechanical pendulums with flexible elements [89] , [90] , the van der Pol oscillator modeling electrical triode circuits [91] , the Lorenz system describing turbulent convection in hydrodynamics [92] , [93] , and the Hodgkin-Huxley model of a neuron [94] . Overlapping or merging mode-locking regions in the parameter space were also discovered in a periodically driven van der Pol oscillator [95] . While reverse periodic-adding cascades appear to be the route to chaos in this study of a periodically forced Jansen and Rit model , in a number of previously investigated systems , cascades of period-doubling led to chaos , for example , in the Duffing oscillator [89] , [90] , [96] , the Lorenz system [93] , [97] , the Rössler system [98] , the Brusselator [99] and the Oregonator [87] . In the van der Pol oscillator , both routes – period-adding [83] , [100] and period-doubling cascades – occur [91] , [101] . On the other hand , our results concerning the route to chaos are in line with findings in a periodically stimulated excitable neural relaxation oscillator [27] and a simple model of the Belousov-Zhabotinsky reaction [102] . Crevier and Meister [103] describe retinal ( electroretinogram , ERG ) and cortical responses ( LFP and visual evoked potential , VEP ) to periodic flashes of light in salamander and humans . They also found complex behavior such as frequency entrainment in experimental data as well as in their model . In contrast to our findings , they found a cascade of period-doubling bifurcations ( in both data and model ) that leads to chaotic regimes in their model . Finally , quasi-periodic solutions have also been reported for various systems , such as the van der Pol oscillator [91] , [101] , the Oregonator [87] , the Rössler system [98] and the Hodgkin-Huxley model [94] , [104] . We applied the concept of a periodically forced oscillator to model brain resonance effects . In the brain , such periodic input might stem from rhythmic stimulation of the brain , such as in the photic driving paradigm , or from the output of other oscillating brain areas . Such coupling between ( oscillating ) processes inside and outside the brain has been discussed as important for the processing of information ( e . g . , [105] , [106] , [107] , [108] , [109] ) . We described the dominant intrinsic brain rhythm using the NMM performing a self-sustained oscillation , generated by an Andronov-Hopf bifurcation [62] . Generally , resonance phenomena such as frequency entrainment in photic driving experiments can be explained by the concept of a periodically forced oscillator . Applying periodic input to an oscillatory system will change the current phase of the oscillation and frequency entrainment ( i . e . , phase locking ) occurs if the sum of phase changes is nonzero over time [75] ( see Results for more details ) . It is expected that the dynamics of the system depend on timing , that is , the ratio between stimulus and intrinsic frequencies , as well as the intensity of stimulation . While the general trend of the frequency-detuning curves is similar for our model and experimental data , there are numerous deviations ( see Figure 2 ) . These might be explained by the simplicity of the model . In the brain , many neuronal circuits are likely to be concurrently active and deviating behavior might be canceled out . In our simulations , we found that the dynamics of the periodically forced extended Jansen and Rit NMM feature a rich mosaic of complex behavior ( Figure 3 ) . From the parameter space analysis presented in Figure 4 , it can be seen that both flicker intensity and frequency are critical parameters . As expected based on theory [75] , [76] , the state space analysis reveals that the system is entrained by the stimulus frequency ( see Figure 1 ) where the entrainment regions ( i . e . , plateaus in the frequency-detuning curve ) around the intrinsic frequency become wider with increasing stimulus intensity ( results not shown ) . Also , stimulus frequencies below the intrinsic frequency lead to decelerations of the intrinsic rhythm of the modeled cortex and vice versa . This phenomenon is reflected by the ratio of stimulus to response frequencies η/ηresp above and below the diagonal in Figure 1 for stimulus frequencies below or above the intrinsic frequency ( i . e . , η/ηint<1 and η/ηint>1 ) , respectively . In regions of the parameter space without entrainment , complicated interaction between stimulus and intrinsic kinetics leads to periodic , quasi-periodic , and chaotic behavior , as indexed by the largest Lyapunov exponents and the Kaplan-Yorke dimension . Areas with different dynamic behavior form fractal structures in parameter space ( Figure 4 to Figure 6 ) so that rhythmic and chaotic brain states are found virtually next to each other and even small parameter changes can give rise to a switch from one to another . For these parameter configurations , different forms of the extrinsic periodic input would affect the specific pattern of chaotic regimes in the parameter space , but not the qualitative behavior if the parameter of the stimulus shape δ ranges between 109 and 130 , as was found by additional simulations with different stimulus shape parameters ( results not shown ) . On this account , the shape of the extrinsic input is an important model parameter for investigating the occurrence of complex regimes that needs to be investigated in the future . It should , however , be pointed out that this result has been obtained from a purely deterministic model without any added or modulating noise . If noise is added to the input , this would cause jitter in its amplitude and frequency , and thereby impose a blur on the pattern depicted in Figure 4 . However , the gross patterns are expected to survive; that is , areas with a high density of “chaotic” configurations ( e . g . , around 6 Hz and amplitudes between 4 and 5 mV ) will feature a lower degree of predictability than areas without any configurations with positive largest Lyapunov exponents ( e . g . , around 11 Hz , same amplitude range ) . Strikingly , this is corroborated by the fact that the characteristic patterns of unpredictability generated by the model were also found with reasonable accuracy in the noisy experimental data ( Figure 8 ) . We identified a particular stimulus amplitude , where , for all subjects , the Lyapunov exponents are in close agreement between experiment and model ( Figure 8 and Table 1 ) . We found that the profile of the characteristic Lyapunov spectra for the stimulus amplitude that best fits the data is preserved when noise is added to the input for a signal-to-noise-ratio ( SNR ) up to 10 dB ( for more details , see Model in the Materials and Methods section ) . The intensity that best fits our experimental data is located in the upper portion of the effective range for exciting inhibitory interneurons . Since the largest Lyapunov exponent reflects fundamental properties of the current dynamic regime of a system ( as evidenced , for example , by its sensitivity to pathological states of the brain , such as epilepsy , see [72] , [73] , [74] ) , the fact that our model predicts its dependence on the most important stimulus parameter ( frequency ) corroborates the validity of the model . Consequently , we predict that a decrease in stimulus intensity in photic driving experiments would shrink and an increase would broaden the ranges of frequency entrainment ( i . e . , the plateaus in the frequency-detuning curve ) . Our model also predicts that saturation effects become important starting with approximately 1 . 3 times the currently applied stimulus intensity and for intensities close to zero . A stimulus increase between 1 and 1 . 3 times the current intensity could lead to an improved entrainment effect ( i . e . , broadened range ) . Such broadening of the entrainment range is particularly important because in clinical practice , the individual alpha frequency is usually unknown . It is important to know how great an increase in the stimulus intensity still improves the entrainment effect and hence makes the photic driving more reliable . Although the effect of photic driving has long been known , and standard examination in neurology includes intermittent photic stimulation in patients with suspected photosensitive epilepsy , the exact pathomechanism is not well understood . It is known that the photoparoxysmal response ( PPR ) is inheritable . In terms of electrophysiology , photosensitive epilepsy seems to be associated with changes in oscillatory activity . For example , Parra et al . [16] found enhanced gamma band synchrony and hypothesize that “ … some sort of recruitment or dynamic capture of neurons into larger assemblies appears to precede the epileptic chain reaction ( ictal cascade ) that ends in a paroxysmal oscillation , the PPR . ” Likewise , Visani et al . [110] confirmed the potential importance of gamma band activity and found alpha band activity relevant to the PPR . Using transcranial magnetic stimulation , Siniatchkin et al . [111] found evidence that an increased excitability of the occipital but not the motor cortex might be associated with the PPR . The above studies indicate that a model including more than one area might be needed to further elucidate the pathomechanism of the PPR . Our model can be extended to give such experimental predictions or explanations for experimental findings . However , concrete simulations with , for instance , increased excitability in the occipital cortex and regular excitability in a second region are beyond the scope of this paper . In short , we show that our model is capable of accounting for major aspects of the photic driving paradigm . This sets the scene for future work that will explore the predictions of the model in health and disease in more detail based on additional experimental data . Furthermore , a systematic exploration of the parameter space of the model with respect to brain resonance is needed . All this requires substantial efforts and is beyond the scope of the current proof-of-principle paper . A principal limitation of our study is the modeling of the thalamus as independent signal generator , neglecting the cortico-thalamic feedback loop . However , we tested a model of the thalamus according to Robinson et al . [112] and found that , at least for the parameters of the cortical model used in this work , the simple signal generator approach yields a good approximation . Future work will include measurements and explicit modeling of the thalamo-cortical loops . Another issue which must be discussed is whether and to what extent our results support the idea of chaotic dynamics in the brain . The model investigated here describes complex , partially chaotic , dynamics at the mesoscopic spatial scale , which captures mass action of neural ensembles [28] . Chaotic dynamic regimes have been shown before in mesoscopic models of the cortex [113] , [114] and of the olfactory bulb ( e . g . , [115] ) . Concerning the brain , there is evidence for chaotic behavior at different hierarchical levels , from single neurons to entire neural ensembles [116] . A suitable means to experimentally access neural activity at the mesoscopic level is provided by M/EEG , which records the summed activity of 105 to 109 , mainly cortical , neurons [40] , [41] . M/EEG data describe high-dimensional , noisy , nonlinear , and non-autonomous processes [117] , which render it difficult to distinguish between stochastic and complex deterministic processes like deterministic chaos . Accordingly , although there is some evidence for chaos in such data ( e . g . , in epilepsy ) , the issue remains controversial ( for a discussion , see [118] and the references cited therein ) . However , irrespective of whether the complexity of M/EEG fulfils the exact mathematical criteria of deterministic chaos , the parsimonious NMM , as shown here , helps to better describe the dynamics of such data and the underlying brain processes . Apart from brain rhythms in characteristic frequency bands ( e . g . , the alpha rhythm ) , complex behavior with noise-like characteristics causes the continuous spectral components in these data . This can be interpreted as filtered noise ( e . g . , stochastic sensory input ) or described by nonlinear deterministic processes . We have shown that periodically driven NMMs may explain the continuous spectral components of M/EEG without having to consider noisy input processes . Other NMM studies often apply a stochastic input process with the effect that the spectra are more realistically widened around an intrinsic frequency of interest ( e . g . , alpha band ) ( e . g . , [23] , [57] , [58] , [65] ) . It is , however , an advantage , if these continuous spectral components can be modeled and controlled as intrinsic phenomena of the neural circuits , because there is evidence that broad spectral components are also modulated by cognitive processes and hence their generative processes play a role in information processing ( e . g . , [119] , [120] , [121] ) . This is corroborated by the postulated prominent role of chaos in information processing ( see , for example , [115] , [116] ) . Furthermore , the complex behavior of the NMM for certain parameter sets or ranges could be used to explain ordered sequences of dynamic regimes and multi-stability in M/EEG data by producing a temporal hierarchy [62] . Such ordered sequences have been observed in , for example , perception ( e . g . , mono- and binocular rivalry [122] , Necker-cube illusion ) , stages of sleep [123] , [124] , changes in attention or vigilance , learning and training such as odor recognition [125] , [126] , progression of disease such as epilepsy [127] , [128] , [129] , [130] , and effects of medication . State transitions or multi-stabilities appear since the brain is subjected to multiple high-dimensional stimuli from both exogenous ( e . g . , vision or haptic ) and endogenous processes ( e . g . , endocrine or circulatory system ) , and is highly dependent to the current on the current individual state ( e . g . , vigilance , sleep or attention ) . For example , one can interpret the quasi-periodic behavior in Figure 3 as multi-stability . However , the orbits are sensitive to noise , albeit in terms of fine structure and the associated sequences , rather than the overall structure . One way to achieve ordered sequences of dynamic regimes that are sufficiently robust against noise is to adequately change the state space through parameter changes that are slower than the state dynamics producing a temporal hierarchy [62] . For example , one could incorporate a second model with kinetics slower than the NMM ( e . g . , representing metabolic processes or the neuroendocrine system ) that controls a subset of the NMM parameters ( e . g . , couplings in terms of synaptic plasticity ) . In this way , the dynamic behavior of the NMM may change qualitatively through passing bifurcations and thus sequences of the complex regimes will be occur . A fine example of this approach is provided by Steyn-Ross et al . [131] , who modeled the succession of slow wave and REM sleep phases in humans using a mean field model . The parameters of the model were controlled by the states of a low-kinetics model describing the levels of acetylcholine and somnogens ( such as adenosine ) . Note that the directions of parameter changes play an important role in parameter ranges of the system where a hysteresis occurs ( see , for example , in [62] Figure 2 and Figures 4 to 6: branch type-I A and B , and type-II AB to CC ) . This previous study [62] provides a catalogue of regimes that is potentially helpful to prevent the system from hysteretic behavior or , quite the reverse , to perform hysteresis . The present study is the first to find complex types of behavior like entrainment , chaos , and periodic and quasi-periodic motion in a periodically forced Jansen and Rit NMM for a single cortical area for biologically plausible parameter ranges without considering noise processes . Such dynamics are observable in brain data and relevant to the explanation of brain function . We demonstrate that with the NMM , one can explain brain resonance phenomena like frequency entrainment in a clinically relevant photic driving experiment . It should be pointed out that , at this stage , the aim of our model has not been to directly improve the diagnostics of mental illnesses , but rather to allow deeper understanding of the mechanisms underlying a diagnostic tool and thereby pave the way for future new treatments and diagnosis techniques . As a logical next step , the model should be applied to pathological cases in order to specify what disease-specific inferences can be made . As any model , our model features a number of simplifications with respect to reality . The mean-field model studied embodies structural ( e . g . , local neural circuitry ) as well as functional approximations ( e . g . , mean postsynaptic potential ( PSP ) , mean firing rates and its conversions ) of neural circuits to describe brain dynamics at the mesoscopic and the macroscopic levels , which are accessible , for instance , to LFP and M/EEG . Simplifications appear at all levels of modeling: the description of single cell behaviors , the modeling of neural masses ( NMs ) based on a single cell description ( i . e . , firing rate neuron ) , the description of the local neural circuitry and the description of networks of brain areas . On the single cell level , we consider the firing rate instead of individual action potentials . Moreover , only two types of synaptic kinetics are modeled , which leads to two types of neurons that either excite or inhibit other neurons . In the brain , there is a great diversity of electrophysiological neuron types that differ in their specific input and output operations [43] . On the population level , the distribution of states ( i . e . , PSPs and firing rates ) is described by their means , while variances and higher-order statistics are left out of consideration . The local neural circuitry of the cortex is characterized by a wealth of distinguishable populations and their interconnections ( see , for example , [55] ) . In the model structure used here , this is simplified by simply considering pyramidal cells and two feedback loops established by inhibitory and excitatory interneurons . Finally , in this work , we deal with a cortical area mean-field model , without considering projections to the rest of the brain . In particular , the thalamus , which might play a role here , is modeled only in terms of its output . Moreover , since the retina is fully illuminated by the flicker that drives much of the visual cortex ( see Materials and Methods for further explanation ) , we consider the entire primary visual cortex as a single source using a simple NMM for a single cortical area . Of course , the primary visual cortex is much more complicated than a single Jansen and Rit circuit , not to speak of its incorporation in brain-wide networks . Hence , our model can only represent a subset of the dynamics of the entire system . Nonetheless , our results show that the model can account for the main phenomena in the photic driving paradigm . However , one must be aware that the future availability of new or more detailed data might necessitate an extension of the model . In this study , we show that a simple local cortical area model is already capable of performing relevant complex dynamics , particularly in response to periodic inhibitory feed-forward stimulations . Based on the fact that such a local cortical circuitry of neural populations is embedded in large-scale networks that can span the whole brain and also include subcortical structures such as the thalamus , the question arises to what extent network interactions might contribute to the complexity of brain signals such as M/EEG . The present work might also contribute to the understanding of large-scale networks . In particular , the present results can be applied to inhibitory feed-forward interactions in networks between two local area models , where one model periodically performs spikes that drive the other model . The frequency entrainment or locking phenomena we found here can thus be interpreted as an effect of network interactions , which might have an impact on functional or effective brain connectivity measures such as the phase correlation ( e . g . , [109] , [132] ) or the Granger causality ( e . g . , [133] , [134] ) . From the modeling perspective , one can obtain a dynamic regime of a local cortical area such as quasi-periodic behavior within a network by frequency locking through feed-forward inhibition from another local area by considering the following steps: ( i ) tuning the driving cortical area so that it performs ( spiky ) rhythms , ( ii ) selecting the dynamic regime depending on stimulus amplitude and frequency ( see Figure 4 to Figure 6 ) , ( iii ) adjusting the characteristic time constant of the driver to tune the frequency of the driver to the required stimulus frequency , and ( iv ) selecting the characteristic potential and/or the coupling parameter so that amplitude of the driver fits . The parameters for the driving cortical area can be taken from the catalogue of regimes presented in our previous work [62] . Moreover , in order to best reproduce a specific phenomenon , for instance in M/EEG data , this catalogue helps to balance the complexity of a network , in particular , whether a single area model is sufficient or not . The effective extrinsic input ranges of a cortical area model [62] can be used to determine the coupling parameters between areas in order to prevent a network or individual cortical areas from saturating . The use of these approaches to control or set up a network depends on the complexity of the graph . For instance , several bidirectional connections or feedback loops usually make a setting more difficult . In such complex graphs , one can expect more complicated behavior than for a single local area model , such as hyperchaos or phase locking of several ( chaotic ) regimes . However , one might to have to perform a separate analysis for the network .
A generative model for brain measurements such as M/EEG can be specified by two separate systems: the state system f explaining the usually hidden neuronal states x ( e . g . , the mean postsynaptic potentials ( PSPs ) of neuronal populations that potentially generate M/EEG ) , and the observer system g relating the neuronal states to the measurements z: ( 1 ) and ( 2 ) where L ( ∂/∂t ) is a temporal differentiation operator , p denotes the extrinsic inputs , and sx and sz parameterize state and observer system , respectively . For the state system f , we use a neural mass model ( NMM ) of a cortical area . For the observer system g , we use a simple linear relationship , as we simply consider one area ( i . e . , source ) , because the retina is fully illuminated during the photic-driving experiment that presumably drives much of primary visual cortex in parallel , and therefore no elaborate forward modeling is needed . The state system will be explained in more detail in the following paragraphs . The NMM of Jansen and Rit [23] , [52] describes a local network representing a cortical area . This basic circuit , consisting of three interacting neural masses ( NMs ) , namely pyramidal cells ( PCs: NM 3 ) with feedback loops mediated by excitatory and inhibitory interneurons ( EINs and IINs: NMs 1 and 2 ) , has been described in a number of previous studies ( e . g . , [23] , [44] , [52] , [53] , [57] , [58] , [59] ) . Note that the feedback loops may also be modeled dynamically ( see , for example , [26] , [48] , [112] when also considering propagation delays , see , for example , [135] , [136] , [137] ) . However , here we assume connections within a single area , resulting in transmission times which are shorter than the characteristic ( dendritic ) time constant τ = 10 ms . Therefore , it is sufficient to describe the feedback connection by a gain constant . With this NMM , the mean neuronal states can be described by a system of six nonlinearly coupled first-order ordinary differential equations: where the state vector x = ( x03 , x31 , x32 , y30 , y31 , y32 ) T contains the normalized mean PSPs xba and currents yba at NM b caused by NM a . The extrinsic afferents T projected to NM b are denoted by xbT . The average synaptic gains or the average numbers of synaptic contacts established between the two NMs a and b are represented by the constants αba . Furthermore , β is the ratio of excitatory to inhibitory dendritic time constant β = τe/τi and , in the formulas ( 3 ) to ( 5 ) , the dot indicates the derivatives with respect to the normalized time κ = t/τ , where τ is the characteristic time scale . The transfer function O ( xb ) that converts the normalized mean PSP xb = Σa xba ( i . e . , the normalized potential at the axonal hillock ) to the normalized mean firing rate is taken to have a sigmoidal shape O ( xb ) = 1/ ( 1+γ exp ( −xb ) ) , where γ represents the distribution of firing thresholds within a NM . The normalized and generalized Equations ( 3 ) to ( 5 ) correspond to the Jansen and Rit model [23] with the characteristic time constant τ = τe , the coupling parameter αba = 2e0 r cba He , i τ2/τe , i , the sigmoid parameter γ = exp ( υ0 r ) , and the states xba ( κ ) = rbυba ( τκ ) , with the following parameters: maximum firing rate 2e0 , the slope of the sigmoid r , the mean number of synaptic contacts cba , and the excitatory and inhibitory synaptic gains He , i ( for more details , see [23] , [62] ) . Note that we use normalized parameters and variables in the rest of this paper without further indicating this . In this work , we explore the dynamics of the single-area model as a function of amplitude and frequency of a periodic input . This input consist of brief pulses similar to ones used by Jansen and Rit for eliciting visual evoked potentials [23] , [52] , or used in dynamic causal modeling ( e . g . , [63] , [66] ) . These pulses are meant to represent the impulse response of the visual pathway , which has been investigated experimentally by a number of researchers ( see [52] , and the references cited therein ) and described analytically by Watson and Nachmias [138] . In the following , we will specify the parameter space to be investigated . The system described by Equations ( 3 ) to ( 5 ) has nine parameters , namely couplings αba with ba = {13 , 23 , 31 , 32} , kinetic ratio β , sigmoid parameter γ , and extrinsic inputs xbT with b = {1 , 2 , 3} . Jansen and Rit [23] proposed a specific parameter set for the NMM of a cortical area , based on a thorough discussion of the literature . The normalization of time and potentials by Jansen's excitatory dendritic time constant and sigmoid slope ( τ = 10 ms and r = 0 . 56 mV−1 , respectively ) leads to the following dimensionless parameters in our model: couplings α13 = 12 . 285 , α23 = α13/4 , α31 = 4α13/5 , α32 = −11α13/13 , kinetic ratio β = 0 . 5 and sigmoid parameter γ = 28 . 7892 . The extrinsic inputs on the three NMs are taken to be constant for EINs x1T = 0 and PCs x3T = 3 . 36 , and time-variant for IINs in the form of periodic pulses x2T = ζ exp ( −2δ cos2 ( θ ) ) , with the angle θ ( 6 ) specified by stimulus amplitude ζ and stimulus frequency η ( δ controls the shape and is set to δ = 110 ) . Such a peaky waveform has been found in the lateral geniculate nucleus of the thalamus in response to square visual stimuli [139] . Interestingly , a very similar waveform can be generated using a NMM of the thalamus , as proposed by Robinson et al . [112] , which takes into account the intra-thalamic and thalamo-cortical feedback loops ( e . g . , [140] ) . In this model , a strong inhibitory influence of the reticular nucleus on the thalamic relay cells during the relaying of external sensory stimulation , such as an on/off waveform of flickering lights , sharpens the cortical input to render it pulse-like . The time-variant input to the IINs may represent thalamic feed-forward input . This type of disynaptic feed-forward inhibition has been described as crucial for bottom-up processing in the somatosensory ( e . g . , [141] , [142] ) , auditory ( e . g . , [143] ) , and visual ( e . g . , [144] , [145] ) systems of rodents . Moreover , the literature provides evidence that feed-forward inhibition ( e . g . , from layer IV IINs driven by thalamus ) dominates excitation ( from thalamus ) ( e . g . , [144] , [146] ) . Also , our previous model analysis of the Jansen and Rit circuit reveals the importance of input on IINs for controlling cortical behavior [62] . In the absence of stimulation ( i . e . , x2T = 0 ) , the system intrinsically performs limit cycle oscillations arising from Andronov-Hopf bifurcations , appearing as harmonic oscillations with a frequency of approximately ηintr = 0 . 108 ( see bifurcation diagram and phase portraits , Figure 2 and Figure 3 in [62] ) . Applying the characteristic dendritic time constant τ = 10 ms as defined above , this corresponds to the parameter set proposed by Jansen and Rit [23] and an actual oscillation frequency of f = 10 . 8 Hz , and can be used to describe alpha rhythms in brain signals . This characteristic time constant is used in all results reported in this work . Note , however , that varying the characteristic time constant τ only scales the neuronal states x ( κ ) in time t = τ κ and thus the frequency f = τ−1 η while the states x , the form of time signals and the underlying mechanisms such as bifurcations remain unaffected . Hence , the frequency depends on the choice of the characteristic time constant τ and thus the normalization embraces all cases of τ . In order to study the system with periodic stimulation around the intrinsic frequency ( ηintr = 0 . 108 ) , the stimulus frequency η is taken to range from 0 to 0 . 19 . The stimulus frequency is nonlinearly sampled ensuring ( η Δκ ) −1∈N1 with the sampling interval Δκ , so that the pulses are well sampled . The stimulus amplitude ζ is linearly sampled from 0 to 4 . 1 to cover the effective range of excitatory inputs on IINs within the limit cycle which exists when extrinsic input on IINs is constant ( see Figure 8 in [62] ) . Since the specification of the effective extrinsic input ranges is based on an analysis of the invariant transfer function ( sigmoid function ) [62] , these ranges are valid for any type of input , no matter whether it is constant or time-variant . In summary , for analysis , we consider a system of seven first-order ordinary differential equations ( Equations ( 3 ) to ( 6 ) ) describing the ( neuronal ) states x* = ( x03 , x31 , x32 , y30 , y31 , y32 , θ ) T specified by two parameters p = ( ζ , η ) T . We study the differential equations ( 3 ) to ( 6 ) numerically using the fourth-fifth order Runge-Kutta method over κ = 30·103 in time ( which equals 5 minutes for τJR = 10 ms , according to Jansen and Rit [23] ) with a relative tolerance of 10−11 , and then linearly sampled with an interval Δκ = 10−2 for further analysis . From the last 6·103 samples ( last minute if τJR = 10 ms ) , the histograms of each state were computed using the optimal number of bins [147] . Using the state equations , we also compute the characteristic mean frequency of each attractor [85] . Characteristic mean frequency is the time average of a trajectory over the angle velocity at points along an n-dimensional curvature forming an attractor in state space . To study the complex behavior , we computed the power density spectra of the time series ( last 6·103 samples ) using the fast Fourier transform , especially for the time series of the PSPs of the PCs , which are reflected in M/EEG . We also compute the characteristic Lyapunov spectra , that is , all six Lyapunov exponents λ1>λ2 …>λ6 directly from the differential equations ( 3 ) to ( 6 ) , using the Fortran algorithm by Chen et al . [148] , integrated for κ = 1073 , 742 . 00 using a constant sample interval Δκ = 10−3 . Chen et al . [148] used a constant time-step fourth-order Adams-Bashforth integration method and a QR-reorthoginalization that also preserves orthogonality for higher-dimensional systems . The time interval is sufficiently long to stably estimate the characteristic Lyapunov spectra ( error<10−6 ) . The Lyapunov spectrum gives a quantitative measure of the sensitivity of the states of the system dependent on the initial conditions , or , more precisely , the average rate of divergence or convergence of two neighboring trajectories in the state space . Furthermore , the whole Lyapunov spectrum enables statements of hyperchaos . Hyperchaos is a higher form of chaos with at least two rather than one directions of hyperbolic instability on the attractor [80] indicated by two or more positive Lyapunov exponents and by a Kaplan-Yorke dimension larger than two . Such a hyperchaotic attractor appears as a ‘folded-towel’ structure through a continuous stretching and folding in , at least , two independent directions of the state space [149] . Such behavior was first reported by Rössler in 1979 [80] . Generally , a system that performs hyperchaos must be of at least four dimensions . Due to the computational effort required , only the largest exponent or a few of the largest ones are calculated in most of the existing literature . Here , we compute the whole characteristic Lyapunov spectra running on a massive parallel cluster system of the advanced computing unit at the Computer Center , Ilmenau University of Technology . We also select several regions from the parameter space with scattered , presumably fractal , patterns of chaotic regimes ( i . e . , positive largest Lyapunov exponents ) for recomputing at a finer stimulus amplitude and frequency resolution ( see Figure 4 through Figure 6 ) . We probe the stability of the characteristic Lyapunov spectra by adding a stochastic term to the stimulus x2T . Although the Gaussian noise process that we used is not autocorrelated and could lead to errors due to the constant integration step size of the Adam-Bashforth method , the estimation of the characteristic Lyapunov spectra ( for the stimulus amplitude that fits the experimental data best ) is stable up to a signal-to-noise-ratio ( SNR ) of 10 dB , especially for the 1∶1 entrainment region ( i . e . , η≈ηint ) . However , the stochastic term changes the characteristic Lyapunov spectra specifically , for instance , at stimulus frequencies η around 2/3 of the intrinsic frequency ηint . For this stimulus frequency range ( 0 . 5817<η/ηint<0 . 7632 ) , the profile is qualitatively preserved for mild noise with a SNR up to 17 dB . We determined the SNR as the ratio of the variances of the deterministic and stochastic portions of the stimulus . The variance of the deterministic terms σ2 ( x2T ) ( i . e . , periodic pulses ) is given as follows ( 7 ) where I0 is the modified Bessel function of the first kind , δ is the shape parameter and ζ is the amplitude of the stimulus xT2 . The knowledge of the whole spectrum enables us to derive the Kaplan-Yorke dimension [150] given by ( 8 ) where k is such that ( 9 ) The Kaplan-Yorke dimension measures the upper bound of the Hausdorff dimension and is similar to the information dimension ( entropy ) or correlation dimension of an attractor . The Hausdorff dimension quantifies the complexity of the geometry of the attractor . For example , the Hausdorff dimension of a point is zero , of a line is one , of a plane is two , but irregular sets , such as fractals or the attractors found in this work , can feature non-integer Hausdorff dimensions . We divide the state space by classifying the behavior of the system qualitatively . To that end , we specify a Poincaré map P by choosing a suitable hyperplane transverse to the limit cycle of the unperturbed system . A Poincaré map P considers the intersections of a trajectory existing in the d-dimensional state space with a hyperplane of dimension d−1 . The resulting discrete series of intersection points allow the characterization of the dynamics near periodic solutions . Finally , to study the relationship between system perturbation and system response in terms of synchronization and frequency entrainment , we compute the frequency-detuning curves [76]; that is , the difference of the response frequency ( characteristic frequency or largest peak in the spectrum ) and the stimulus frequency plotted against the stimulus frequency . Experimental data were obtained by performing a photic driving experiment that was adapted to the individual alpha frequency of the subjects . Data were previously published by Schwab et al . [81] . The aim of this former study was the quantification of frequency entrainment in the alpha rhythms that was most effective in the region around individual alpha and half alpha . Ten healthy participants ( 22 to 40 years of age , 5 male; 5 female ) were stimulated by an intermittent flickering light , while EEG ( 32 channels , enhanced 10-20 system with a 10-10 system over the occipital region , Compumedics Neuroscan , El Paso , USA ) was recorded . EEG was sampled at 1000 Hz and hardware-filtered between 0 . 1 Hz and 300 Hz . An initial resting condition of 60 seconds was recorded to define the individual alpha rhythm of each participant . The individual alpha frequency measured ranged from 9 . 5 Hz to 11 . 8 Hz . After this period , flicker stimulations were conducted for 15 fixed frequencies with an alpha ratio ( stimulus/individual alpha frequency ) ranging from 0 . 4 to 1 . 6 in each participant ( randomized order of presentation ) . The flicker stimuli were generated by two LEDs outside the measurement chamber and were delivered via optical fibers to about 9 cm in front of the closed eyes of the subjects in order to ensure relatively stable luminance over subjects and a fully illuminated retina . The closed eyelid diffuses the flickering light from the optical fiber ( with its viewing angle ) so that the whole retina is illuminated ( e . g . , [151] , [152] ) . Each stimulation frequency was presented in a sequence of 20 trains . A single train contained 40 flashes and was followed by a resting period ( 4 s ) . The complete experimental design is summarized in Figure 7 . One EEG channel located in the occipital region ( O1 ) was examined per participant . Data were filtered and down-sampled to 200 Hz . For each participant , periods of 62 . 5 s ( n = 12500 data points ) were analyzed for the 15 flicker frequencies presented ( the shortest available data length of the individual flicker blocks F1 through F15 is 62 . 5 s over all participants investigated ) . The estimation of the largest Lyapunov exponent was based on the approach of Wolf et al . [153] . An embedding dimension of 16 , a time delay of 9 ( ≈50 ms ) and an evolving time of 5 ( ≈25 ms ) was used for the investigation of flicker stimulations . Embedding parameters were defined according to Atay and Altintas [154] . Our periodically driven deterministic model exhibits chaos or otherwise complex behavior in certain parameter ranges ( see Figure 4 ) . A chaotic regime can be considered as a source of noise . Empirical data as from the photic driving experiment ( see Experimental data ) generally represent a highly noisy ( nonlinear ) signal . The sources of this noise are diverse and range from technical noise ( e . g . , Johnson-Nyquist noise of sensors , 50/60 Hz powerline interferences ) , via non-brain biological noise ( such as transpiration , muscle activities of the heart or eye movements ) , and unrelated brain activity , to the chaotic behavior of the actual stimulus processing . Thus , mapping experimental data to any biologically motivated model with reasonable accuracy is an extremely challenging task . In particular , comparing identified system variables from model and data with respect to their absolute values needs to be done delicately . Consequently , we chose to compare the Lyapunov exponents as scalar measures of the variations of the regimes of the system with respect to a well-defined “external” parameter , here the stimulus frequency ( available for both model and experiment ) . In contrast to the experimental design , our model has the stimulus amplitude ζ as a parameter in addition to the stimulus frequency η . For this reason , we searched for the stimulus amplitude where the model best fits the data . The calculated 15 largest Lyapunov exponents from our experimental data are all positive due to the background noise . For our model , this is the case only if chaos arises ( see Figure 4 ) . The absolute values of the Lyaponov exponents can therefore not be compared directly . If , however , we assume that the unpredictability of the experimental data is partially due to background noise ( which does not depend on the stimulus frequency ) and partially due to the intrinsic dynamics of the modeled system , it makes sense to compare the pattern of dependency of the Lyapunov exponent from the stimulus frequency instead . For this reason , we compared the largest Lyapunov exponents as computed from the model to the largest Lyapunov exponent computed from the data , normalized to the same range as the model-based exponent ( −9 . 6972·10−2≤λ1 , Model≤−9 . 4435·10−5 ) , by a shift-and-scale transformation u+v·λ1 . For the means and standard deviations of u and v , please refer to supplementary Figure S2 . Interestingly , the three subjects for whom the individual fits were not significant ( number 3 , 6 and 7; see Results section ) are clearly noticeable here in terms of means and standard deviations of u and v . The offset of the experimental Lyaponov exponents u can be regarded as a multiplicative process R of divergence S1 ( κ ) , because |S1 ( κ ) |≈R1•exp ( v κ·λ1 ) with R1 = exp ( u ) . Background activity and more general unspecified processes may be included in R . Since the ratio between the sampling rates on the frequency axis is 4 . 6 between the model ( 69 samples ) and the experimental data ( 15 samples ) , we compare an experimental data point with the four nearest neighbors in the model . The comparison and detection of the model configuration that fits the experimental data best comprise seven steps: ( i ) select the four nearest neighbors along the stimulus frequency axis of a amplitude configuration of our model to a query point in the experiment ( i . e . , the response to an experimentally applied ratio of stimulus to intrinsic alpha frequency ) , ( ii ) calculate the Euclidean distances in the plane spanned by frequency and largest normalized Lyapunov exponent between each of the nearest neighbors and the experimental data point ( this way an agreement between model and data Lyaponov exponents is weighted according to the agreement between the frequencies they belong to ) , ( iii ) determine the maximum Euclidean distance between the nearest neighbors and the experimental data point , where the largest normalized Lyapunov exponent of a nearest neighbor is set to the lower ( min ( λ1 , Model ) = −9 . 6972·10−2 ) or to the upper bound ( max ( λ1 , Model ) = −9 . 4435·10−5 ) of the model-based largest Lyapunov exponent if the query point in the experiment is greater than −4 . 8533·10−2 ( i . e . , min ( λ1 , Model ) /2+max ( λ1 , Model ) /2 ) or not , respectively , ( iv ) calculate the relative error as the ratio of the distance of a nearest neighbor to its maximum distance , ( v ) detect the nearest neighbor with the minimum relative error for each experimental data point and average these errors over all 15 data points ( for the different experimental frequencies ) , ( vi ) repeat steps ( i ) to ( iv ) for each amplitude of the model stimulation , and ( vii ) find the model configuration with the stimulus amplitude that fits the data best by detecting the minimum of the averaged minimum relative errors ( i . e . , mean error ε in graph B of Figure 8 ) . In order to test the significance of the comparison results , we ( i ) compute Pearson's ( linear ) correlation coefficient for each stimulus amplitude ζ as well as for each subject ( and also for the average over subjects ) between the largest normalized Lyapunov exponents of model and data as function of the ratios of stimulus to intrinsic alpha frequency , and ( ii ) test its significance by applying a Student's t-test . Due to the multiple comparisons of the 106 amplitudes and each four nearest neighbors , we used the Bonferroni correction for the significance level p = 0 . 05 corrected by p′ = p/ ( 4·106 ) . Moreover , in order to further substantiate our findings , we performed a bootstrap test . We randomized the sequence of experimental Lyapunov exponents along the frequency axis so that their distribution remained the same but any putative frequency-dependence was destroyed . Then we applied our fit method and recorded the fit error . We repeated this 5000 times and obtained an estimate of the error distribution . Counting the occurrences of errors that are below the one obtained with the true sequence of frequencies , we obtained an estimate of the probability that such an error could have been achieved by chance .
|
Neuroscience aims to understand the enormously complex function of the normal and diseased brain . This , in turn , is the key to explaining human behavior and to developing novel diagnostic and therapeutic procedures . We develop and use models of mean activity in a single brain area , which provide a balance between tractability and plausibility . We use such a model to explain the resonance phenomenon in a photic driving experiment , which is routinely applied in the diagnosis of various diseases including epilepsy , migraine , schizophrenia and depression . Based on the model , we make predictions on the outcome of similar resonance experiments with periodic stimulation of the patients or participants . Our results are important for researchers and clinicians analyzing brain or behavioral data following periodic input .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"neuroscience",
"cognitive",
"neuroscience",
"mathematics",
"computational",
"neuroscience",
"electroencephalography",
"circuit",
"models",
"clinical",
"neurophysiology",
"biology",
"nonlinear",
"dynamics",
"visual",
"system",
"diagnostic",
"medicine",
"physiology",
"sensory",
"systems",
"computational",
"biology"
] |
2011
|
Modeling Brain Resonance Phenomena Using a Neural Mass Model
|
Opisthorchis felineus , O . viverrini , and Clonorchis sinensis ( family Opisthorchiidae ) are parasitic flatworms that pose a serious threat to humans in some countries and cause opisthorchiasis/clonorchiasis . Chronic disease may lead to a risk of carcinogenesis in the biliary ducts . MicroRNAs ( miRNAs ) are small noncoding RNAs that control gene expression at post-transcriptional level and are implicated in the regulation of various cellular processes during the parasite- host interplay . However , to date , the miRNAs of opisthorchiid flukes , in particular those essential for maintaining their complex biology and parasitic mode of existence , have not been satisfactorily described . Using a SOLiD deep sequencing-bioinformatic approach , we identified 43 novel and 18 conserved miRNAs for O . felineus ( miracidia , metacercariae and adult worms ) , 20 novel and 16 conserved miRNAs for O . viverrini ( adult worms ) , and 33 novel and 18 conserved miRNAs for C . sinensis ( adult worms ) . The analysis of the data revealed differences in the expression level of conserved miRNAs among the three species and among three the developmental stages of O . felineus . Analysis of miRNA genes revealed two gene clusters , one cluster-like region and one intronic miRNA in the genome . The presence and structure of the two gene clusters were validated using a PCR-based approach in the three flukes . This study represents a comprehensive description of miRNAs in three members of the family Opistorchiidae , significantly expands our knowledge of miRNAs in multicellular parasites and provides a basis for understanding the structural and functional evolution of miRNAs in these metazoan parasites . Results of this study also provides novel resources for deeper understanding the complex parasite biology , for further research on the pathogenesis and molecular events of disease induced by the liver flukes . The present data may also facilitate the development of novel approaches for the prevention and treatment of opisthorchiasis/clonorchiasis .
Opisthorchis felineus , O . viverrini , and Clonorchis sinensis ( class Trematoda; order Plagiorchiida; family Opisthorchiidae ) are parasitic flatworms with complex life cycles , which include three hosts , with human and piscivorous mammals as definitive hosts [1] . These three flukes cause diseases of the hepatobiliary system , referred to as opisthorchiasis/clonorchiasis . These diseases are characterized by chronicity and severe consequences , some of which are cancers of the biliary tract and liver [2–5] . C . sinensis is endemic in China , Taiwan , Vietnam , Korea , Japan , the Lao People's Democratic Republic and the Russian Far East; O . viverrini is found in Cambodia , the Lao People's Democratic Republic , Thailand , and Vietnam; and O . felineus is spread in the former Soviet Union ( Ukraine , Belarus , Kazakhstan , the Baltic Republics and Russia , particularly Western Siberia ) and some European countries [6 , 7] . Recently , many studies focusing on the developmental biology of the opisthorchiid flukes and the molecular mechanism of their pathological effects on host organisms were conducted using advanced genomic and transcriptomic techniques . For example , protein-coding transcriptomes have been well characterized for O . felineus [8] , O . viverrini [9 , 10] and C . sinensis [9 , 11 , 12] , allowing investigations of diverse issues of the host-parasite interaction at the molecular and cellular levels as well as indicating the diagnostic potential of particular proteins from the excretory secretory products ( ESP ) of the flukes . However , the microRNA-containing transcriptomes , which are known to dramatically influence many protein patterns , have not been comprehensively studied to date in opisthorchiid flukes . It is well known that microRNAs ( 18–22 nucleotide , non-coding RNAs ) are able to down-regulate target mRNA expression at the post-transcriptional level in multicellular animals and thus play important roles in many biological processes including development , differentiation , viral defense and apoptosis [13] . A miRNA becomes mature after processing of its stem-loop precursors by RNase III enzymes with short miRNA duplex generation . In addition , miRNA becomes functionally active upon detachment from its complement ( miRNA* ) in the duplex during integration into RNA-induced silencing complexes ( RISC ) [13 , 14] . Both the miRNA and the miRNA* are potentially functional in the RISC [15–18]; however , only one miRNA remains functional , and the other degrades [19–21] . The RISC-containing miRNA induces translational repression or the degradation of the target mRNA by binding to its 3’-UTR [14 , 22] . Increasing evidence shows that the action of miRNAs has great importance and broad roles in pathogen-host interactions and the regulation of immunity against infectious agents [23] . Recently , miRNAs have been detected circulating outside of cells in the serum within exosomes or in association with specific proteins [24] . These extracellular RNAs are stable in bodily fluids [24] and are involved in cell-to-cell communication [25 , 26] . Therefore , they have attracted attention as biomarkers of disease [26 , 27] . Moreover , parasite-derived miRNAs have recently been identified in the serum of hosts infected with Schistosoma mansoni [28] and in exosome-like vesicles in the ESP ( Dicrocoelium dendriticum ) [29] . MiRNA manipulation in parasites has been also proposed as a new strategy for controlling schistosomiasis and cystic echinococcosis [23] . Parasite miRNA studies have thus become promising for elucidating the molecular mechanisms of parasitic diseases and for the development of more specific diagnostic tools [30] . In the last decade , numerous miRNAs have been discovered in several flatworms species , such as Schmidtea mediterranea [31–33] , Dugesia japonica [17 , 34] , Orientobilharzia turkestanicum [35] , S . mansoni [36 , 37] , S . japonicum [38–40] , C . sinensis [41] , Eurytrema pancreaticum [42] , Echinococcus granulosus , E . multilocularis [43] , Fasciola gigantica , F . hepatica [44] , D . dendriticum [29] , Hymenolepis microstoma [45] , Taenia saginata [46] and Gyrodactylus salaris [47] . Most of the miRNAs of E . granulosus , E . multilocularis , S . japonicum , S . mansoni and S . mediterranea have been described and are well annotated in miRBase ( Release 21: June 2014 ) . All proteins necessary for miRNA maturation and miRNA-induced silencing were identified in several flatworms species , for example , in S . mansoni [48] . The set of orthologous proteins were also found in opisthorchiid species [10 , 12 , 49] . So the description of miRNA transcriptomes of opisthorchiids is necessary for understanding gene expression and function in these parasites . The aims of the present study were to identify the miRNAs of O . felineus , O . viverrini and C . sinensis , describe respective miRNA genes and provide a basis for further investigations of the roles of miRNAs in the regulation of gene expression in liver flukes .
For the detection of small RNAs of the three opisthorchiids , an enrichment technique consisting of the selective fractionation of RNA ( 18–200 nt ) in polyethylene glycol solutions of various concentrations was used as described by Wang et al . [50] . The size distribution of the RNA molecules was analyzed by micro-electrophoresis with a BioAnalyzer ( Agilent ) . The miRNA libraries were constructed using an Ambion® SOLiD Small RNA Expression Kit . For each sample , three libraries ( technical replicates ) were sequenced: two with Adaptor Mix A ( yields the template for SOLiD sequencing from the 5' end of the sense strand ) and one with Adaptor Mix B ( yields the reverse complement sequence ) . The cDNA libraries were produced using 200 ng of the small RNA fraction , following the protocol supplied with the kit , and amplified using barcoded primers and 17 PCR cycles for Mix A libraries and 15 PCR cycles for Mix B libraries . Amplified products were concentrated using the Fermentas® GeneJET PCR Purification Kit and gel purified using 6% acrylamide gels . Gel pieces containing PCR products of ~105–150 bp were excised , libraries were eluted by 5M ammonium acetate and cleaned by ethanol precipitation . Each library was diluted to a concentration of 0 . 5 pM for full-scale template bead preparation . Approximately 40 million beads for each sample were deposited on ¼ slide of the SOLiD 3 . 5 System and sequenced in 35-base runs . Sequencing was performed at the Siberian Branch of Russian Academy of Science ( SB RAS ) Genomics Core Facility . The library designations with corresponding GenBank database accession numbers are: C . sinensis—A1 ( SRX817942 ) , rA1 ( SRX817990 ) , B1 ( SRX817989 ) O . viverrini—A2 ( SRX817991 ) , rA2 ( SRX817993 ) , B2 ( SRX817992 ) O . felineus AdultNoEggs—A3 ( SRX817994 ) , rA3 ( SRX817996 ) , B3 ( SRX817995 ) Metacercaria—A4 ( SRX817997 ) , rA4 ( SRX817999 ) , B4 ( SRX817998 ) Adult+Eggs—A5 ( SRX818000 ) , rA5 ( SRX818002 ) , B5 ( SRX818001 ) The pipeline of the computational search for conserved and novel miRNAs in the opisthorchiid species is presented in Fig 1 . First , quality filtering of the sequences was performed using the SOLiD preprocess filter [51] using the following parameters: Min count for Polyclonal Analysis—1 , Min QV for Polyclonal Analysis—25 , Max count permitted errors—100 , Max QV to consider an error—10 , Removal of reads with negative QV score—y , and Truncation—off . The adapter fragments were removed by cutadapt v . 0 . 9 . 5 [52] with a maximum error rate of 12 . 0% and a minimum read length of 18 bp . To remove possible fragments of messenger and non-microRNA sequences , we mapped the reads to mRNA sequences in Refseq ( rel . 106 ) [53] , mRNA sequences of plathyhelmints and nematode taxa from the GenBank database ( December , 2011 ) [54] , and sequences from Rfam ( rel . 10 ) , [55] excluding miRNAs using BFAST [56] . The BFAST program was chosen , since it allows the mapping of short reads and uses the Smith-Waterman method , with gaps to support the detection of small indels at its final processing stage . This step improves the sensitivity of alignment , which , in our case , is important for mapping reads to genomes from different species . A significant advantage of this approach is that the alignment of sequences in the SOLiD 2-base color coding reduces the influence of sequencing errors . The following BFAST parameters were used: editing distance ( the number of substitutions/insertions/deletions allowed in read alignment ) ≤ 2 , multiple mapping of reads was allowed , and other parameters were set as default . All reads mapped to these databases were removed from further analysis . To identify conserved miRNAs , the remaining reads were mapped to animal pre-miRNA sequences in miRBase ( Release 21: June 2014 ) [57] using BFAST with the following parameters: editing distance ≤ 4 , multiple mapping of read was allowed , and other parameters were set as default . To identify genome-specific sequences of known miRNAs , we performed additional mapping of reads similar to miRBase sequences onto C . sinensis [58] , S . mansoni ( rel . 4 ) [59] and S . japonicum ( rel . 2 ) [60] genomes with editing distances ≤ 2 . To verify that these sequences can form pre-miRNA hairpins within their genomic context , the secondary structures of these candidate pre-miRNAs were reconstructed using the UNAFold program [61] . Two variants of the candidate pre-miRNA sequences were selected . The first variant spans from 50 bp upstream to 10 bp downstream of the miRNA region . The second variant spans from 10 bp upstream to 50 bp downstream of the miRNA region . We inferred miRNA sequences that met the following criteria: ( 1 ) ΔG ≤ -20 kcal / mol; ( 2 ) the fraction of paired nucleotides in the hairpin corresponding to the mature miRNA is > 70%; ( 3 ) no branching interactions for the hairpin forming nucleotides are allowed; ( 4 ) the sequence of miRNA is not in the terminal loop; and ( 5 ) the difference in the side lengths of internal loops and bulge size is not more than two nucleotides [62 , 63] . To identify novel miRNAs , the small RNA-like reads without similarity to sequences in miRBase were mapped to the genomic sequences of C . sinensis , S . mansoni , and S . japonicum . RepeatMasker ( http://www . repeatmasker . org/ ) was used to mask repeats and regions with low complexity in the genomes . We used BFAST with an editing distance of ≤ 2 , filtered out multiple mapped reads , and other parameters were as default . Genomic regions with lengths of ≤ 25 bp that were covered by at least three reads were considered as candidates for novel species-specific miRNAs . To verify the stem-loop pre-miRNA secondary structure of these sequences , we applied UNAFold analysis for their two extended sequence variants . The sequences meeting the above mentioned secondary structure criteria were considered as novel miRNA candidates . To estimate reproducibility of technical replicates , the Spearman's rank correlation coefficients of normalized ( RPKM ) expression level of several conserved miRNAs ( that are common for three flukes ) were established using Past3 [64] ( S1 Table ) . Conserved miRNAs were used in reproducibility analysis because new miRNAs have low non-normalized expression levels ( around three reads were mapped to the genome for each new miRNA; therefore , the novel miRNAs were not detected in all technical replicates ) . Additional similarity searches were performed using the BLAST [65] . To detect violations of one of the criteria of the conservative cluster definition ( cluster of miRNAs should be a group of miRNA precursors expressed as a polycistronic unit [66] ) we applied the protein coding gene-finding procedure using the Fgenesh program [67] . The alignments of some miRNAs ( two miR-71/ miR-2 clusters , miR-1 , miR-133 , and miR-190 ) with sequences of these miRNAs orthologs ( obtained from S . mediterranea , G . salaris , S . mansoni , S . japonicum , E . granulosus , E . multilocularis , H . microstoma and T . solium genomes ) were performed using the program CLUSTALW [68]; miRNA sequences of T . solium , namely miR-1 , miR-2b , miR-2c , miR-71 , miR-133 , miR-190 , were obtained by homology search of these miRNAs in T . solium genome ( http://www . genedb . org/Homepage/Tsolium ) using the BLAST [65] . All time-consuming computations were performed using a high-throughput computing system at the Joint Access Center for Bioinformatics and a computational cluster at the Novosibirsk State University . The following primers were used for the amplification of genomic regions hosting the miRNA genes of the three opisthorchiid species: clust1-for1 ( 5'-CACAGCCAGTATTGATGAAC-3' ) , clust1-for2 ( 5'-ACAGCCCTGCTTGGGACAC-3' ) , clust1-rev ( 5'-CCAAAGCTTGGACTGTGAT-3' ) , clust2-for ( 5'-AAAGACTTGAGTAGTGAGACGCT-3' ) , clust2-rev ( 5'-TCGTCACCTAAGCAGGACT-3' ) , Cl1-F ( 5'-CGCAAGTGATCAATGTTTTCCTC-3' ) and Cl1-R ( 5'-GCGCACCAACGGCCTAA-3' ) . The amplification was conducted using a DNA thermal cycler ( Mastercycler gradient Eppendorf ) as follows: initial denaturation at 95°C for 2 min , followed by 35 amplification cycles ( 95°C for 25 s , 56°C for reactions with clust1-rev , clust1-for1 , clust1-for2 , Cl1-F and Cl2-R and 53°C for reactions with clust2-for and clust2-rev for 30 s , 72°C for 30 s ) and a final extension cycle ( 72°C for 5 min ) . PCR products were analyzed by agarose gel ( 2% ) electrophoresis . Purification of PCR products was performed by the method of Exo-TsAP . To 20 μl of PCR product were added 1 μl of Exonuclease I and 1 μl of Thermosensitive Alkaline Phosphatase , followed by an incubation for 15 min at 37°C and then 15 min at 80°C . Sequencing reactions were performed using the BigDye ® Terminator v3 . 1 Cycle Sequencing Kit according to the manufacturer's instructions and analyzed at the SB RAS Genomics Core Facility .
A three-step mapping and filtering procedure was applied to the reads ( Fig 1 ) generated from the 15 libraries to obtain the pool of small RNA-like sequences for the three opisthorchiid species . The results of filtering are given in S2 Table . For O . felineus , the sequencing of nine libraries generated 446 million reads that were distributed as follows: 131 millions for three Adult+Eggs libraries , 152 millions for three AdultNoEggs libraries , and 162 millions for three Metacercaria libraries . For C . sinensis , three libraries were sequenced and 126 million reads were obtained . For O . viverrini , three libraries were sequenced with 150 million reads obtained . After filtering low quality tags , including 5′ and 3′ adaptors and adaptor-adaptor ligation products , a total of 279 million reads with high quality were retained for O . felineus ( Adult+Eggs ( 84 million reads ) , AdultNoEggs ( 108 million reads ) , Metacercaria ( 87 million reads ) ) , 75 million reads for C . sinensis , and 83 million for O . viverrini . Among the clean reads , an average of 13 . 6% were found to be rRNA , tRNA , snRNA , and snoRNA , when searched against the Refseq/Rfam databases . The percentage of the remaining reads mapping to miRBase sequences averaged 2 . 85% . Spearman's rank correlation coefficient analysis showed high reproducibility between A and rA libraries ( ~ 0 . 9 ) and somewhat less reproducibility between B and either A or rA libraries ( ~ 0 . 8 ) , which might be explained by the fact that rA libraries were exact technical replicates of A libraries whereas B libraries were created using another adaptor . The miRNA was regarded as conserved if it had an ortholog in another animal species . The ortholog search for the miRNAs of the three opisthorchiids yielded 19 conserved miRNAs belonging to 13 families ( bantam , let-7 , miR-1 , miR-2 , mir-7 , miR-10 , miR-36 , miR-46 , miR-71 , miR-124 , miR-125 , miR-133 , and miR-190 ) ( Fig 2A , Table 1 , S3 Table ) . Most families included one miRNA variant , but the miR-71 family consisted of two variants and the miR-2 family comprised five variants . Interestingly , the expression of miRNA* from the duplex carrier strands for two miRNAs ( let-7 and miR-10 ) was also found ( S3 Table ) . Sixteen conserved miRNAs were identified as common in all three opisthorchiids . Additionally , miR-281 ( miR-46 family ) was found in two species—O . felineus and C . sinensis . There were also conserved miRNAs either in O . felineus only ( miR-10 ) or in C . sinensis only ( miR-36b ) ( Fig 2A; Table 1 ) . Eighteen conserved miRNAs were identified for O . felineus when combining the Adult+Eggs , AdultNoEggs , and Metacercaria samples . Individual analyses of the O . felineus samples ( Adult+Eggs , AdultNoEggs , Metacercaria ) revealed differences in miRNA composition between the samples . Fourteen of the eighteen O . felineus miRNAs were identified in all three samples . Two miRNAs ( bantam and miR-281 ) were identified in AdultNoEggs and Adult+Eggs samples only , but not in the Metacercaria sample . miR-7 was detected in AdultNoEggs and Metacercaria samples but not in the Adult+Eggs sample , and miR-10 was found in the Metacercaria samples only ( Fig 2B , Table 2 ) . The mapping results demonstrated that most of the conserved miRNA sequences identified in the present study are common among opisthorchiid and schistosome species , which was expected . Candidate sequences for novel miRNAs ( S4 Table ) were selected from reads without matches to miRBase sequences after mapping them to the C . sinensis genome and processing the genomic fragments encompassing the resultant hits through the secondary structure filter ( see Materials and Methods ) . We identified 43 such miRNAs for O . felineus , 20 for O . viverrini and 33 for C . sinensis . The occurrence of novel common and species-specific miRNAs in the samples from the three Opisthorchiidae species is presented in Fig 2C and S4 Table . Interestingly , most of these novel miRNAs were species-specific . Only one miRNA ( new_miR-001 ) had orthologs in all three species . The greatest number of novel specific miRNA candidates was identified for O . felineus ( 83% ) ; however , the fraction of unique species-specific miRNAs was highest for O . viverrini ( 95% ) . Forty-three novel miRNAs were obtained for O . felineus when combining the Adult+Eggs , AdultNoEggs and Metacercaria samples . The distribution of stage-specific and stage-nonspecific novel miRNA candidates in O . felineus demonstrated that no common miRNAs were identified in all three sample types , and only two of the 43 novel miRNAs were identified in more than one stage/body part ( Fig 2D ) . Mapping the conserved miRNAs onto the C . sinensis genome provided evidence supporting the presence of two miRNA clusters: miR-71a/miR-2a/miR-2b/miR-2e ( miR-71a/2 ) and miR-71b/miR-2d/miR-2c ( miR-71b/2 ) . Homologous clusters have been previously described for seven flatworms ( E . granulosus , E . multilocularis , G . salaris , H . microstoma , S . mediterranea , S . japonicum and S . mansoni ) [45 , 47] , and in the current study , were also found in the T . solium genomic sequences ( Fig 3 , S1 Appendix ) . We compared the structures of these miRNA clusters from the C . sinensis genome with homologous sequences from the flatworm genomes mentioned ( Fig 3 ) . All mature miRNA sequences of miR-71 family were located in the 5’ arm of their precursors , while all mature miRNAs of the miR-2 family were located in the 3’ arm [69] . In the miR-71a/2 cluster group , the distance between the mature miR-71 and the nearest miR-2 varied from 104 to 121 bp; in the miR-71b/2 cluster group , this distance ranged from 121 to 150 bp . The minimal distance between the two mature sequences of miR-2 family was found in the miR-71a/2 group ( 71 bp between miR-2a and miR-2b in S . japonicum ) ; the maximal distance was found in the miR-71b/2 group ( 101 bp between miR-2a and miR-2d in G . salaris ) [45 , 47] . The miR-71a/2 cluster mapped to C . sinensis contig 2339 and spanned 441 nucleotides ( from 103849 to 104290 ) , with the miRNA order the same as in both Schistosoma genomes . A comparative analysis of the miR-71a/2 cluster genomic organization among the flatworms revealed three distinct types ( Fig 3A ) . The first type , comprising the precursors for miR-71a and the three miR-2 isoforms , was exemplified by clusters from the genomes of C . sinensis , S . japonicum , and S . mansoni . The second type , consisting of the precursors for miR-71 and the two miR-2 isoforms , was represented in the genomes of the cestodes H . microstoma , E . granulosus and E . multilocularis . The third type , with the precursors for miR-71 and only one miR-2 isoform , was observed in the monogenean G . salaris and the planarian S . meditteranea ( in three genomic copies ) [45 , 47] . The miR-71b/2 cluster mapped to C . sinensis contig 2957 and spans 416 nucleotides ( from 323569 to 323984 ) . Detailed analysis of the cluster sequences in three trematode , one monogenean and one turbellarian genome resulted in the discovery of the precursor for miR-2f in C . sinensis ( Fig 3B , S5 Table ) . This miRNA was previously described for two schistosomes [70 , 71] . The miRNA order in these orthologous clusters was also well conserved in the C . sinensis , S . japonicum and S . mansoni genomes ( Fig 3B ) . It should be noted that sme-miR-752 , although not formally assigned to the mir-2 family , is recognized as having evolved from miR-2 [47] . Because the mature miRNA sequences of the two clusters were identical among the three opisthorchiid species , we designed two primer sets to experimentally prove the presence of the clusters and partially structure the clusters using PCR amplification of corresponding regions in the three genomes . To amplify a fragment of cluster miR-71a/2 , the primer set clust1-for1 , clust1-for2 , clust1-rev was used ( Fig 4A , S2 Appendix ) . For the miR-71b/2 cluster , the primer set clust2-for , clust2-rev was employed ( Fig 4C , S2 Appendix ) . The electropherogram presented in Fig 4D , 4E and 4F ) show the PCR products generated using these primer sets with DNA templates prepared from C . sinensis , O . felineus and O . viverrini . The sequence alignments of the corresponding genomic regions of the three opisthorchid species revealed specific variable positions: 8 per 387 nucleotides in cluster miR-71a/2 and 11 per 299 nucleotides in cluster miR-71b/2 ( S3 Appendix ) . These variable positions were located mainly in the regions corresponding to the ends of the pre-miRNA , the terminal loops of the pre-miRNA and the spacers between miRNA precursors . It is worth noting that miR-2f from cluster miR-71b/2 of C . sinensis , S . japonicum and S . mansoni ( Fig 3 ) was also discovered in the respective clusters of O . felineus and O . viverrini . Furthermore , the alignment of this genomic region demonstrated high conservation among the three opisthorchiid species: only four variable positions ( which were located closer to the precursor of miR-2d ) per 154 nucleotides were found . To experimentally prove the overall structure of the miR-71a/2 clusters in the three species , we designed primers Cl1-F and Cl1-r , which are capable of amplifying the genomic regions encompassing the clusters , using the only available sequences for C . sinensis ( Fig 4B , S1 Appendix ) . The results are presented in the electrophoretogram ( Fig 4G ) . The sequencing of the three species-specific amplicons ( S4 Appendix ) allowed us to determine the four pre-miRNA sequences for each of the three flukes . The secondary structures of these pre-miRNAs were estimated by UNAFold ( S2 Fig ) . The results of UNAFold demonstrated that the nucleotide substitutions discriminating the pre-miRNA sequences of each of the opisthorchiid species exerted minor or no effects on the pre-miRNA secondary structures . Upon analysis by Jin et al . [45] , the genomic regions with matches for miR-1 and miR-133 were designated as orthologous miRNA gene clusters in three flatworms , namely the cestodes E . granulosus , E . multilocularis and H . microstoma . We extended this list of flatworm species by demonstrating that C . sinensis and S . mansoni also have similar genomic regions . It should be mentioned that miR-133 were not annotated for S . mansoni in previous reports [36 , 37 , 71] . However , we found sequences highly similar to this miRNA in read archives ( ERR278825 , ERR278826 , ERR278827 , ERR278828 ) using a BLAST search . The UNAFold secondary structure prediction for the precursors of the conserved miRNAs showed no canonical structure for the putative S . mansoni pre-miR-133 , which could possibly explain the delay in sma-miR-133 annotation ( S6 Table ) . Our alignment analysis did not show complete conservation over these regions of the five genomes . Remarkably , large spacers were detected between the sites matching the miRNAs , ranging from 11705 bp in E . multilocularis to 34008 bp in C . sinensis ( Fig 5 , S5 Appendix ) . Hence , we referred to the regions as “cluster-like regions miR-1/miR-133” . To elucidate the content of the spacers in genomes of five parasitic flatworms , we employed the gene prediction program Fgenesh [67] using S . mansoni-specific gene-finding parameters and found few unannotated ORFs without significant similarity among the species ( S6 Appendix ) . We then explored the genomic context beyond the cluster-like regions miR-1/miR-133 in the five flatworm species using information from the C . sinensis database ( http://fluke . sysu . edu . cn/CsinGeno/home . php ) , NCBI ( http://www . ncbi . nlm . nih . gov ) ( for S . mansoni ) and Genedb ( http://www . genedb . org/Homepage ) ( for E . granulosus , E . multilocularis and H . microstoma ) . We found that miR-133 is located near a gene encoding one of several Mind bomb proteins in all five genomes . We also found that miR-1 mapped near a gene encoding another Mind bomb protein in the genomes of S . mansoni , E . granulosus and E . multilocularis ( S7 Table ) . Although these miRNA sites were conservatively linked ( forming a putative synteny group ) , the inter-microRNA distances exceeded 10 kb and contained putative genes . Altogether , the features suggested that the expression of these two miRNAs was unlikely as a single transcriptional unit in either genome . Therefore , we concluded that the case under consideration did not adhere to the conservative definition for a miRNA gene cluster [66] . The mapping of miR-190 , which was also identified in the three opisthorchiid species , to the available flatworm genomic sequences showed that this miRNA is located in an intron of the gene encoding the talin protein . Therefore , we could classify this miR-190 as intronic [72] . It is noteworthy that , despite some variability in the nucleotide content of the talin exons surrounding the intronic miRNA ( S7 Appendix ) , the overall protein structure was conserved enough ( S8 Appendix ) to ensure a reliable comparative analysis of the gene structure ( Table 3 ) . The intronic miRNA showed the motives corresponding to both mature miR-190 and miR-190* ( S9 Appendix ) . The alignment depicted the sites with high conservation in either the flatworm class and those with evident inter-class variations , which , nevertheless , likely did not hamper the intronic miRNA’s ability to form the necessary secondary structure and effectively undergo maturation .
Using deep sequencing with SOLiD technology , we have identified 88 novel and 19 conserved miRNAs in three liver flukes of the family Opisthorchiidae—C . sinensis , O . felineus , and O . viverrini . The discovery of the novel opisthorchid-specific miRNAs is interesting , since they could be responsible for some opisthorchid-specific features of their parasitic life style including some pathogenicity features in definitive hosts . Interestingly , the number of the novel species-specific candidate miRNAs identified in the opisthorchiid flukes was larger than that of conserved miRNAs . This may relate to a low coverage of individual novel miRNAs ( three reads were mapped to the genome for each new miRNA ) . Nevertheless , it is worth noting that similar species-specific/conserved miRNA families ratios ( miRBase Release 21: June 2014 ) are also observed for other trematodes—S . japonicum ( 28/22 ) and S . mansoni ( 82/22 ) . The same is seen also for free living planarian S . mediterranea ( 45/44 ) . The addition of our data on the miRNAs of opisthorchiid flukes likely raises the question as to whether an excess of novel ( species-specific ) miRNAs compared with conserved ( family- , class- , and phylum-specific ) miRNAs could relate to a difference in life style ( free vs . parasitic ) , as proposed previously [36] . It seems that even a less significant difference in parasitic style between schistosomes or opisthorchiids is associated with generation of numerous family- , genus- , and species-specific miRNAs . More additional data for flatworms’ miRNAs are needed to elucidate the evolutionary and biological significance of species-specific miRNAs of parasitic flatworms . The identification of 19 conserved miRNAs in three liver flukes of the family Opisthorchiidae has strengthened the results of previous attempts to explore miRNAs of liver flukes . It should be noted that , in a previous study of C . sinensis , numerous miRNA-like sequences were found among reads generated with high-throughput sequencing using Solexa/Illumina technology [41] . However , the authors had no opportunity to carry out mapping the miRNA-like sequences on to the C . sinensis genome to achieve a confident assignment of their miRNA-like sequence sets to miRNA families annotated in miRBase . Therefore , we now provide the results of miRNA-like sequence mapping on to the C . sinensis genome , thus improving the reliability of miRNA identification for members of the Opisthorchiidae . Furthermore , we provide the results of the miRNA family classification . The occurrence of the 19 conserved miRNAs in organisms of various taxa ( including 10 miRNAs out of 34 ones arisen after “bilaterian expansion” [72] ) is presented ( S8 Table ) . We should mention two curious C . sinensis miRNAs: the reads corresponding to csi-miR-36b were found in our study but were not found by Xu et al . [41] , and the reads corresponding to miR-10 as indicated by Xu et al . [41] were readily mapped in the C . sinensis genome , but were detected in our study for O . felineus only . Perhaps these cases need further investigation . The mapping of 19 conserved miRNAs on to the three genomes available for Trematoda ( C . sinensis , S . mansoni and S . japonicum ) are presented in Table 4 . Interestingly , there was some shortfall in hits for few miRNAs after the mapping of sequencing data . This could be due to the incompleteness of either genome assembly ( miR-125 was not found in C . sinensis genome ) or indeed by the species specificity of miRNA genes ( we did not find the opisthorchid miR-1 in S . japonicum genome , we also did not locate opisthorchid miR-36b in either schistosome genome ) . Our analysis of the genomic organization of the opisthorchid miRNA genes confirmed the presence of gene clusters and intronic miRNAs . It is known that the miR-71/miR-2 cluster , which we experimentally proved to be in two copies in opisthorchiids ( like in other parasitic trematodes studied ) is present as one copy in parasitic cestodes , and five copies in the free-living planarian S . mediterranea [45] . This variation in the number of miR-71/miR-2 clusters in the genomes of representative flatworms of different classes could not be explained by the biology of the organism or by the reduction of targets for these miRNAs . The parasitic nematode Ascaris suum and Brugia malayi display one miR-71/miR-2 cluster , while the freeliving Caenorhabditis species have either one or no such cluster ( miRBase Release 21: June 2014 ) . Therefore , it seems that the miR-71/miR-2 cluster evolution proceeded differently in the Nematoda compared with the Platyhelminthes , and the details of the evolution remains to clarify in further studies . Both clusters miR-71a/miR-2a/miR-2b/miR-2e and miR-71b/miR-2f/miR-2d/miR-2c were conserved , suggesting their functional importance in all three opisthorchiid species ( Fig 3 ) . To date , some miRNAs belonging to miR-71 and miR-2 families are known to have female-biased expression in S . mansoni [71] and to play an important role in regenerative processes in planarian [73] . Also the miR-2 family miRNAs are probably involved in neural development and maintenance in Drosophila melanogaster and C . elegans [69] . Their detection in exosome-like vesicles in the ESP of the liver fluke D . dendriticum leads to a speculation about the possible implications of trematode miRNAs in the modulation of parasite-host interactions by a new means of regulating host gene expression [29] . The fact that expressed sequences ( reads ) corresponding to miR-2f have not been detected in opisthorchiids requires further studies to explain why the expression pattern of this putative miRNA is so strikingly different from that of its neighbors . In previous papers , the combination of the miR-1/miR-133 miRNA genes was described also as a miRNA cluster for many animal species ( see data in miRBase ) [74] including flatworms [45] . However , this combination in Drosophila genomes has been shown to escape the conserved cluster definition [66] . Hence , due to the distance between the sites corresponding to the miRNAs in flatworms , as well as the capability to predict protein-coding genes in between these sites , we suggest referring to these regions as “cluster-like regions miR-1/miR-133” , which form a putative synteny group . The next miRNA cluster that should be discussed is let-7/miR-100/miR-125 . Its main characters are conserved in almost all Deuterostomia taxa . However , in Protostomia , many variations of its structure have been discovered , while in some animals ( Annelida , Trichinella , Arthropoda ) , its general structure is conserved . Important is that the cluster was shown to be disintegrated in flatworms with a complete loss of miR-100 [75] . We can support this conclusion for opisthorchiids also . First , mir-100-like sequences were not detected in the three opisthorchiid species . Second , the combination of let-7/ miR-125 genes is unlikely to exist as a synteny group , as the two miRNA genes map to different chromosomes in S . mansoni ( S9 Table ) . The present analysis corroborates the classification of the miR-190 gene as intronic within the talin gene . The intronic nature of the miR-190 gene has been described for many animals [36 , 45 , 76] . High conservation of the structural ( and maybe functional ) association between miR-190 and the talin protein in platyhelminths appears very interesting and is worthy of further elucidation . Just prior to submission of this manuscript , an article on the O . viverrini genome was published [49] . In the article , the authors predicted in silico 178 conserved miRNA genes . These data will give us the opportunity for a more detailed analysis of O . viverrini miRNA genes , in particular for a comparison of our data based on miRNA real expression with the results of in silico prediction based on genomic sequence analysis . In conclusion , the present study presents the results of large-scale identification and characterization of miRNAs sets encoded in the genomes of O . felineus , O . viverrini and C . sinensis . This first comprehensive comparative analysis of the miRNA genes of these species allowed us to reveal the conserved and species-specific miRNAs in these sets . For several conserved opisthorchiid miRNAs , the genomic organization was analyzed by comparison with orthologous genes in other platyhelminths . The structures of two miRNA gene clusters were experimentally validated for the three opisthorchiid species . The differences in expression level found for some conserved miRNA among the three species and among the three stages of O . felineus stimulate studies to more precisely profile the expression of miRNAs . Finally , the present data provide a sound basis for further studies of the molecular mechanisms of host interactions of opisthorchiids and for development of novel methods to control these neglected parasites .
|
Liver flukes of the family Opisthorchiidae cause diseases of the hepatobiliary system , known as opisthorchiasis/clonorchiasis . The chronic forms of these diseases greatly increase the risk of cancer developing in the biliary ducts . Much has been elucidated regarding the developmental biology of opisthorchiid flukes and the molecular pathological effects on the definitive host; however , the role of microRNAs ( short non-coding RNAs ) capable of influencing the pathogenic process and host-parasite interactions have not yet been comprehensively studied . The aim of the present work was to identify the miRNA genes of the liver flukes and provide a basis for further investigating the roles of these miRNAs in the complex opisthorchiidae life cycle and the pathogenesis of disease .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Identification of microRNA Genes in Three Opisthorchiids
|
KRAS mutant lung cancers are generally refractory to chemotherapy as well targeted agents . To date , the identification of drugs to therapeutically inhibit K-RAS have been unsuccessful , suggesting that other approaches are required . We demonstrate in both a novel transgenic mutant Kras lung cancer mouse model and in human lung tumors that the inhibition of Twist1 restores a senescence program inducing the loss of a neoplastic phenotype . The Twist1 gene encodes for a transcription factor that is essential during embryogenesis . Twist1 has been suggested to play an important role during tumor progression . However , there is no in vivo evidence that Twist1 plays a role in autochthonous tumorigenesis . Through two novel transgenic mouse models , we show that Twist1 cooperates with KrasG12D to markedly accelerate lung tumorigenesis by abrogating cellular senescence programs and promoting the progression from benign adenomas to adenocarcinomas . Moreover , the suppression of Twist1 to physiological levels is sufficient to cause Kras mutant lung tumors to undergo senescence and lose their neoplastic features . Finally , we analyzed more than 500 human tumors to demonstrate that TWIST1 is frequently overexpressed in primary human lung tumors . The suppression of TWIST1 in human lung cancer cells also induced cellular senescence . Hence , TWIST1 is a critical regulator of cellular senescence programs , and the suppression of TWIST1 in human tumors may be an effective example of pro-senescence therapy .
Lung cancer is responsible for more cancer deaths in the US than colorectal , breast , and prostate cancer combined with a dismal overall survival of 15% [1] . The majority of human lung cancers are adenocarcinomas carrying somatic mutations in the genes that encode the EGFR/KRAS/BRAF pathway [2] . Observations in both experimental mouse models and human lung tumors strongly suggest that these pathways are causally responsible for lung tumorigenesis [3] , [4] , [5] , [6] , [7] . KRAS mutant lung adenocarcinomas are generally refractory to conventional cytotoxic therapies [8] and currently available small molecule targeted agents [9] , [10] . Difficulties in pharmacologically targeting K-RAS have resulted in some labeling the protein “undruggable” [11] . Approaches such as using farnesyl transferase inhibitors to prevent prenylation of Ras for its membrane localization have not shown clinical efficacy [12] , [13] . Other potential kinase targets for KRAS mutant tumors have been identified through RNAi screens including: TBK1 , STK33 and PLK1 [14] , [15] , [16] . Rational candidate based approaches that target key pathways required during the process of tumorigenesis for KRAS mutant cancers have not been exhaustive . One such pathway is oncogene-induced senescence ( OIS ) , a failsafe program that prevents normal cells from progressing towards malignancy following introduction of a mutant form of an oncogene such as KrasG12D [17] . OIS is an irreversible cell cycle arrest that is characterized by cells displaying an enlarged , flattened cytoplasm , increase in senescence associated beta-galactosidase ( SA-β-Gal ) activity , increased chromatin condensation and changes in gene expression associated with DNA damage checkpoint proteins or cell cycle checkpoint proteins . OIS is thought to be triggered early during tumorigenesis in order to inhibit aberrant cell cycle progression , preventing pre-malignant tumors from progressing to malignancy [17] . OIS seems to be dependent on the p53-p19ARF , p16-Rb and Atf4-p27 pathways to enforce the senescent phenotype , but the requirement of any or all these pathways is highly context dependent [18] , [19] . Whether these latent OIS programs can be activated in KRAS mutant cancers to result in a clinical effect has only recently been examined [20] , [21] . Recently , Twist1 , a basic helix-loop-helix transcription factor that is central to embryogenesis [22] , has been shown to suppress OIS associated with KrasG12D and EGFR2 oncogenes in vitro in MEFs [23] and pancreatic epithelial cells [24] . Twist1 protein expression is usually undetectable in most adult tissues , but has been shown to be overexpressed in cancers including prostate , bladder pancreatic , osteosarcomas , melanomas and breast [25] , [26] , [27] , [28] , [29] , [30] , [31] . The high expression of Twist1 in cancers strongly correlates with invasive and metastatic tumor cells . Twist1 is thought to regulate epithelial-mesenchymal transition ( EMT ) through the down-regulation of key proteins that maintain epithelial cell characteristics and up-regulation of proteins that confer a mesenchymal phenotype [31] . Thus , Twist1 may act both to induce malignancies early in tumorigenesis and also promote tumor progression [32] . To date , there has yet to be reported an autochthonous model to study the role of Twist1 overexpression in the initiation and maintenance of tumorigenesis . Here we report the generation of such a model and through this demonstration we show an important role of Twist1 in suppressing cellular senescence programs .
To produce a useful tool to address Twist1 functions in vitro and in vivo we generated a transgenic founder line , Twist1-tetO7-luc ( T ) , that harbored the mouse Twist1 cDNA under the control of a bidirectional tetracycline operator sequence ( tetO7 ) also regulating the firefly luciferase gene ( luc ) [33] ( Figure 1A ) . This T founder was crossed to Clara cell secretory protein-reverse tetracycline transactivator protein ( CCSP-rtTA or C ) mice to generate inducible , double-transgenic ( CT ) mouse cohorts . CT mice contain both the rtTA activator expressed primarily in lung alveolar Type II pneumocytes [34] and the tetracycline inducible Twist1-tetO7-luc transgene allowing for spatial and temporal expression of Twist1 and luc ( Figure 1A ) . Inducible regulation in CT mice was verified using serial small animal bioluminescence imaging ( BLI ) and Western blotting , respectively ( Figure 1B–1C ) . Doxycycline drinking water given to CT mice ( CT ON ) induced luciferase and Twist1 expression specifically in the lung only ( Figure 1C ) which reverted to background luciferase and Twist1 expression by 3–7-days after doxycycline withdrawal [34] , [35] , ( Figure 1B–1C ) . To address the functional significance of ectopic Twist1 expression in the lung epithelium global gene expression microarray analysis was performed with induced CT mouse lungs versus wildtype mouse lungs . Notably , after performing gene set enrichment analysis ( GSEA ) [36] with this dataset , we found CT ON lungs had a global gene expression pattern that had a highly significant similarity to two overlapping gene sets for EMT [37] ( Figure 1D and Figure S1A ) and three EMT related phenotypes ( hypoxia , metastasis and wound healing [38] , [39] , [40]; Figure S1B–S1D ) . CT ON lungs showed a subset of epithelial cells appeared to lose E-cadherin and gain vimentin staining by immunofluorescence consistent with an EMT ( Figure 1E and Figure S1E ) , strongly supportive of the gene expression data . Thus , our lung specific CT mouse model is capable of enforcing a Twist1-dependent transcriptional program in lung epithelial cells that is consistent with cells that have undergone EMT . Twist1 has been strongly implicated in tumor progression , but no studies have examined the effect of Twist1 alone for autochthonous tumorigenesis . Twist1 was not a strong oncogene when expressed alone in the lung epithelium . CT ON mice did not develop lung tumors at an increased frequency compared to wildtype FVB/N mice ( Figure 2A ) [41] . Twist1 cooperated dramatically with KrasG12D expression in the lung . CCSP-rtTA/tetO-KrasG12D ( CR ) mice developed multiple synchronous lung tumors , mostly adenomas , with a median tumor latency of 32 weeks [3] , [35] ( Figure 2A and 2E ) . Triple transgenic mice , CCSP-rtTA/tetO-KrasG12D/Twist1-tetO7-luc ( CRT ) , demonstrated a greatly reduced lung tumor latency compared to CR mice , 15 versus 32 weeks ( p<0 . 0001 by log-rank analysis ) ( Figure 2A ) . CRT mice developed numerous lung tumors ( Figure 2B ) that appeared to be from a type II pneumocyte origin based on CCSP negative and proSpC positive immunohistochemistry ( IHC ) ( Figure 2C ) . Twist1 cooperation with KrasG12D increased the number and size of lung tumors that developed . At six months of oncogene induction there was a large difference in the total lung tumors per mouse for CRT versus CR , 40 tumors versus 2 tumors ( p = 0 . 03 by t-test ) ( Figure 2D ) . Twist1 co-expression with KrasG12D in the lung also appeared to promote transformation of the predominantly benign lung adenoma tumor phenotype of CR mice [3] to a malignant phenotype composed mostly of adenocarcinomas as determined by a veterinarian pathologist [42] ( Figure 2E , p<0 . 0001 Fisher's exact test ) . A more sensitive marker of this conversion from adenoma to adenocarcinoma was the proliferative rate as we observed much higher proliferative index in CRT versus CR tumors ( Figure 2E , p = 0 . 021 Chi-square and Figure S2A ) . Although , we observed a strong genetic interaction between Twist1 and KrasG12D for lung tumorigenesis , we did not see a pronounced effect on distant lung tumor metastases . One CRT mouse did exhibit a macroscopic metastasis to the liver confirmed by pathology ( data not shown ) . However , in general the CRT cohort of mice ( n = 33 ) did not demonstrate increased distant metastasis compared to CR mice ( n = 55 ) when followed for up to 9 months of oncogene induction ( data not shown ) . Taken together , these data suggest that Twist1 does not appear to be a strong oncogene when over-expressed alone in the lung . Twist1 is capable of strong cooperation with KrasG12D for lung tumorigenesis and progression . Despite markedly accelerating tumorigenesis , Twist1 did not promote increased numbers of circulating tumors cells as detected by qPCR specific for the luc transgene ( data not shown ) and nor did Twist1 promote distant metastasis from primary lung tumors . CR lung tumors were fully reversible following 2–3 weeks of KrasG12D oncogene inactivation with the mechanism of tumor regression being a combination of tumor cells undergoing proliferative arrest and apoptosis [3] . We inactivated both Twist1 and KrasG12D from a cohort of CRT lung tumor moribund mice by the removal of doxycycline and monitored them for lung tumor regression at multiple time points using serial non-invasive imaging techniques in addition to final pathologic analysis ( n = 4 ) . CRT lung tumors showed dramatic tumor regression by gross examination ( compare Figure 3A versus Figure 2B ) that could be demonstrated serially with microCT ( Figure 3B ) and microPET-CT ( Figure 3C and Figure S3A ) after as little as 1 week of dual oncogene inactivation . By 4 weeks , CRT OFF lungs typically demonstrated no evidence of viable tumor cells on histologic analysis ( Figure 3D ) even despite CRT mice having considerably more advanced lung tumors than CR at similar time points ( Figure 2D ) . CRT mice with heavy initial tumor burden did have residual fibrotic scars remaining ( white spots in Figure 3A and trichome collagen staining in Figure S3B ) . To gain insight into the mechanism of tumor regression , we performed a time course analysis of CRT OFF lung tumors during the first week of oncogene inactivation . CRT OFF lung tumors demonstrated a prominent decrease in proliferation and increase in apoptosis following 5 days of doxycycline withdrawal as measured by Ki-67 and cleaved caspase 3 ( CC3 ) IHC , respectively ( Figure 3E–3F , p<0 . 0001 Chi-square for both Ki-67 and CC3 and Figure S3C–S3D ) . As mentioned previously , Twist1 has been shown in vitro to suppress KrasG12D oncogene-induced senescence ( OIS ) [23] . However , we did not see any appreciable increase in senescence associated beta-galactosidase ( SA-β-Gal ) staining following simultaneous inactivation of Twist1 and KrasG12D in CRT OFF lung tumors ( Figure 3G , p = 0 . 68 Chi-square ) or by assessing for markers of cell cycle arrest ( data not shown ) . These data suggest that although CRT lung tumors demonstrate more aggressive histologic appearance than CR tumors , CRT lung tumors are still strictly dependent on initiating oncogenes for tumor maintenance . Furthermore , Twist1 did not alter the mechanism of tumor regression between CR OFF and CRT OFF lung tumors . The strong dependency or addiction of KrasG12D-initiated lung tumors for KrasG12D [3] , [35] may have precluded us from observing any activation of OIS in CRT OFF lung tumors . Furthermore , given the genetic configuration of the CRT mouse model we were not able to examine the effects on lung tumors following inactivation of Twist1 alone . We addressed in vitro whether activation of rasG12V-induced senescence could be driven by inactivation of Twist1 by using mouse embryonic fibroblasts ( MEFs ) generated from β-actin-rtTA/Twist1-tetO7-luc ( BT ) mice . BT MEFs were shown to be inducible with doxycycline in vitro by Western blotting ( Figure S4A ) . As reported previously [23] , we found Twist1 was able to fully suppress rasG12V-induced senescence in vitro as shown by proliferation curves and SA-β-Gal staining ( Figure S4B–S4D ) . We removed doxycycline from the media of BT MEFs infected with rasG12V to downregulate expression of Twist1 at Day 12 . These de-induced BT MEFs activated OIS in vitro as shown by decreased proliferation and increased SA-β-Gal staining relative to cells maintained in the presence of doxcycline ( p = 0 . 0025 for proliferation and p = 0 . 0294 for SA-β-Gal; Figure S4E–S4F ) . These data suggested that at least in vitro inhibition of Twist1 can activate rasG12V-induced senescence . To examine whether Twist1 inhibition could be a viable therapeutic target in vivo for Kras mutant autochthonous lung cancers , we generated mice in which only Twist1 expression was doxycycline-dependent ( Figure 4A ) . The LSL-KrasG12D ( LSL ) model allows for conditional activation of an endogenous KrasG12D allele in the lungs following intranasal adenoviral delivery of Cre recombinase ( AdCMVCre ) [43] . The strain background difference between CT ( FVB/N ) and LSL ( C57BL/6 ) transgenic models forced us to use first generation progeny from these crosses for all our experiments . We generated tri-transgenic CT-LSL animals ( Figure 4A ) , activated Twist1 expression with doxycycline , then conditionally activated the KrasG12D allele with AdCMVCre and followed these CT-LSL ON mice and similarly treated littermate controls for lung tumor development . Twist1 accelerated conditional KrasG12D-induced lung tumorigenesis in CT-LSL mice ( CT-LSL versus LSL , p = 0 . 0121 by log-rank analysis , Figure 4B–4C , similar to CRT mice , although to a lesser degree . CT-LSL lung tumors were similar to CRT tumors based on histology , expression of type II pneumocyte markers , and increase in the proportion of lung tumors with a higher proliferative index ( Figure 4D–4F , p<0 . 0001 Chi-square ) . Recently , two groups have demonstrated in a similar LSL-p53 model system that p19ARF is a critical sensor of oncogenic stress from MAPK signaling in adenocarcinomas [44] , [45] . We similarly observed overlap of activated p19ARF ( nucleolar localization ) with areas of intense pErk1/2 staining by IHC in our CT-LSL ON tumor model ( Figure 4G ) . We next inactivated the expression of Twist1 alone in a cohort of CT-LSL lung tumor moribund mice by withdrawal of doxcycline ( CT-LSL OFF , n = 4 ) . Twist1 levels were confirmed in CT-LSL OFF tumors by qPCR ( Figure 5A , p = 0 . 004 by t-test ) and by serial BLI ( data not shown ) to return to levels in wildtype lungs . Interestingly , CT-LSL OFF lung tumors showed tumor stasis by serial microCT over the course of the 4 weeks of Twist1 inactivation in stark contrast to the progressive tumor growth seen for the control LSL OFF tumors ( Figure 5B; 18% versus 220% growth , p<0 . 0001 t-test ) . To further characterize in an unbiased manner the mechanism by which Twist1 suppression was inducing tumor stasis we performed microarray analysis . We compared CT-LSL OFF lung tumors with normal lung and microdissected lung tumors from CR , CRT , LSL , CT-LSL ON and CT-LSL OFF mice . The analysis of 2 , 163 annotated pathways using single sample GSEA ( ssGSEA ) , an algorithm designed for modest samples sizes [as used previously in [16]] , found gene sets representing p21 ectopic overexpression to be highly correlated with the CT-LSL OFF lung tumor transcriptional program ( Figure S5A–S5B ) . We used the complimentary Ingenuity Pathway Analysis ( IPA ) to identify canonical pathways from the differentially expressed genes between CT-LSL ON and CT-LSL OFF tumors . Consistent with ssGSEA we found Twist1 regulated key drivers of cellular senescence ( genes encoding p21 , p16 , p27 and IL-6 ) and EMT ( genes encoding cadherins , vimentin and alpha-catenin ) in the context of KrasG12D-driven lung tumors ( Figure S6 ) . Directed IHC analysis of CT-LSL OFF tumors confirmed ssGSEA and IPA that molecular changes consistent with activation of OIS , such as marked decreases in proliferation by Ki-67 and pronounced increases in staining for SA-β-Gal , p21 and p16 ( Figure 5C–5F and 5H–5K , p<0 . 0007 Chi-square for 5H–5K ) . The single inactivation of Twist1 in our CT-LSL OFF tumors also appeared to decrease the number of adenocarcinomas as shown with decrease in tumors with high proliferative rate to a frequency similar to LSL alone ( compare Figure 5C and 5H to LSL from Figure 4F , p = 0 . 93 Chi-square ) . In addition MAPK signaling intensity decreased in CT-LSL OFF significantly compaed to CT-LSL ON ( Figure 5G and 5L , p<0 . 015 Chi-square ) . The decrease of highly proliferative adenoncarcinomas with active MAPK signaling in KrasG12D-induced lung tumors was also seen following p53 restoration [44] , [45] . Lastly , apoptosis increased only very slightly in a subset of the CT-LSL OFF tumors as demonstrated by cleaved caspase 3 IHC ( data not shown ) . These data provide the first in vivo evidence that Kras mutant lung adenocarcinomas can be clinically impacted by activating a latent program of cellular senescence via the inhibition of Twist1 . The relevance of TWIST1 as a potential therapeutic target in human lung cancers was evaluated by examining public gene expression microarray datasets . We found seven independent human lung cancer gene expression datasets that in total consisted of 394 tumor samples and 159 normal lung samples [46] , [47] , [48] , [49] , [50] , [51] , [52] ( Figure 6A ) . Six out of the 7 datasets , as well as aggregate analysis of all 7 datasets demonstrated TWIST1 overexpression in lung cancers ( p = 0 . 04 for aggregate ) . The analysis included tumors of adenocarcinoma and squamous cell carcinoma histology which comprise the two most common subtypes encountered in human lung cancer . This microarray expression data was directly validated using quantitative PCR ( qPCR ) for TWIST1 on human lung cancer samples . In total we screened by qPCR 164 human lung tumor samples and confirmed that TWIST1 was indeed overexpressed ( 100/164 or 61% demonstrate at least 3-fold upregulation , 43/164 or 26% at least 10-fold overexpression and in some cases as high as 536 fold overexpression was observed , p<0 . 0001 by t-test; Figure 6B ) . TWIST1 was similarly overexpressed in all the histologies examined including adenocarcinoma and squamous cell carcinoma ( p<0 . 0001 by ANOVA ) ( Figure 6C ) . The range of relative TWIST1 overexpression observed by qPCR in our 164 primary human lung cancer samples ( range 3–536 fold TWIST1 overexpression ) was similar to the Twist1 overexpression observed in our mouse KrasG12D-Twist1-induced lung tumors ( range 5–960 fold Twist1 overexpression , n = 6 ) . Together these data demonstrate that TWIST1 is commonly overexpressed in human lung cancer and that our KrasG12D-Twist1 mouse models do reflect human lung cancer . The overexpression of TWIST1 in human lung cancers and our in vivo data from the CT-LSL OFF mouse lung tumors strongly suggested that TWIST1 may be a relevant therapeutic target in human lung cancer . The consequences of knocking down TWIST1 using shRNA technology was tested in human KRAS mutant H460 lung cancer cells . We screened various published shRNAs and found three sequences that were capable of knocking down human TWIST1 as shown by qPCR ( Figure 7A , p<0 . 029 by ANOVA ) and at the protein level by Western ( Figure 7B ) . TWIST1 knockdown in H460 cells resulted in marked inhibition of proliferation using all three shRNAs ( Figure 7C ) and increased staining for the cellular senescence marker SA-β-Gal ( Figure 7D–7E , p<0 . 023 by ANOVA ) . Other OIS markers p21 and p27 showed upregulation with a subset of the shRNAs examined ( Figure 7F ) . We extended these results in two other human non-small cell lung cancer cell lines , H727 and A549 , showing that TWIST1 knockdown resulted in decreased proliferation and increased expression of markers consistent with activation of senescence ( Figure S7 ) . We then confirmed that the TWIST1 shRNA was not having off target effects by performing rescue experiments with mouse Twist1 infected into H460 and A549 cells ( Figure S8A and data not shown ) . Notably , the three shRNAs used in our study were not predicted to knockdown mouse Twist1 cDNA , which was confirmed by qPCR ( Figure S8B and data not shown ) . The anti-proliferative effects of shRNA mediated knockdown of human TWIST1 in H460 and A549 cells was completely rescued by expression of mouse Twist1 ( Figure S8C and data not shown ) . These data provide evidence that inhibition of TWIST1 can activate latent OIS in multiple different human KRAS mutant lung cancer cell lines . To evaluate if the tumorigenic potential of human NSCLC cells required TWIST1 overexpression , we performed subcutaneous xenografting experiments with A549 cells in immune-compromised NOD-SCID mice . A549 cell infected with sh-Scrambled control shRNA and subcutaneously injected into NOD-SCID mice produced large tumors with high efficiency ( 5/6 mice developed tumors by ≤4 weeks ) necessitating humane euthanasia of the mice . In stark contrast , the identical experiment using sh-TWIST1 produced no tumors in any of the mice injected ( Figure 7G , p = 0 . 015 by Fisher's exact test ) . These xenografting results confirm that TWIST1 overexpression is required for tumorigenicity in vitro and in vivo in human NSCLC cells .
Our results dramatize that suppression of TWIST1 may be an effective pro-senescence therapy for human lung cancer . We provide the first in vivo demonstration that Twist1 plays an important role in both the acceleration and maintenance of KrasG12D-induced autochthonous lung tumorigenesis . Our results illustrate that TWIST1 may be an important target for the treatment of human lung adenocarcinoma . We generated two novel autochthonous transgenic mouse models to demonstrate that Twist1 overexpression cooperates with KrasG12D to markedly accelerate the onset of lung adenocarcinoma . Suppression of Twist1 expression to physiological levels is sufficient to induce lung tumor stasis that was associated with the activation of cellular senescence programs . Importantly , through the transcriptional analysis of over 500 human tumors , human TWIST1 was found to be frequently overexpressed and hence highly relevant to primary human lung cancers . Finally , the knockdown of TWIST1 in human KRAS mutant lung tumor cells was also associated with the loss of their neoplastic properties and the induction of cellular senescence . The generality of our results using different cell types and across species suggest TWIST1 is a potential therapeutic target in KRAS mutant lung cancers . Oncogene-induced senescence and oncogene-induced apoptosis represent early tumor suppressive barriers that must be overcome for premalignant cells to ultimately emerge as neoplastic . It had been reported previously that Twist1/2 could suppress mutant Kras-induced OIS in vitro [23] , [24] , but we report for the first time the ability of Twist1 to suppress OIS in vivo using a novel Twist1 lung model in combination with two complementary KrasG12D–induced autochthonous lung tumor models: the inducible transgenic KrasG12D ( CR ) model and the conditional endogenous KrasG12D ( LSL ) model . Our results are confirmed by an accompanying paper demonstrating that Twist1 can also accelerate KrasG12D-induced autochthonous breast tumorigenesis ( Morel et . al . ) . Twist1 co-expression accelerated tumorigenesis relative to KrasG12D alone in both lung tumor models . Twist1 acceleration was more pronounced in the CRT model than the CT-LSL model . One explanation for this difference is the greater strength of oncogenic signaling by transgenic KrasG12D versus endogenous KrasG12D [53] . An alterative explanation is that cell type specific chromatin regulation of tumor suppressor loci such as the Ink4a/Arf locus is a key determinant of whether mutant Kras elicits tumor suppressive responses resulting in apoptosis and/or senescence [54] . Another explanation are strain difference effects as we had to use a mixed background for the CT-LSL mouse experiments . These alternatives are not mutually exclusive and further study using additional tissue specific models of KrasG12D and Twist1 expression are needed to define the mechanistic basis for the differences we observed in oncogenic synergy observed between Twist1 and KrasG12D . The acceleration and progression of KrasG12D -induced lung tumors by Twist1 is reminiscent of that seen with p53 deficiency [3] , [44] , [45] , [55] . Notably , Twist1 may inhibit p53 through several independent mechanisms [23] , [56] , [57] , [58] , [59] , [60] , including direct Twist1-p53 antagonism [61] . One straightforward interpretation of our results is that Twist1 overexpression can phenocopy Trp53 deletion . Twist1 may also accelerate and promote KrasG12D-induced lung tumors by the direct transcriptional regulation of BMI1 [62] . As mentioned above , the control of tumor suppressor loci by chromatin regulatory complexes , such as those containing Bmi1 , may be a strong determinant of responses to oncogenic signaling [54] . Interestingly , ectopic expression of Twist1 in lung epithelial cells was associated with the induction of an EMT program . Whether the transdifferentiation program might contribute to accelerated tumor initiation , as proposed by Morel et . al . , is also an intriguing possibility . Additional studies are required to define the mechanisms by which Twist1 accelerates KrasG12D-induced lung tumors , as well as explain why different tissues exhibit differing cancer susceptibilities despite harboring the same initiating oncogenic event . Twist1 has been commonly implicated in metastasis [32] . Thus , our finding that Twist1 expression did not seem to confer increase distant metastases in either the CRT or CT-LSL autochthonous lung tumor models was surprising . We note that Twist1 appears to confer increased prometastatic ability in other models of tumorigenesis as predicted ( D . I . Bellovin , P . T . Tran and D . W . Felsher , unpublished data ) . Hence , Twist1 may have specific effects on metastatic potential . Our study dramatically illustrates that it is possible to activate a latent senescence program in Kras mutant tumors in vivo by targeting the collaborating oncogene , Twist1 . We uncover a newly defined synthetic interaction between mutant Kras and Twist1 resulting not in cell death , but cellular senescence . The activation of this program is evident at the molecular level and most importantly results in marked inhibition of Kras mutant lung tumor growth in vivo . We realize that a possible caveat to this approach is that we first overexpressed Twist1 prior to KrasG12D activation and lung tumor formation and thus may have biased tumors towards dependency for Twist1 . However , simply overexpressing an oncogene during tumorigenesis does not per se make tumors dependent or “addicted” to that oncogene as we have shown , in particular for lung tumorigenesis [35] , [63] . Finally , we validate that knocking down endogenous TWIST1 in human lung cancer cell lines in vitro and in vivo also results in activation of senescence . An alternative approach to inducible overexpression using the TET system as we used in our study would be to use genetic ablation of endogenous Twist1 using the Cre-LoxP or a inducible shRNA system following development of KrasG12D–induced lung tumors . As KrasG12D–induced lung tumors are primarily adenomas with low proliferative rates ( Figure 2E and Figure 4F ) , endogenous Twist1 ablation or knockdown would not likely have an effect as has been shown for p53 restoration in adenomas [44] , [45] . From a clinical standpoint complete ablation of a gene , such as in mice using the Cre-LoxP system , is therapeutically not possible in humans . In contrast , the TET model system where we can suppress Twist1 overexpression to physiologic levels is more clinically relevant to what is done in the clinic with inhibitors . Others have shown senescence can arise in vivo in established tumors by targeting an initiating oncogene or reconstitution of a tumor suppressor [21] , [64] , [65] , [66] . Our work further highlights the activation of a latent cellular senescence program or pro-senescence therapy as an innovative avenue for cancer therapy [67] . Our results may extend beyond KRAS-mutant lung cancers . Notably , TWIST1 was found to be overexpressed in a majority of human lung cancer samples we tested . This includes not only adenocarcinoma , in which KRAS mutation is commonly observed , but also other major lung cancer histologies including squamous cell carcinoma , in which KRAS mutation is rare . Our preliminary data suggests that TWIST1 knockdown can result in activation of OIS in KRAS wildtype lung cancer cell lines in vitro , but further characterization of these lines for mutations in other components of the EGFR/KRAS/BRAF pathway are needed ( T . F . Burns , P . T . Tran and C . M . Rudin , unpublished data ) . Furthermore , additional preliminary findings suggest that TWIST1 may have a larger role in suppressing OIS following activation of other key driver mutations using other transgenic mouse lines ( P . T . Tran and D . W . Felsher , unpublished data ) . This hypothesis will be further explored in lung cancer through introduction of our inducible Twist1 construct into other relevant transgenic models of lung tumorigenesis . Importantly , regardless of whether there is an exclusive association between KRAS mutation and TWIST1 overexpression in human lung cancer cells , the data presented strongly support that TWIST1 upregulation in KRAS mutant lung cancer represents a novel and particularly promising therapeutic target . These observations have important and immediate translational implications for this particularly refractory subset of lung cancers . The consequences of systemic transient inhibition of Twist1 in the adult has not been well defined and thus side-effects of such treatment are unknown . Germline deletion of Twist11 in mice is embryonic lethal [22] and loss of function mutations in humans cause a severe developmental disorder . However , postnatal expression of TWIST1 appears to be tightly restricted to a subpopulation of mesoderm derived tissues and limited studies suggest that Twist1 inhibition systemically may be well tolerated [68] . We conclude that TWIST1 may be an effective target for “pro-senescence” therapy for human lung cancers [67] . Our results suggest that it will only be necessary to suppress TWIST1 to a physiological level which may preclude toxicity . Our mouse model will be useful to identify agents that target TWIST1 for the treatment of human cancer .
The human non-small cell lung cancer cell lines , H460 , H727 and A549; and embryonic kidney cell line HEK 293 T were obtained from ATCC and grown in media as recommended . MEFs were isolated from E13 . 5 embryos and propagated as described previously [18] . MEFs were grown for two population doublings and then frozen for future experiments . MEFs were grown in DMEM plus 10% fetal calf serum . The Twist1 cDNA was PCR cloned into the bidirectional tetO7 vector S2f-IMCg [33] at EcoRI and NotI sites , replacing the eGFP ORF . The resultant construct , Twist1-tetO7-luc , was sequence confirmed , digested with KpnI and XmnI to release the bidirectional transgene and then used for injection of FVB/N pronuclei by the Stanford Transgenic Facility . We ultimately obtained three founders from 25 pups after screening by tail genotyping using PCR as described below . These three founders were mated to CCSP-rtTA mice to screen for functional Twist1-tetO7-luc founders . One founder failed to pass the transgene germline and one founder did not report inducible Twist1 or luc expression . The remaining founder was used for all the experiments in this study . We use the β-actin-rtTA , CCSP-rtTA , tetO-Kras4bG12D and LSL-K rasG12D transgenic lines [3] , [34] , [69] . Twist1 and/or K-rasG12D expression was activated in the CT , CR , and CRT lung lines by administering doxycycline ( Sigma ) to the drinking water weekly [2 mg/mL] starting at the age of 3–5 weeks . The conditional LSL-K rasG12D lines were activated by intranasal delivery of adenoviral CMV-Cre [43] . All procedures were performed in accordance with APLAC protocols and animals were housed in a pathogen-free environment . DNA was isolated from mouse tails using the Qiaprep DNeasy kit ( Qiagen ) . The CCSP-rtTA , tetO-K-rasG12D and LSL-K rasG12D transgenic lines were screened as described previously . The Twist1-tetO7-luc line was detected with the following primers: mTwist1-Luc . S2 5′- CCTTATGCAGTTGCTCTCCAG -3′ and mTwist1-Luc . AS2 5′- GCTTGCCTATGTTCTTTTGGA -3′ . DNA was amplified using PCR and PCR products were resolved on a 2% agarose gel . Total RNA was isolated from tissue using the Qiaprep RNAeasy Kit ( Qiagen ) according to the manufacturer's directions . Samples were treated with RQ1 RNase-Free DNase ( Promega ) . cDNA was generated from 1 µg of total RNA using the Superscript II kit ( Invitrogen Technologies ) . Control reactions were run without RT enzyme . 50 ng of cDNA equivalents were amplified for the transcript described below in an ABI-prism 7700 for 40 cycles using SYBR green PCR Master mix ( Perkin Elmer Applied Biosystems ) . PCR reactions were performed in duplicate-triplicate in a final volume of 20 µL . Following amplification , the data was processed with the analysis program Sequence Detection Systems v2 . 2 . 2 ( Perkin Elmer Applied Biosystems ) . For each sample , the level of RNA for the genes of interest was standardized to a housekeeping gene ( ubiquitin or 18S rRNA ) within that sample; subsequently , the level of a transcript of interest was normalized to the expression of that transcript from the appropriate comparator sample . Primers for qPCR are listed in the Text S1 . Human normal lung and lung tumor qPCR tissue arrays and TWIST1 qPCR oligos were purchased from OriGene . All relevant clinical information can be found ( http://www . origene . com/qPCR/Tissue-qPCR-Arrays . aspx ) . Cells were lysed on ice for 60 min in radioimmunoprecipitation assay buffer supplemented with protease and phosphatase inhibitors ( Sigma-Aldrich ) and clarified by centrifugation . Protein concentrations were determined by Bradford proteinassay ( Bio-Rad Laboratories ) . Equal protein concentrations of each sample were run on NuPAGE bis-Tris gels ( Invitrogen ) and transferred to membranes . After being blocked with 5% dried milk in TBS containing 0 . 2% Tween 20 , the filters were incubated with primary antibodies . The following primary antibodies were used: goat anti-Actin ( C-11 , Santa Cruz ) , mouse anti-Twist1 ( TWIST2C1a , Santa Cruz ) , mouse anti-p21 ( Ab-1 , Calbiochem ) ) , mouse monoclonal anti-p27 ( F-8 , Santa Cruz ) After washing and incubation with horseradish peroxidase ( HRP ) -conjugated anti-Goat or anti-mouse IgG ( Amersham ) , the antigen-antibody complexes were visualized by chemiluminescence ( ECL detection system; Perkin Elmer ) . Tissues were fixed in 10% buffered formalin for 24 h and then transferred to 70% ethanol until embedded in paraffin . Tissue sections 5 µm thick were cut from paraffin embedded blocks , placed on glass slides and hematoxylin and eosin ( H&E ) or Masson's trichrome staining was performed using standard procedures . Antibodies used in our study: p21 , p27 , p16 , vimentin ( BD Pharmingen ) and E-Cadherin ( Cell Signaling ) . We performed IHC , measured K-i67 and CC3-staining as described previously [35] . For immunofluorescence ( IF ) , Alexa488-conjugated anti-mouse and Alexa594-conjugated anti-rabbit ( 1∶300 dilution , Invitrogen ) were used as secondary antibodies and incubated at room temperature for 30 minutes . DAPI was used as a nuclear stain and slides were mounted in aqueous mounting media ( Vector Laboratories ) . For EMT IF analysis double immunofluorescence was used . Vimetin-expressing cells were labelled with Alexa488 ( green ) and E-cadherin-expressing cells were labeled with Alexa 594 ( red ) . To quantify cells undergoing EMT , cells that were red ( low ) green ( high ) were manually counted . A minimum of seven different fields of view per section from greater than four different animals were analyzed in total . 293T cells were seeded ( 2 . 5×106 cells ) in T25 flasks . shRNA constructs were obtained from the Broad RNAi Consortium . pLKO . 1-shRNA scramble vector was used . Lentivirus was made using a three-plasmid system and infected using the TRC Library Production and Performance Protocols . Twenty-four hours after infection , cells were treated with 1 mg/ml puromycin and passaged once 80% confluent . Retroviral production used ecotropic and amphotropic Phoenix packaging lines . Early passage MEFs were transduced with pWZL-Hygro vectors expressing HrasG12V or with empty vector for two successive times over a 36-h period and then followed by selection with hygromycin ( 100 µg/ml ) for 4 days . Retroviral infections on H460 cells used pWZL-Hygro vector and pWZL-Hygro/mTwist1 constructs , for two successive times over a 36-h period and then followed by selection with hygromycin ( 250 µg/ml ) for 4 days . On Day 6 after infection with the indicated shRNA lentiviruses , cells were plated in 12-well plates at a density of 5E3 , 10E3 and 15E3 cells/well . On Day 12 , the cells were stained with crystal violet ( 0 . 5% in 95% ethanol ) . Similar low passage MEFs were used for all proliferation assays . Retroviral infections were performed as above , selection carried out for 4 days and stably selected cells were plated and then treated with or without 2 µg/ml doxyxycline for proliferation assays ( Day 1 ) . Sets of cells were removed for trypsinization and counting every 4 days . Values are normalized with Day1 readings . Cells were washed twice with phosphate-buffered saline ( PBS ) and then fixed with PBS containing 2% formaldehyde and 0 . 2% glutaraldehyde for 5 min . The cells were then incubated at 37°C for 20 hr with staining solution ( 40 mM citric acid sodium phosphate , pH 6 . 0 , 1 mg/ml 5-bromo-4-chloro-3-isolyl-β-D-galactoside [X-gal , Fisher] , 5 mM potassium ferricyanide , 5 mM potassium ferrocyanide , 150 mM NaCl , 2 mM MgCl2 ) . After incubation , cells were washed twice with PBS and viewed with bright-field microscopy . Micro-computed tomography ( μCT ) and PET scans were performed on a custom GEHC ( London , Ontario ) eXplore RS150 cone-beam scanner and an R4 microPET ( Siemens Medical Solutions USA , Inc . ) , respectively , as described previously [35] , [70] . Mice were screened serially every 1–2 weeks following doxycycline activation or intranasal adenoviral CMV-Cre and images were reviewed by a board certified radiation oncologist ( PTT ) . PET images were reconstructed using the ordered-subsets expectation maximization algorithm with a spatial resolution of 1 . 66 to 1 . 85 mm . No attenuation correction or partial volume corrections were applied . Micro-computed tomography ( μCT ) images were reviewed by a board certified radiation oncologist ( PTT ) on multiple index tumors in a blinded fashion ( n = 2–5 tumors per mouse ) . Bi-dimensional measurements were made on tumors using serial examinations and tumor volumes calculated using the following equation vol = pi/6×1 . 65 ( length×width ) ×3/2 . Volumes were normalized to the starting volume , t = 0 before doxycycline treatment , and percent tumor volume growth was then calculated by ( normalized tumor vol . ×100% ) −100% . Female NOD-SCID mice 4–5 weeks old were purchased from Harlan Laboratories . Mice were maintained under pathogen-free conditions and given food and water ad libitum in accordance with guidelines from the Johns Hopkins Animal Care and Use Committee . A549 infected with sh-Scrambled control or sh-TWIST1 shRNA , selected for 4 days as described above and then 5×105 million cells in 100 µL of Hank's solution and Matrigel ( BD Biosciences ) mixed 1∶1 were injected subcutaneously in the right flank . Tumor measurements were taken every 2–3 days .
|
Lung cancer is the most common cause of cancer death worldwide . The Twist1 gene encodes for an essential transcription factor required for embryogenesis and overexpressed in many cancer types . It has yet to be shown in vivo whether Twist1 plays a role in the initiation or maintenance of cancer . Here we demonstrate using novel transgenic mouse models that Twist1 cooperates to induce lung tumorigenesis by suppressing cellular senescence programs . Moreover , the suppression of Twist1 in murine tumors elicited cellular senescence and the loss of a neoplastic phenotype . We found that TWIST1 is commonly overexpressed in human lung cancers . Finally , the inhibition of TWIST1 levels in human lung cancer cells was associated with loss of proliferation , induction of cellular senescence , and the inability to form tumors in mice . Hence , we conclude that TWIST1 is a key regulator of cellular senescence programs during tumorigenesis . The targeted inactivation of TWIST1 may be an effective pro-senescence therapy for human lung adenocarcinomas .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"animal",
"models",
"medicine",
"oncology",
"cancer",
"genetics",
"model",
"organisms",
"lung",
"and",
"intrathoracic",
"tumors",
"genetics",
"cancer",
"treatment",
"biology",
"cancers",
"and",
"neoplasms",
"mouse",
"genetics",
"and",
"genomics"
] |
2012
|
Twist1 Suppresses Senescence Programs and Thereby Accelerates and Maintains Mutant Kras-Induced Lung Tumorigenesis
|
Human metabolism involves thousands of reactions and metabolites . To interpret this complexity , computational modeling becomes an essential experimental tool . One of the most popular techniques to study human metabolism as a whole is genome scale modeling . A key challenge to applying genome scale modeling is identifying critical metabolic reactions across diverse human tissues . Here we introduce a novel algorithm called Cost Optimization Reaction Dependency Assessment ( CORDA ) to build genome scale models in a tissue-specific manner . CORDA performs more efficiently computationally , shows better agreement to experimental data , and displays better model functionality and capacity when compared to previous algorithms . CORDA also returns reaction associations that can greatly assist in any manual curation to be performed following the automated reconstruction process . Using CORDA , we developed a library of 76 healthy and 20 cancer tissue-specific reconstructions . These reconstructions identified which metabolic pathways are shared across diverse human tissues . Moreover , we identified changes in reactions and pathways that are differentially included and present different capacity profiles in cancer compared to healthy tissues , including up-regulation of folate metabolism , the down-regulation of thiamine metabolism , and tight regulation of oxidative phosphorylation .
Genome-wide Metabolic Reconstructions ( GEMs ) computationally model the molecules and reactions responsible for metabolism in any given organism , and have been applied across a variety of fields including metabolic engineering and evolutionary analysis [1] . Computational methods developed to study GEMs [2] have generated novel hypotheses about the structure of metabolic networks in microorganisms , and helped elucidate gaps in our knowledge of metabolism [3 , 4] . Since the publication of the comprehensive human metabolic reconstruction Recon1 [5] , human GEMs have enabled the study of human metabolism at a genome level [6] . These studies include the prediction of novel metabolic functions [7] , prediction of metabolic biomarkers for congenital genetic disorders [8 , 9] , context analysis of omics data [10–12] , comparison between humans and other mammals through gene homolog mapping [13 , 14] , and prediction of suitable cancer drugs [15 , 16] and drug targets [17–19] . A particularly prolific subfield of human GEMs is the development of tissue-specific reconstructions . Different groups of metabolic reactions occur in different cell types . Hence , numerous studies have been dedicated to generating tissue specific or cell specific models of metabolism [20 , 21] . These tissue-specific reconstructions can be built by piecing together the model based on previously established biological evidence obtained by reviewing the literature [22–26] , through the integration of omics data and computational methods in order to tailor generic , published human reconstructions [5 , 9 , 27–29] to the desired cell type [15 , 16 , 30–33] , or through a combination of computational algorithms and manual curation [27 , 28 , 34–36] . Automated tissue-specific reconstruction algorithms developed to date can be broadly categorized into two groups [20]: “flux-dependent” and “pruning” methods . Flux dependent methods find an optimal flux distribution through the general reconstruction which contains the maximum number of high confidence reactions ( i . e . reactions whose presence is supported by significant experimental data ) [15 , 31 , 32 , 37–39] . These algorithms have been successfully used to predict gene essentiality in cancer tissues [19 , 33] , cancer specific metabolic pathways [31] , metabolic biomarkers for congenital genetic disorders [8 , 9] , and cancer specific anti-growth factors [15 , 16] . One of the main advantages of flux-dependent methods is the fact that they predict a flux distribution along with the tissue-specific model [20] . While this characteristic can be desirable , it also renders flux-dependent reconstructions “snapshots” of the metabolic state defined by the data , as opposed to comprehensive , functional metabolic models [15 , 20] . The second category of tissue-specific reconstruction methods are pruning algorithms , which include MBA [34] , mCADRE [30] and fastCORE [40] . Models generated using these algorithms have been used to calculate metabolic flux values in hepatocytes [34] , identify pathways specific to cancer [30] , and predict cancer drug targets [17 , 18] . These algorithms start with a core set of reactions , obtained through literature review or experimental data , and proceed by removing the remaining reactions in the generalized human reconstruction while maintaining functionality in the core set . In these algorithms , a tradeoff can be defined between maintaining the model as concise as possible and including all core reactions . That is , if a core reaction requires too many undesirable reactions to carry flux , the algorithm may remove this core reaction from the tissue model , a tradeoff referred to as flexible core . There are two main advantages to defining a core set of reactions before performing the tissue-specific algorithm . The first advantage is the possible inclusion of multiple sources of data and biochemical information [20 , 34] . The definition of the reactions core is left to the user’s discretion , allowing for both the combination of data sources and the manual inclusion of reactions . Secondly , reactions with overwhelming evidence are always included in the final tissue model , since a non-flexible set of high confidence reactions can be defined [20] . This pruning approach then allows for the construction of comprehensive tissue models , containing all reactions that may be in a tissue’s metabolism , as opposed to a snapshot of the metabolic state returned by the flux-dependent methods [15 , 20] . Current pruning methods are also accompanied by two major limitations , however . First , the order in which reactions are removed from the model plays a major role in the final reconstruction . Second , similar to flux-dependent methods , current algorithms aim to keep the final tissue-reconstruction as concise as possible , an approach referred to as parsimonious . These algorithms aim to remove from the tissue-specific model all reactions for which experimental data is unsupportive or unavailable , such as reactions with low levels of gene expression or non-gene associated reactions . While a concise tissue-specific reconstruction is desirable , keeping the reconstruction as parsimonious as possible may lead to the removal of fundamental reactions and physiologically unlikely flux distributions . In Recon 1 , for instance , oxygen and H2O exchange reactions can be removed from the reconstruction with no effect on model functionality ( Fig 1A ) . During simulations , however , these would be replaced by the uptake of the toxic metabolites superoxide anion and hydrogen peroxide respectively , leading to the prediction of physiologically inaccurate flux distributions ( Fig 1A ) . The oxygen exchange reaction is in fact not present in the MBA and mCADRE liver reconstructions , and the water exchange reaction is not present in the mCADRE liver reconstruction . Hence , in order to ensure our algorithm did not rely on alternative , physiologically unlikely pathways , and that it was independent of any ordering assignments , we chose to take an approach which was not parsimonious . Here we introduce a novel tissue-specific reconstruction algorithm based on Cost Optimization Reaction Dependency Assessment ( CORDA ) . CORDA returns a concise , functional tissue-specific reconstruction , and features a flexible reactions core . CORDA does not depend on Flux Variability Analysis [41] or Mixed Integer Linear Programming ( MILP ) problems , but only on Flux Balance Analysis [42] ( FBA ) , which is dependent on Linear Programming ( LP ) . This characteristic renders CORDA considerably faster than previous , similar methods . Finally , the CORDA algorithm returns reaction associations that assist in any manual curation to be performed following the automated reconstruction process . In line with previous studies [43] , we apply CORDA to generate a library of 76 healthy and 20 cancer-specific metabolic reconstructions . These reconstructions enabled us to identify metabolic similarities amongst healthy tissues as well as key differences between healthy and cancerous tissues . Furthermore , by sampling the feasible solution space in cancer and healthy models , this library can be used to predict the up- and down-regulation of cancer-specific pathways in cancer metabolism .
The CORDA algorithm is based on a novel approach to identify the dependency of desirable reactions ( i . e . reactions with high experimental evidence ) on undesirable reactions ( i . e . reactions with no experimental evidence ) , a method referred to here as dependency assessment . In the dependency assessment approach , the metabolic network is modified in four ways ( Fig 1B ) . First , reversible reactions are split into forward and backward components . Second , a pseudo-metabolite is added as a product for every reaction in the model . At this point , undesirable reactions will carry a higher stoichiometric coefficient for this added metabolite , assigning these reactions a higher “cost” . Third , a reaction consuming this pseudo-metabolite is added to the model . Finally , a positive lower bound is set for the reaction being tested in order to force that reaction to carry flux . After modifying the network , FBA ( Materials and Methods ) is performed while minimizing the flux through the cost-consuming reaction ( Fig 1B ) . The flux distribution returned will then use high cost , undesirable reactions only as necessary for the reaction being tested to carry flux . Throughout the manuscript , we will refer to high cost reactions predicted to carry flux as associated with the reaction being tested . In order to identify pathways with the same cost ( i . e . same number of undesirable reactions ) , multiple dependency assessment can be performed while adding a small amount of noise to the cost of each reaction . Using this dependency assessment , we have developed the CORDA algorithm for the reconstruction of tissue-specific models ( Fig 1C ) . CORDA takes as input the reactions in the generalized human reconstruction separated into high ( HC ) , medium ( MC ) , and negative ( NC ) confidence groups ( see Materials and Methods section for a detailed description ) . All remaining reactions in the reconstruction ( i . e . non gene associated reactions or reactions for which no data is available ) are designated as others ( OT ) . All HC reactions are included in the model , and the maximum number of MC reactions is included while minimizing the inclusion of NC reactions . While the definition of these four reaction groups are left to the user’s discretion , here we categorize them according to proteomics data from the Human Protein Atlas ( HPA ) [44 , 45] and a methodology used in previous studies [30 , 32 , 37] ( Materials and Methods ) . To begin the algorithm , all HC reactions are moved into the tissue reconstruction ( RE ) . In a first step , MC and NC reactions associated with each RE reaction ( which are the same as the HC group at this point ) are identified using the dependency assessment and moved into the RE group . In a second step , NC reactions associated with a high number of MC reactions are identified and moved into the tissue model , and all remaining NC reactions are blocked ( upper and lower bounds set to zero ) . Next , all MC reactions still able to carry flux are also moved to the RE group . Finally , in the final step of the algorithm , all OT reactions associated with any RE reaction are moved to the RE group for the final tissue-specific model . A detailed description of the CORDA method , including detailed steps , algorithm parameters , and categorization of model reactions is available in the Materials and Methods section . Following the algorithm validation , we generated a library of 76 healthy and 20 cancer tissue-specific models using CORDA . In order to generate the most comprehensive models possible , we used the generalized human reconstruction Recon2 [9] in the calculation of this library . Recon2 is one of the most comprehensive human reconstructions performed to date , containing approximately twice the amount of reactions than Recon1 , 1 . 7 times more unique metabolites , and 1 . 2 times more unique genes . Details of how the reconstructions were calculated can be found in the Materials and Methods section .
Here we introduced a novel tissue-specific algorithm based on Cost Optimization Reaction Dependency Assessment ( CORDA ) . CORDA relies solely on FBA , rendering it more computationally efficient than previous methods . CORDA takes a non-parsimonious approach to the reconstruction process , based on the addition of valuable reactions to the reconstruction as opposed to the removal of non-essential reactions . We showed that the CORDA algorithm provides reconstructions that agree better with experimental data , and that demonstrate better metabolic functionality than prior methods like MBA and mCADRE . Furthermore , CORDA provides reaction associations that can greatly assist subsequent manual curation , while maintaining the reconstructions only slightly larger than previous parsimonious approaches . Monte-Carlo sampling analysis also demonstrates that the CORDA generated models provide better predictions of tissue-specific functionality . In addition to the algorithm validation , we generated a library of 76 healthy and 20 cancer tissue-specific reconstructions , which show considerable agreement with our current knowledge of healthy tissue and cancer metabolism . First , as an initial validation of our cancer and healthy tissue models , we computationally predicted metabolites that are more frequently essential in cancer models than healthy tissues [15 , 16 , 54] . Two metabolites were implicated in this analysis: phosphatidylethanolamine ( pe_hs ) and triglyceride ( tag_hs ) , both of which are part of metabolic pathways previously implicated as cancer specific [15 , 16] . While future work is merited to identify more specific essential metabolites ( e . g . through the inclusion of more comprehensive metabolic tasks in the tissue reconstruction process , and more metabolites in the essential metabolite identification algorithm ) , these results help validate the cancer and healthy tissue reconstructions presented here . Following this analysis , we demonstrated that the tissue models calculated by CORDA cluster largely according to tissue type . Similar clustering patterns , based on gene expression and proteomics data , have been observed experimentally . In particular , based on the expression of over 30 , 000 genes across multiple individuals and tissues , one study found that brain , muscle , and liver tissues , as well as Epstein-Barr virus-transformed lymphocytes , form well defined groups , while skin , adipocytes , and nerve tissues cluster closely together [117] . A separate study used in the generation of the HPA , based on protein evidence from almost 17 , 000 protein-coding genes in 44 major tissues and organs , also showed that tonsils , spleen , appendix , and lymph node tissues cluster closely together , and that bone marrow clusters separately , but close to these lymphoid tissues [45] . Evidence supporting many of the apparent exceptions identified by our clustering analysis is also available . For instance , Uhlén et . al . found that brain and liver tissues , along with testis , cluster considerably separate from other tissues and closer to each other , which is what we observed by clustering the CORDA models . The same study found that prostate tissue clusters closely with salivary glands [45] . It is worth noting that good agreement with the data by Uhlén et . al . is expected , given that a subset of this data was used to generate the tissue-specific models . This agreement , however , suggests that the similarities between tissues shown by Uhlén et . al . [45] and Melé et . al . [117] at the gene expression and protein level are also present in the metabolic enzymes level . Additionally , breast and salivary glands are known to share many morphological features , and studies have shown that both can give rise to tumors with similar morphology [118 , 119] and myoepithelial differentiation [120] . These finding can explain why breast and salivary glands clustered with epithelial and myoepithelial cells , as opposed to glandular cells . Finally , skin cancer and non-Hodgkin’s lymphoma appear frequently as secondary cancers in immunosuppressed individuals [121 , 122] . This could lead to cancers with significantly different metabolic profiles , supporting their separation from the remaining cancer models . Clustering of tissue-specific models according to subsystems has also highlighted many differences between healthy and cancerous tissues at the pathway level ( Fig 5 ) . Evidence for many of these differences are also available in the literature , including: Single reactions included most often in cancer or healthy tissue models were also analyzed , and again literature evidence has been found to support many of them ( Table 3 ) . Two surprising findings stemmed from this analysis . First is the predicted down-regulation of CoA synthesis reactions , implicated in both the subsystem and single reaction analyses . Upon further inspection , we traced this differential inclusion to the gene PPCS , the only gene related to this pathway included in the reconstruction process , which is significantly down-regulated in cancer cells [44 , 45] . Second , the exclusion of ACOAO7p from most cancer models is also unexpected , since this reaction is part of the fatty-acid oxidation pathway , which has been shown to be up-regulated in cancer tissues [123 , 124] . Protein evidence of this reaction’s associated gene , ACOX1 , supports this exclusion from cancer models [44 , 45] , suggesting an alternate pathway for palmitoyl-CoA oxidation in cancer tissues . Finally , Monte-Carlo sampling was also performed in all healthy and cancer tissue models . Sampling results demonstrate that cancer models show an increased capacity through pathways that are largely up-regulated in cancer metabolism , and a reduced capacity through pathways previously shown to be down-regulated . Interestingly , mitochondrial respiration showed a slightly reduced and tightly constrained capacity in cancer over healthy tissue models , despite the presence of a larger number of oxidative phosphorylation reactions in cancer models ( Fig 5 ) . For decades , the role of mitochondrial respiration was thought to be decreased in cancer tissues due to their high glycolytic capacity . In recent years , however , researchers have shown that this pathway actually plays an important role in cancer metabolism [125 , 126] . Our results suggest that although a larger number of oxidative phosphorylation reactions are present in cancer models , the activity of this pathway is tightly regulated by cancer metabolism topology ( Fig 6 ) . On one hand , the low probability of cancer models reaching high cytochrome c oxidase flux values compared to healthy tissues is in line with cancer’s high glycolytic potential . At the other extreme , the low probability of cancer models reaching relatively low cytochrome c oxidase sampled fluxes is in line with the key role played by mitochondrial respiration in cancer metabolism uncovered in recent years . We have also investigated the differences in glycine hydroxymethyltransferase capacity in cancer versus healthy tissue models ( S1 Text ) . This reaction is dependent on two proteins , SHMT1 and SHMT2 , which correspond the cytosolic and mitochondrial isozymes respectively . Both these proteins have been shown to be up-regulated in cancer over healthy tissue models [127] , although SHMT2 has been so to a greater extent [71 , 127] . The over expression of these proteins , however , has been shown to be heavily dependent on cancer type [127] . This claim is supported by the protein expression of SHMT2 in the HPA , where half the cancer types considered have samples with both high and not detected SHMT2 expression . This variability could explain why the distribution of reactions associated with these genes is similar between cancer and healthy tissue models ( S1 Text ) . Some cancer types , however , show a considerable increase in SHMT2 expression when compared to their healthy counterparts , including breast , glioma , head and neck , lung , stomach , testicular , and thyroid cancer . In all but one of these models ( glioma ) , the flux distribution of glycine hydroxymethyltransferase was shown to be considerably shifted towards higher values when compared to their healthy counterparts ( S1 Text ) . These results demonstrate CORDA’s ability to predict cancer type specific functionality , and not only differences between all cancer and healthy tissues taken together . The CORDA tissue-specific reconstruction algorithm , as well as the healthy and cancer tissue-specific reconstructions presented here , introduce a new approach for the development of comprehensive tissue-specific metabolic reconstructions . These reconstructions can generate novel insights into both healthy and diseased human metabolic behavior . Furthermore , the ability of CORDA to generate models based solely on experimental data , along with the computational efficiency of this algorithm , allows for continuous updates of this library of tissue-specific models , both as more experimental data is updated and made available , and as more comprehensive human metabolic reconstructions are developed .
While previous methods determined reaction dependencies using Flux Variability Analysis ( FVA ) , the CORDA algorithm takes a different approach , referred here as dependency assessment . The novelty of this method lies not in the LP formulation itself , which is the same as the widely established Flux Balance Analysis ( FBA ) , but in the model modifications performed prior to the application of FBA , as well as the interpretation of the flux distribution returned . Assuming we want to test whether a given reaction , x , is dependent on the presence of a group of reactions , Y , to carry flux , CORDA proceeds in five steps . The parameters required for the CORDA algorithm are summarized in Table 4: It is worth noting that the high cost reactions implicated in step five are not necessarily essential for x to carry a flux ±ϵ , but are the set of reactions in Y that combined carry the minimal amount of flux . That is , no flux distribution through the metabolic network allows for the predefined flux through x with a lower combined flux through the reactions of Y . For instance , if one of the reactions in Y deemed associated with x were to be removed from the reconstruction , x could still be able to carry a flux ±ϵ , but the combined flux through the reactions in Y would be larger than before . This way , this dependency assessment does not minimize the number of undesirable reactions to allow x to carry flux , but instead the combined flux through them . Naturally , however , a lower number of reactions would more easily allow for a lower combined flux . It is also for this reason that throughout the manuscript we use the term associate instead of dependent . Throughout the literature , referring to one reaction as dependent on another means the removal of the later from the model negates the former’s ability to carry flux , which is not necessarily the case for the reaction associations defined here . Another significant advantage of this dependency assessment over previous pruning algorithms is that it requires only the LP problem solved during FBA , rendering it much faster than previous methods . While MBA and mCADRE used a much faster variation of FVA , it is still considerably more computationally expensive than LP . Although mCADRE is up to three orders of magnitude faster than MBA [30] , the mCADRE model used in this study took about 4 hours to be calculated in a 2 . 34 GHz CPU with 4G RAM using the IBM CPLEX solver [30] . The CORDA reconstruction , on the other hand , using the same data and general human reconstruction , took under 30 minutes in a 2 . 66 GHz CPU with 4G RAM using the Gurobi solver [128] . In order to obtain a tissue-specific metabolic reconstruction using this dependency assessment , we define the Cost Optimization Reaction Dependency Assessment ( CORDA ) algorithm . This algorithm takes as input the reactions in the generalized human reconstruction divided into four categories: Here , we also allow for the inclusion of metabolic tasks in the HC group . That is , during the CORDA algorithm , sinks can be specified for given metabolites , and added to the model when tested to ensure the final tissue model can produce these metabolites . These reactions are added when being tested then immediately removed from the model , so that none of these metabolic task reactions are present when other reactions are being tested , and no two test reactions are present in the model at the same time . The 32 metabolic tasks included in all CORDA reconstructions in this manuscript are available in S1 Table . While the definition of these reaction groups can be left to the user’s discretion , here we defined the four groups according to proteomics data from the HPA [44 , 45] , and boolean gene-reaction rules included in the generalized reconstructions Recon1 and Recon2 . In the HPA , each protein is classified as being Not Detected , or present at Low , Medium or High levels in each tissue . The gene-reaction association rules are composed of gene names and “AND” and “OR” boolean associations . For instance , the reaction r0634 in Recon2 has the boolean rule “HADHB AND ( ACAA2 OR ACAA1 ) ” , and can therefore be considered active if the gene HADHB , as well as ACAA2 or ACAA1 , are active . Using this boolean mapping , gene IDs were first replaced by the numerical values -1 , 1 , 2 , and 3 , corresponding to Not detected , Low , Medium and High protein expression levels respectively . Genes not included in the dataset were assigned a numerical value of zero . Next , AND boolean associations were replaced by the function MIN; OR boolean associations were replaced by the function MAX; and the expression was evaluated . Reactions with a final score of 3 were assigned to the HC group; reactions with scores of 1 or 2 were assigned to the MC group; and reactions with a score of -1 were assigned to the NC group . Reaction scores of -1 , 1 , 2 , and 3 also correspond to Not Detected , Low , Medium , and High expression levels expressed in Fig 2 . As an example , HADHB is expressed at low levels in cerebellum Purkinje cells; ACAA2 is not detected; and ACAA1 is expressed at high levels . The r0634 gene-reaction rule mentioned above was then be replaced by “MIN ( 1 , MAX ( -1 , 3 ) ) ” , which evaluates to 1 . During the Purkinje cells reconstruction , this reaction was then placed in the MC group . Similar approaches have been used by previous studies to assign reaction confidence scores [30 , 32 , 37] . Aside from the four reaction groups , the CORDA algorithm also requires 5 parameters to operate , which are summarized in Table 4 . To begin the algorithm , all HC reactions are moved into the tissue-specific reconstruction ( RE ) , since these are sure to be included in the final model . Given the remaining three reaction groups , the CORDA algorithm proceeds in three steps: It is worth noting that one of the main advantages of CORDA over pruning algorithms is the fact that it is independent of how reactions are ordered . This is due to the fact that reaction associations are calculated for each step , and at the end of each step a decision is made as to which reactions are added to the tissue reconstruction . This way , the order in which reaction dependencies are calculated does not affect the final tissue reconstruction . The CORDA reconstructions used for comparison to previous methods were generated using γ = 105 , the highest cost value tested , κ = 10-2 , the lowest noise value tested , ϵ = 1 , a threshold similar to a previous study [32] , n = 5 , to allow for the inclusion of a larger number of OT reactions , and p = 2 . For a direct comparison to previous methods , the CORDA reconstructions used during the parameter sensitivity analysis , cross-validation , and comparison to previous methods were performed using the same data used for the mCADRE hepatocyte reconstruction . For the Monte-Carlo sampling analysis , a new reconstruction was generated using the most up-to-date data from the HPA . Both of these reconstructions are available in the supplemental material ( S1 File ) . All calculations in this study were performed using the COBRA toolbox [129] and the Gurobi optimizer [128] . The MATLAB function file used for CORDA reconstructions is also available in the supplemental material ( S2 File ) . Finally , an example of the CORDA algorithm , applied to small sample networks , is available in S2 Text . While CORDA requires a number of different parameters , many of these values can be arbitrarily assigned . For instance , γ can be arbitrarily large , while ϵ and κ can be arbitrarily small . In order to demonstrate that the CORDA algorithm is robust to a wide range of parameters , we performed 108 hepatocyte specific reconstructions varying all parameters but p ( which was set to be equal to two ) to a wide range of values . A separate sensitivity analysis of p was performed and is included in S1 Text . The parameter p can be set in order to define a more or less flexible MC and NC core , and can be set to the user’s discretion . These 108 reconstructions were based on the generalized human reconstruction Recon1 [5] , using the same set of protein expression data ( total of 560 ) and 32 of the metabolic tests used in the mCADRE hepatocyte specific reconstruction [30] . The data used in this step , as well as the metabolic tests and calculated reaction groups , are available in the supplemental information ( S1 Table ) . Metabolic tests were included as single reactions in the reconstruction in order to assure the model was able to produce certain metabolites . Each metabolic test was added to the model when being tested then immediately removed , so that no two tests were present in the model at the same time , and no metabolic test reaction was included when other reactions were being assessed . Details of this analysis are available in S1 Text . During the metabolic tasks validation analysis , the exchange rate of the basal inputs carbon dioxide ( co2[e] ) , water ( h2o[e] ) , protons ( h[e] ) , oxygen ( o2[e] ) , phosphate ( pi[e] ) , hydrogen peroxide ( h2o2[e] ) , superoxide anion ( o2s[e] ) , bicarbonate ( hco3[e] ) and carbon monoxide ( co[e] ) were unconstrained . All other uptake reactions were blocked unless otherwise specified . For each of the 20 amino-acid recycling tests , the uptake rate of the given amino acid and glucose were set to an arbitrary value , so that the amino-acid being tested was the only source or nitrogen . Next , the production of urea was set to a strictly positive value , and FBA was performed while optimizing the production of urea . The same test was also performed for ammonium . For each of the 21 glucogenic tests , the uptake rate of the given metabolite was set to an arbitrary value , and the production of glucose was optimized . For both the amino-acid and glucogenic tests , if the model returned a feasible flux distribution the test was considered passed , otherwise it was considered failed . If the exchange reaction of the given metabolite was not present in the model , the result was considered inconsistent . The generalized Recon1 reconstruction failed two of the glucogenic tests , so the results of the remaining 19 tests are reported in the main text . For the eight nucleotide production tests , a sink consuming the given nucleotide was added to the cytosolic compartment . The model was allowed to uptake glucose and ammonium ( as a source of nitrogen ) , and the flux through the sink was optimized . If the model was able to produce the given nucleotide , the test was considered passed . Following the validation of the CORDA algorithm , we generated a library of 76 healthy and 20 cancer tissue-specific reconstructions using the generalized human reconstruction Recon2 [9] and the most recent proteomics data from the HPA [44 , 45] . All reactions used to generate the tissue-specific models are available in S1 Table , and tissue-specific models are available in SBML and MATLAB format at [130] . The healthy tissue models were calculated using the same classification as described in the algorithm description section , since data for each protein was categorized as not detected , low , medium or highly expressed in each cell type . For cancer models , the same classification was available for any number of samples for each protein in each cancer type . In this case , values of -1 , 1 , 2 and 3 were assigned to each sample according to not detected , low , medium or high expression levels respectively , and these values were averaged for a final protein score in that particular cancer type . These protein values were then used in the gene-reaction boolean association as described in the algorithm description for a final reaction score . Reactions with a score equal to or greater than 2 . 5 were assigned to the HC group , less than 2 . 5 but greater than 1 to the MC group , and less than or equal to -0 . 5 to the NC group . For instance , in renal cancer samples , protein HADHB has been analyzed in 12 different samples in the HPA , and was found to be expressed in high levels in 2 of them , medium levels in 8 , and in low levels in 2 . The protein score associated with HADHB in renal cancer is then calculated as ( 2 · 3 ) + ( 8 · 2 ) + ( 2 · 1 ) 12 = 2 . Similarly , ACAA1 expression was calculated as medium in 5 samples , low in two samples , and not detected in four samples of renal cancer , yielding a score of ( 5 · 2 ) + ( 2 · 1 ) + ( 4 · ( - 1 ) ) 11 = 0 . 73 . Finally , ACAA2 is present in high levels in one sample , medium level in 5 samples , low levels in one sample and not detected in 3 samples of renal cancer , giving this protein a score of ( 1 · 3 ) + ( 5 · 2 ) + ( 1 · 1 ) + ( 3 · ( - 1 ) ) 10 = 1 . 1 . With that , the score for r0634 is calculated as “MIN ( 2 , MAX ( 0 . 73 , 1 . 1 ) ) ” , which is 1 . 1 , putting this reaction in the MC group during the renal cancer reconstruction . Data and reaction distributions used during these calculations can be found in S1 Text . Healthy and cancer specific models were clustered according to reactions present in each model . For that , 4 , 205 reactions present in at least one , but not all models were obtained . A binary vector was then calculated for each model indicating whether reactions were present ( 1 ) or not present ( -1 ) . These vectors were then clustered using hierarchical clustering with Hamming distance as the similarity metric , and average linkage . Leaf orders were also calculated in order to maximize the similarity between neighbors in the hierarchical binary cluster tree dendrogram . These results are summarized in Fig 4 . Next , in order to divide the clusters according to subsystem expression , a total of 4 , 751 reactions present in any of the models was obtained . These reactions were then divided by subsystem according to their classification in the Recon2 reconstruction . For each of the clusters of models calculated in the previous step , the average number of reactions from each subsystem included in the cluster’s models was then calculated . Finally , this number was divided by the total number of reactions in that subsystem which were included in any of the models for a final score between zero and one . These values were then clustered using hierarchical clustering with Euclidean distance as the similarity metric , and average linkage . Leaf orders were again organized to maximize similarity between neighbors to yield Fig 5 . Perhaps the most widely used method to analyze GEMs is Flux Balance Analysis ( FBA ) [42] . FBA predicts a flux distribution through the metabolic network which optimizes ( maximizes or minimizes ) a given objective function , defined as a single reaction or group of reactions in the network . This flux distribution is subject to upper- and lower-bound constraints , which include exchange reactions , and a steady state assumption for all model metabolites , so that no metabolite has a net production or consumption rate . The mathematical formulation of GEMs are defined at the core by a stoichiometric matrix S , where each row defines a metabolite , each column defines a reactions , and each entry the stoichiometric coefficient of that metabolite in that particular reaction . Vectors defining lower ( lb ) and upper ( ub ) bounds for each reaction , as well as an objective vector ( c ) of the same length , are also defined . Given this model , FBA finds a flux vector v through all reactions in the GEM such that: S ⋅ v = 0 l b i ≤ v i ≤ u b i o p t i m i z e : v ⋅ c T During the dependency assessment described here , the stoichiometric matrix S is altered to reflect the changes described above . Given a reaction j being tested , a group of undesirable reactions Y , and a matrix S of size m by n , let κ ¨ denote a random number drawn uniformly between 0 and κ . The GEM is modified in the following ways: With these constraints in place , FBA is performed as described above while minimizing the objective function . For each reaction in the reconstruction , if i ∈ Y and vi ≠ 0 , the reaction i is deemed associated with j . Monte-Carlo sampling was performed in a manner similarly to Bordbar et . al . [25] and Lewis et . al . [49] . This sampling method is a slight variation of the Artificially Centered Hit and Run ( ACHR ) algorithm developed by Kaufman and Smith [131] . In this algorithm , warmup points are initially generated at random corners of the solution space by solving an LP problem with objective vectors containing randomly generated ones and negative ones . The center point between all points is then computed . Next , for each point sampled , a random direction is selected as the difference between a randomly selected point and the center point . By selecting the direction this way , the direction is biased in the longer direction of the solution space , speeding up the rate of mixing while maintaining uniformity . After a direction is chosen , the limit of how far the current point can travel in that direction is calculated , and a new point is randomly chosen along that line . After several iterations , the set of generated points will be well mixed and approach a uniform sampling of the solution space . The termination condition imposed on the ACHR algorithm here is the same imposed by Bordbar et . al . [25] and Lewis et . al . [49] , introducing the concept of mixed fraction . For that , a partition is created over the set of points by drawing a line at the median value , with half the points on either side of the partition . The mixed fraction is the number of points that cross this line during mixing . Initially , the mixed fraction is one as all the points are on their original side of the line . As the sample solutions are mixed , the probability of each point crossing the median line approaches 0 . 5 asymptotically . The sampled points were initially mixed using the warmup points created as described above until the mixed fraction reached a particular threshold . Following that , the samples were mixed two more times , using the previous iteration’s final points as warmup points , until the same mixed fraction was reached . For the comparison between CORDA and other tissue-specific algorithms , a mixed fraction threshold of 0 . 52 was chosen as the termination condition . For the cancer and healthy tissue-specific models , a mixed fraction threshold of 0 . 6 was chosen to make the 96 sampling experiments computationally feasible . Due to the heterogeneity between tissue-specific models , sampled flux values were evaluated between all cancer and healthy tissue models separately . That is , all sampled flux values for the given reaction were obtained from all cancer models that contain that reaction , and compared to all sampled values from healthy tissue models that contain the reaction . Results of this analysis are presented in Fig 6 . In some cases , two or more reactions were combined: MTHFD2* combines reactions MTHFD2 and MTHFD2m , GHMT2r* combines reactions GHMT2r and GHMT2rm , and SPODM* combines reactions SPODM , SPODMe , SPODMm , SPODMn and SPODMx . These are the same reactions taking place in different cellular compartments . For these , flux values from each of these groups of reactions were added within each sampled flux distribution when plotting Fig 6 .
|
Cellular metabolism is defined by a large , intricate network of thousands of components , and plays a fundamental role in many diseases . To study this network in its entirety , metabolic models have been built which encompass all known biochemical reactions in the human metabolism . However , since not all metabolic reactions take place in any given tissue , these generalized models need to be tailored to study specific cell types . Algorithms developed to date to perform this tailoring process have focused on keeping tissue-specific models as concise as possible . This approach , however , can remove essential reactions from the model and hamper subsequent analysis . Here we present CORDA , a tissue-specific building algorithm that yields concise , but not minimalistic , tissue-specific models . CORDA has many advantages over previous methods , including better agreement with experimental data and better model functionality . Using CORDA , we developed a library of 76 healthy and 20 cancer-specific models of metabolism , which we used to identify similarities between healthy and cancerous tissues , as well as metabolic pathways that are unique to cancer . Results of this work provide a broadly applicable tool to model cell- and tissue-specific metabolism , while highlighting potential new pathway targets for cancer therapies .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"physiology",
"medicine",
"and",
"health",
"sciences",
"liver",
"applied",
"mathematics",
"metabolic",
"networks",
"cell",
"metabolism",
"simulation",
"and",
"modeling",
"algorithms",
"mathematics",
"metabolites",
"network",
"analysis",
"exchange",
"reactions",
"pharmacology",
"drug",
"metabolism",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"animal",
"cells",
"metabolic",
"pathways",
"hepatocytes",
"chemistry",
"pharmacokinetics",
"biochemistry",
"cell",
"biology",
"anatomy",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"chemical",
"reactions",
"metabolism"
] |
2016
|
Reconstruction of Tissue-Specific Metabolic Networks Using CORDA
|
Secretory polymorphic serine/threonine kinases control pathogenesis of Toxoplasma gondii in the mouse . Genetic studies show that the pseudokinase ROP5 is essential for acute virulence , but do not reveal its mechanism of action . Here we demonstrate that ROP5 controls virulence by blocking IFN-γ mediated clearance in activated macrophages . ROP5 was required for the catalytic activity of the active S/T kinase ROP18 , which phosphorylates host immunity related GTPases ( IRGs ) and protects the parasite from clearance . ROP5 directly regulated activity of ROP18 in vitro , and both proteins were necessary to avoid IRG recruitment and clearance in macrophages . Clearance of both the Δrop5 and Δrop18 mutants was reversed in macrophages lacking Irgm3 , which is required for IRG function , and the virulence defect was fully restored in Irgm3−/− mice . Our findings establish that the pseudokinase ROP5 controls the activity of ROP18 , thereby blocking IRG mediated clearance in macrophages . Additionally , ROP5 has other functions that are also Irgm3 and IFN-γ dependent , indicting it plays a general role in governing virulence factors that block immunity .
Toxoplasma gondii is an obligate intracellular parasite that infects a wide range of vertebrate animal hosts and causes zoonotic infection in humans , leading to potentially severe congenital infections and risk of reactivation in immunocompromised patients [1] . In North America and Europe , T . gondii exists as four distinct clonal lineages that show marked virulence differences in laboratory mice , which serve as a model for infection [2] , [3] . Forward genetic analyses have been used to map the genes responsible for virulence in laboratory mice [4] , [5] . Remarkably , this complex trait is largely mediated by a few members of a large family of polymorphic serine threonine ( S/T ) protein kinases secreted from rhoptries ( ROP ) into the host cell during invasion [6] , [7] . The ROP kinase family consists of ∼20 active members , as well as a similar number of putative pseudokinases that are predicted to lack kinase activity [8] . The structures of several ROP pseudokinases reveal they contain a typical kinase fold and yet they are structurally and phylogenetically diverse [9] , [10] . Most strains of T . gondii survive within naïve macrophages; however , when previously activated by exposure to IFN-γ macrophages acquire the ability to kill or inhibit parasites [11] . During primary infection , inflammatory monocytes are recruited to the site of infection where they are critical for control of intracellular T . gondii [12] , [13] . Macrophages control T . gondii through induction of iNOS , which leads to stasis [14] , reactive oxygen intermediates , which leads to killing of opsonized parasites [15] , and upregulation of immunity related GTPases ( IRGs ) , which destroy intracellular parasites [16] , [17] . Recruitment of IRG effectors to the parasite containing vacuole results in destruction of the parasite residing within it [18] , [19] . Compared to most mammals , the IRG gene family is highly amplified in rodents [20] , where it plays a major role in natural resistance to T . gondii [21] , [22] . Not all strains of T . gondii are susceptible to clearance in IFN-γ-activated macrophages: highly mouse virulent type I parasites resist IRG recruitment and consequently avoid clearance , while intermediate virulent type II and avirulent type III parasites are unable to block IRG recruitment and are destroyed [14] , [23] . Recent studies have revealed the mechanism for this escape: the S/T kinase ROP18 phosphorylates a number of IRGs on key threonine residues in switch region I of the GTPase domain , thereby preventing assembly on the vacuole and blocking clearance in activated macrophages [24] , [25] . The IRG system is also dependent on the autophagy protein Atg5 , although the molecular basis for this requirement remains unclear [23] , [26] . Genetic mapping studies have also implicated the pseudokinase ROP5 in acute virulence [27] , [28] , a result that is unlikely to be due to kinase activity as ROP5 lacks a key catalytic residue and binds ATP in an unconventional manner [29] . Comparison of several pair-wise genetic crosses suggest that ROP5 interacts with ROP18 [27]; however , the molecular basis for the dramatic effects of ROP5 on virulence remains uncertain . Pseudokinases in other systems have recently been shown to perform regulatory roles [30] , raising a similar possibility for ROP5 . Herein , we explore the mechanism of action for ROP5 and demonstrate that it controls ROP18 activity , while also serving a separate and essential role in acute virulence .
Previous studies have shown that ROP5 deficient ( RHΔku80Δrop5 ) parasites grow normally in vitro , but are highly attenuated in laboratory mice [27]; however , the molecular basis of this phenotype is not understood . To examine growth in vivo , CD-1 outbred mice were infected with either wild type ( RHΔku80 ) or ROP5 deficient ( RHΔku80Δrop5 ) parasites expressing luciferase and growth was followed over time . Virulent RHΔku80 parasites expanded rapidly , as detected by luciferase activity , until the mice succumbed to infection between days 6–8 ( Figure 1A , Figure S1 ) . In contrast , ROP5 deficient ( RHΔku80Δrop5 ) parasites expanded normally for the first few days and then dramatically decreased by the end of the first week , resulting in survival of the mice ( Figure 1 , Figure S1 ) . To examine innate immune responses , we measured inflammatory cytokines in serum during the first week post infection . The levels of IFN-γ , IL-6 , and MCP-1 increased at day 3 and were all significantly higher at day 5 ( P≤0 . 01 ) in mice infected with wild type ( RHΔku80 ) parasites as compared to ROP5 deficient ( RHΔku80Δrop5 ) parasites ( Figure 1A , B ) . Consistent with a rise in IFN-γ , we also observed an increase in IL-12p40 at day 3 and 5 in mice infected with wild type ( RHΔku80 ) parasites . In contrast , IL-10 and TNFα were essentially unchanged . Overall , changes in cytokine levels appeared to track closely with parasite burden . To test whether ROP5 modulates host cell transcription , we infected human foreskin fibroblasts ( HFF ) with either wild type ( RHΔku80 ) or ROP5 deficient ( RHΔku80Δrop5 ) parasites and examined gene expression using the Affymetrix HG-U113A_2 Human Array . There were no significant differences in gene expression in HFF infected with wild type ( RHΔku80 ) or ROP5 deficient ( RHΔku80Δrop5 ) parasites ( NCBI GEO record GSE32104 ) , indicating ROP5 does not directly modulate host gene expression , at least under the conditions tested in vitro . The failure of ROP5 deficient parasites to expand in vivo could emanate from some alteration in the immune response . For example , if ROP5 were normally immunosuppressive , in its absence , a stronger or more potent immune response might provide more effective control of infection . To test whether ROP5 deficient parasites lack a normally suppressive function , mice were infected separately , or coinfected with wild type ( RHΔku80 ) and ROP5 deficient ( RHΔku80Δrop5 ) parasites , and followed for 30 days to assess survival . Consistent with previous studies , mice infected with wild type ( RHΔku80 ) or ROP5 complemented ( RHΔku80Δrop5 Complement ) parasites succumbed in 9–10 days ( Figure 2A ) . Although ROP5 deficient parasites failed to cause lethal infection , coinfection with wild type parasites led to rapid death of all mice by day 10 ( Figure 2A ) . Moreover , mice immunized with ROP5 deficient parasites were able to generate a normal adaptive immune response and survive a secondary challenge with a normally lethal dose of wild type parasites ( Figure 2B ) . These findings indicate that it is unlikely that ROP5 is globally immunosuppressive . Alternatively , it was possible that ROP5 deficient parasites were metabolically restricted in vivo , similar to the previously reported pyrimidine biosynthesis mutants ( i . e . cpsII mutants ) , which are unable to propagate in vivo due limitations in uracil for salvage [31] . cpsII mutants fail to expand in both wild type and Ifng−/− mice [31] , consistent with their metabolic limitation . To determine if ROP5 deficient ( RHΔku80Δrop5 ) parasites were controlled by an IFN-γ-dependent mechanism , we tested the virulence of ROP5 deficient ( RHΔku80Δrop5 ) parasites in Ifngr1−/− mice , which are unable to respond to IFN-γ and hence , are highly susceptible to toxoplasmosis [32] . Ifngr1−/− mice were completely susceptible to infection with ROP5 deficient parasites , succumbing in the same time frame as wild type C57BL/6 or Ifngr1−/− mice infected with virulent wild type parasites ( Figure 2C ) . To further clarify the role of ROP5 in immune evasion we tested the virulence of ROP5 deficient ( RHΔku80Δrop5 ) parasites in mice lacking the inducible nitric oxide synthase ( iNOS ) ( Nos2−/−mice ) , the superoxide-generating NADPH-oxidase gp91phox ( X-CGD mice ) , or the recombination activating gene 1 ( Rag1−/− mice ) , important for B and T cell function . iNOS−/− and X-CGD mice survived infection with ROP5 deficient ( RHΔku80Δrop5 ) parasites ( Figure 2D ) and did not present symptoms of illness or weight loss ( data not shown ) , similar to C57BL/6 control mice ( Figure 2C ) . Although Rag1−/− mice succumbed to ROP5 deficient ( RHΔku80Δrop5 ) infection , death was delayed compared to wild type ( RHΔku80 ) infection ( Figure 2D ) . Collectively these findings indicate that IFN-γ signaling is necessary for controlling ROP5 deficient parasites and suggest that ROP5 is critical to the parasites' ability to resist innate immune effectors generated by the IFN-γ response . Previous studies have shown that inflammatory monocytes , which are recruited to the peritoneal cavity following i . p . infection [13] , are critical for controlling toxoplasmosis in mice [12] . To determine if the altered growth kinetics of ROP5 deficient ( RHΔku80Δrop5 ) parasites in mice were due to differences in cellular recruitment , we examined the frequency of myeloid cells in the peritoneum at intervals after infection by FACS . Infection with wild type ( RHΔku80 ) , ROP5 deficient ( RHΔku80Δrop5 ) and ROP5 complemented ( RHΔku80Δrop5 Complement ) parasites all induced robust recruitment of inflammatory monocytes to the peritoneal cavity by day 3 ( Figure 3A , B middle gate ) . Although the numbers of inflammatory monocytes were similar at day 3 , the number of resident monocytes drastically decreased in mice infected with all three strains of T . gondii compared to uninfected mice ( Figure 3B , Figure S2 ) , a result that is likely due to cell lysis as a consequence of high parasite replication at this time point ( Figure 1A ) . The number of inflammatory monocytes at day 5 also correlated with parasite growth , declining in mice infected with wild type or ROP5 complemented parasites ( Figure 3A ) , while remaining elevated in mice infected with ROP5 deficient parasites . Increased numbers of neutrophils were also observed in infected mice ( Figure S2 ) , a phenomenon that has been seen previously in association with high parasite burdens [12] . Collectively these findings indicate that ROP5 deficient parasites initially expand in resident macrophages , but that following recruitment of inflammatory monocytes , the infection is controlled . Based on the differences in initial growth in vivo ( Figure 1A ) , we tested the ability of ROP18 deficient and ROP5 deficient parasites to survive in resident peritoneal macrophages ( Gr1− F4/80+ ) vs . inflammatory monocytes ( Gr1+ F4/80+ ) . Naïve macrophages isolated from the peritoneal cavity of normal mice showed limited ability to clear either wild type ( RHΔku80 ) , ROP18 deficient ( RHΔku80Δrop18 ) , ROP5 deficient ( RHΔku80Δrop5 ) or ROP5 complemented ( RHΔku80Δrop5 Complement ) parasites ( Figure 4A ) . In contrast , parasites that were deficient in either ROP18 or ROP5 were efficiently cleared by Gr1+ monocytes in vitro ( Figure 4B ) , a result that was not accompanied by loss of cells from the monolayer . Survival was completely restored to wild type levels in a strain genetically complemented for ROP5 expression ( Figure 4B ) . Since ROP18 is known to enhance survival through disrupting IRG recruitment to the parasite containing vacuole [24] , we examined the cellular localization of Irgb6 after infection with ROP18 or ROP5 deficient parasites . Although Irgb6 remained diffuse in cells infected with either wild type ( RHΔku80 ) and ROP5 complemented ( RHΔku80Δrop5 Complement ) parasites , both the ROP18 and ROP5 deficient parasites readily accumulated Irgb6 on the vacuole membrane . The level of Irgb6 recruitment was lower than the extent of clearance in the overnight assay , a result that may be due to kinetic differences ( Figure 4C , D ) . Collectively , these results show that inhibition of recruitment of IRGs by virulent strains of T . gondii requires both ROP18 and ROP5 . Given that ROP18 has previously been shown to be both necessary and sufficient to subvert IRG clearance in murine macrophages infected with T . gondii [24] , it became important to determine how ROP5 influences this pathway . The dependence of ROP18-dependent functions on ROP5 might result from either altered expression or localization of ROP18 . Western blot analysis demonstrated that ROP18 was expressed at near wild-type levels in the absence of ROP5 ( Figure 5A , B ) . Immunofluorescence analysis detected ROP18 properly localized at the vacuolar membrane in both wild type cells ( RHΔku80 ) and ROP5 deficient ( RHΔku80Δrop5 ) parasites ( Figure 5C , D ) . Collectively , these findings indicate that absence of ROP5 is not responsible for altered expression or localization of ROP18 . Some mammalian pseudokinases have been reported to allosterically regulate active kinases [30] , suggesting that ROP5 may regulate ROP18 . To test this hypothesis in vitro , recombinant ROP18 was used to phosphorylate the artificial substrate dMBP ( Figure 5E ) , or the natural substrate Irgb6 ( Figure 5F ) , in the presence or absence of recombinant ROP5 . To determine whether a similar interaction occurs in vivo , ROP18 was immunoprecipitated from wild type ( RHΔku80 ) parasites or those that were ROP5 deficient ( RHΔku80Δrop5 ) ( Figure 5G , I ) and used to phosphorylate dMBP ( Figure 5H ) or Irgb6 ( Figure 5J ) in vitro . In the absence of ROP5 , both recombinant ROP18 ( Figure 5E , F ) and endogenous ROP18 ( Figure 5H , J ) , demonstrated a greatly diminished capacity to phosphorylate both dMBP and Irgb6 substrates based on 32PO4 labeling . Activity of recombinant ROP18 increased ∼12 fold ( Figure 5 E , F ) , while a ∼35 fold increase in endogenous ROP18 activity was observed in the presence of ROP5 ( Figure 5J ) . Restored expression of ROP5 in the complemented clone resulted in phosphorylation of dMBP or Irgb6 by immunoprecipitated ROP18 ( Figure 5 H , J ) . The enhanced activity of ROP18 in the presence of ROP5 did not result from a stable complex between these proteins , as they failed to co-immunprecipitate in lysates of infected , IFN-γ activated cells ( Figure S3A ) . Enhanced activity of ROP18 in the presence of ROP5 also did not result from an interaction between ROP5 and Irgb6 ( Figure S3B ) , nor did the previously demonstrated interaction between ROP18 and Irgb6 [24] , require the presence of ROP5 ( Figure S3C ) . Although ROP5 activated ROP18 , it failed to demonstrate catalytic activity of its own in vitro ( Figure 5E , F ) , consistent with the prediction that it encodes a pseudokinase [27] , [28] . Taken together , these results indicate that the catalytic activity of ROP18 is regulated by the predicted pseudokinase ROP5 , and that this pathway is required for avoidance of IRG clearance in inflammatory monocytes . Our results indicate that ROP5 and ROP18 are both required for escape from the IRG pathway , consistent with the finding that ROP5 regulates the activity of ROP18 . We sought to determine whether the defects in ROP5 and ROP18 could be compensated by defects in the IRG pathway . However , of the two IRG proteins that are shown to be targeted by ROP18 , there is presently no knockout available for Irgb6 , and the phenotype of Irga6 mutants challenged with T . gondii is very modest , especially in cell-autonomous control of parasite survival [33] . Nevertheless , Irgm proteins are known to regulate the proper recruitment of Irga6 and Irgb6 to the vacuole surrounding susceptible strains of T . gondii [34] , [35]; absence of Irgm1 or Irgm3 alters the targeting and function of Irgb6 and Irga6 and effectively cripples the IRG system . However , the use of mice lacking Irgm1 is complicated by pleomorphic effects of Irgm1-deficiency on T cell development [36] and macrophage motility [37] . In contrast , Irgm3-deficient mice have shown normal immune cell development , yet are highly susceptible to infection by T . gondii [17] . Heterologous expression of tagged proteins indicates that Irgm proteins are necessary for proper recruitment of Irga6 and Irgb6 to the vacuole surround susceptible strains of T . gondii [35] , suggesting the same requirement might be true for endogenous proteins . To explore the interaction between ROP5 and ROP18 and the IRG pathway , we examined the recruitment of Irgb6 and parasite clearance in IFN-γ-activated , bone-marrow-derived macrophages derived from wild type C57BL/6 and Irgm3−/− mice . To provide a more complete set of parasite strains for this comparison , we complemented the RHΔku80Δrop18 mutant described previously [24] by reintroducing a single copy of ROP18 that restored normal expression and reversed the virulence defect seen in outbred mice ( Figure 6A , B , Figure S4 , S5 ) . Parasites deficient in ROP18 or ROP5 demonstrated decreased survival in IFN-γ activated wild type macrophages compared to their respective complemented lines or the wild type strain ( RHΔku80 ) , which is resistant to clearance as previously reported [24] ( Figure 6C ) . The enhanced clearance of ROP5 or ROP18 deficient parasites was largely reverted in IFN-γ activated Irgm3−/− macrophages ( Figure 6D ) . Vacuoles containing both ROP18 deficient ( RHΔku80Δrop18 ) and ROP5 deficient ( RHΔku80Δrop5 ) parasites showed enhanced Irgb6 accumulation that was restored to normal in the complemented parasite strains in wild type macrophages ( Figure 6 E , F ) . Additionally , the enhanced recruitment of Irgb6 seen in ROP5 or ROP18 deficient mutants was restored to normal in the absence Irgm3 ( Figure 6 E , F ) . Deletion of Irgm3 also affected the abundance and pattern of Irgb6 , which tended to aggregate in clusters in the absence of Irgm3 ( Figure 6F ) . These studies reinforce the model that recruitment of Irgb6 is dependent on Irgm3 and establish that ROP18 and ROP5 have indistinguishable phenotypes when it comes to survival in activated macrophages in vitro . To examine survival in vivo , wild type and Irgm3−/− mice were challenged with parasites and luciferase activity and survival were recorded . Following s . c . challenge of C57BL/6 mice , wild type ( RHΔku80 ) parasites rapidly expanded while ROP5 deficient ( RHΔku80Δrop5 ) parasites were controlled as shown by luciferase imaging studies ( Figure 7A ) . Interesting , ROP18 deficient ( RHΔku80Δrop18 ) parasites expanded with delayed kinetics in C57BL/6 mice and tissue burdens had begun to recover when animals succumbed to infection ( Figure 7A ) . A similar response was also seen in CD1 outbred mice challenged with ROP18 deficient ( RHΔku80Δrop18 ) parasites ( Figure S4 ) . In the absence of Irgm3 , both wild type ( RHΔku80 ) and ROP18 deficient ( RHΔku80Δrop18 ) parasites underwent rapid expansion and reached high tissue burdens as reflected by luciferase activity , while ROP5 deficient ( RHΔku80Δrop5 ) parasites showed a delayed expansion and reached lower total levels ( Figure 7A ) . The expansion of parasites observed by bioluminescence imaging mirrored survival outcomes . Challenge with wild type strain ( RHΔku80 ) parasites led to rapid and complete mortality of both wild type and Irgm3−/− mice ( Figure 7B ) . Wild type mice infected with ROP18 deficient ( RHΔku80Δrop18 ) parasites exhibited a delayed death phenotype similar to that seen in outbred mice ( Figure 6B ) , while Irgm3−/− mice succumbed rapidly , similar to wild type parasite infection ( Figure 7B ) . In contrast , ROP5 deficient ( RHΔku80Δrop5 ) parasites were completely avirulent in wild type C57BL/6 mice , while they caused 100% mortality in Irgm3−/− mice , albeit with a delay in time to death ( Figure 7B ) . The susceptibility of Irgm3−/− mice infected with ROP5 deficient ( RHΔku80Δrop5 ) parasites was more rapid when injected i . p . with a similar time to death as wild type parasites ( Figure S6 ) .
Although the pseudokinase ROP5 was previously shown to be essential for acute virulence of T . gondii in laboratory mice , the basis for this was initially unclear , especially given the predicted lack of catalytic activity of this protein . Here we demonstrate that ROP5 regulates the activity of ROP18 , an active S/T kinase that phosphorylates IRGs , thus blocking their accumulation on the parasite containing vacuole . ROP5 was necessary for the full enzymatic activity of ROP18 , although it was not required for stable expression or normal trafficking to the parasite-containing vacuole . Studies using Irgm3 deficient macrophages revealed that the inability of ROP5 and ROP18 deficient parasites to avoid IRG recruitment was fully reverted in vitro . Moreover , the attenuation of the ROP deficient mutants was fully reversed in Irgm3 deficient mice . These findings reveal that ROP5 is a multifunctional pseudokinase that regulates acute virulence in T . gondii in part by governing the active kinase ROP18 , and by affecting additional effectors that are both IFN-γ and Irgm3-dependent . ROP5 is a member of a polymorphic family of secretory S/T kinases that are highly divergent from human kinases and which have been amplified in the genome of T . gondii [8] . Forward genetic mapping revealed that ROP5 is primarily responsible for differences in mouse virulence between highly virulent type I strains and intermediate virulent type II strains [27] , and also between type II and avirulent type III strains , although paradoxically the type III ROP5 locus is positively associated with virulence [28] . In all strains , the ROP5 locus encodes a cluster of predicted pseudokinases all of which lack the central conserved Asp residue of the catalytic triad typical of S/T kinases [38] . Our findings demonstrate that the virulence defect in ROP5 deficient parasites is completely reversed in mice lacking Ifngr1−/− or Rag1−/− , indicating that ROP5 mediates escape from IFN-γ-dependent effector mechanisms . Alternative models , such as ROP5 deficient parasites being auxotrophic for nutrients that may be limiting in vivo , or ROP5 being a global suppressor of immune responses , are not supported . Previous studies have shown that IFN-γR1 is required for control of T . gondii in both hematopoietic and non-hematopoietic cells [32] , and both compartments likely contribute to IRG-mediated clearance , which in the mouse provides one of the most effective means of control [21] , [22] . At the level of survival in macrophages in vitro , ROP5 and ROP18 were both required for avoidance for recruitment of IRGs and clearance . In previous studies we have shown that ROP18 deficient parasites exhibit normal survival in naive macrophages , but are restricted in IFN-γ activated peritoneal macrophages and that survival correlates with avoidance of Irgb6 recruitment [24] . Here we extend these findings to show that ROP18 or ROP5 deficient parasites show enhanced Irgb6 recruitment and clearance in Gr1+ monocytes , and in bone marrow derived macrophages activated in vitro with IFN-γ . In separate studies , others have shown that ROP18 or ROP5 deficient parasites also fail to block recruitment of Irga6 and Irgb6 in IFN-γ-activated MEFs [39] . In the present study , the increased susceptibility of ROP18 and ROP5 deficient parasites to clearance by IFN-γ-activated macrophages was completely dependent on Irgm3 , a regulatory protein required for homeostasis of IRGs . Moreover , deficiency in Irgm3 reverted the phenotype of both the Δrop18 and the Δrop5 mutants in vivo . Previous genetic analyses of acute virulence in type I strains of T . gondii indicated that ROP5 and ROP18 interact to control virulence [7] , [27] . We now demonstrate that the basis for this relationship is that ROP5 controls the kinase activity of ROP18 , thus affecting its ability to phosphorylate substrates . ROP5 activation of ROP18 activity contributes to avoidance of IRG recruitment in IFN-γ activated macrophages , hence promoting survival . ROP18 actively phosphorylates a number of IRG proteins in a common motif in switch region I [24] , thereby affecting GTPase activity , and oligomerization [25] . Following phosphorylation , IRG proteins are unable to load onto the parasite containing vacuole , thus blocking this interferon-mediated clearance pathway [40] . Previous studies have indicated that Irgm3 is recruited to the PVM [19] , [41] , and it contains the conserved motif [24] , and therefore is a potential substrate of ROP18 . Additionally , the activity of ROP18 in phosphorylating ATF6β [42] , is likely to also dependent on ROP5 . ATF6β has been proposed to affect a later step in resistance mediated through dendritic cell activation of T-cells , a process likely important in adaptive immunity [42] . At an earlier stage , ROP18 is essential for controlling avoidance of the IRG pathway , a process that participates primarily in innate immunity [43] . Collectively , these two pathways likely control the ROP5-dependent activities of ROP18 in mediating virulence . Our findings are consistent with a model whereby ROP5 acts as an allosteric regulator of ROP18 . ROP5 was required for full catalytic activity of ROP18 using endogenous enzyme immunoprecipitated from cells and in vitro testing against the heterologous substrate dMBP or the natural substrate Irgb6 . ROP5 also directly activated the kinase activity of recombinant ROP18 in vitro against both dMBP and Irgb6 . These results are reminiscent of recent reports of a role for pseudokinases in mammalian cells regulating their active partners [30] , [44] . For example , the pseudokinase STRADα regulates LKB1 , a S/T protein kinase that regulates AMP activated kinase and acts as a tumor suppressor [45] . Although catalytically inactive , STRADα adopts a closed conformation typical of an active kinase and together with the adaptor MO25α promotes the active conformation of LKB1 [46] , [47] . Unlike the situation with LKB and STRADα , the activation of ROP18 occurs despite there not being a strong interaction with ROP5 , which does not coIP from cell lysates [48] , and present report . However , the failure to observe a stable complex under these conditions does not preclude ROP5 and ROP18 from interacting at a lower affinity , or in a complex that depends on local interactions on the PVM . It is conceivable that transient binding of ROP5 to ROP18 facilitates auto-catalytic activation , which results from phosphorylation in helical extensions of the N-lobe of the kinase domain [10] . In separate studies , using a more sensitive approach based on tandem-affinity purification ( TAP ) [49] of ROP5 , ROP18 was one of the major components to copurify in a complex with ROP5 , and this was validated by reciprocal TAP-tagging of ROP18 ( Etheridge , Sibley unpublished ) . Moreover , we observed that ROP5 is found in a complex with other ROP kinases , suggesting it may activate other kinases similar to ROP18 ( Etheridge , Sibley unpublished ) . Although the precise mechanism of regulation is yet unclear , our data indicate that ROP5 is an allosteric activator of ROP18 , thus establishing a new role for pseudokinases in controlling pathogen virulence factors . Our findings differ from a recent report that also examined the interaction of ROP5 and ROP18 [48] . This prior study reported that co-expression of a cosmid contain the locus of ROP5 from the type I strain was not able to enhance the activity of ROP18 from a type II strain , when co-transfected into a recombinant strain called S22 [48] . Interpretation of this experiment is complicated by the fact that S22 is the product of recombination between types II and III and it also contains the type II ROP5 locus , which has previously been associated with avirulence [27] . Either due to this complex mixture of ROP5 alleles , or another undefined locus , this strain may harbor an epistatic activity that suppresses activation of ROP18 . In contrast we demonstrate that the major allele of ROP5 from the type I strain activates the kinase activity of ROP18 from a type I strain in vivo , using isogenic strains , and in vitro using purified recombinant protein to phosphorylate both heterologous and endogenous substrates . This later result was also observed using GST-Irga6 as a substrate in vitro [39] , confirming the ability of ROP5I to enhance the activity of ROP18I . Two recent studies also reported that ROP5 binds directly to some IRGs , notably Irga6 , affecting its oligomerization and GTPase activity in vitro [39] , [48] . This result suggests that ROP5 may also inhibit oligomerization in vivo , thus decreasing IRG accumulation on the parasite-containing vacuole . However , this activity alone is unlikely to be sufficient for parasite survival , given the inability of type III strains to avoid IRG recruitment despite expressing the same complement of ROP5 alleles as seen in type I strains [27] . Additionally , although ROP5 binds reasonably well to Irga6 , it binds less efficiently to other IRGs such as Irgb6 ( [39] and present study ) . As such , the ability of ROP5 to directly activate the kinase activity of ROP18 may be more important for targets such as Irgb6 . Collectively , the binding of ROP5 to IRGs and inhibition of oligomerization is expected to work cooperatively with its ability to enhance the catalytic activity of ROP18 , thus disrupting IRG function . The IRG pathway has been described as a major immunity mechanism in the murine system due to the expansion of this family of innate immune effectors in the rodent lineage [20] . The nearly exclusive expression of IRGs in the murine system has led some to question its relevance to human infection . However , this view overlooks the obvious importance of rodents in the natural transmission of toxoplasmosis , which is a zoonotic disease that humans acquire from infected food animals and cats , although not directly from rodents . Additionally , there are several reasons to believe IRGs are also directly relevant to humans . Humans express only two IRG family members: IRGC , which is testis specific and unlikely to be involved in immunity , and IRGM , which is truncated [20] . Despite likely not being a functional GTPase , IRGM has been implicated in autophagy-mediated control of Mycobacterium tuberculosis [50] and Salmonella typhimurium [51] . Additionally , both humans and rodents express a second family of related GTPases called guanylate binding proteins ( GBPs ) , which are also strongly upregulated following treatment with IFN-γ [52] . GBPs have recently been shown to be required for the control of Listeria monocytogenes and M . tuberculosis in the mouse [53] . GBPs are also recruited to vacuoles containing T . gondii in a strain-dependent manner [52] , [54] , suggesting a role in pathogen control . Direct evidence for such a role was recently provided by a study reporting deletion of a locus on chromosome 3 in the mouse , encoding 5 GBPs , impairs immunity to challenge with a type II strain of T . gondii [55] . In separate studies , we have shown that these effects are partially dependent on Gbp1 and that recruitment of Gbp1 to the PVM is mediated in a ROP5 and ROP18-dependent manner ( Selleck , Sibley submitted ) . Homeostasis of GBPs requires Irgm proteins in the murine system [56] , hence the dramatic reversal of the ROP5 deficient parasites in Irgm3−/− mice may be due to defects in both the IRG and GBP systems . Further studies will be needed to determine the relationship between these different IFN-γ induced systems and to define the role of ROP kinases and IRG-dependent immunity mechanisms in control of human infection . Previous genetic crosses have implicated only a few loci in controlling acute virulence in the mouse model [6] , [22] , [28] , [57] . Notably , only type I parasites are efficient at blocking IRG clearance [14] , [23] and they have a combination of a type I allele at ROP18 and a type I allele at ROP5 , the latter of which is also expressed by type III parasites [7] , [27] . Consistent with their extremely low level of ROP18 expression , type III parasites are avirulent , a phenotype that is fully reverted with transgenic expression of type I [7] or type II [6] ROP18 . Although type II strain parasites have a functional ROP18 , they have a type II ROP5 locus that is associated with avirulence . Although we have not tested the ability of type II ROP5 to regulate the activity of the type II allele of ROP18 , genetic studies indicate that this interaction is not sufficient to promote full virulence [28] , [57] , nor is it sufficient to mediate avoidance of IRG recruitment and clearance [14] , [23] . The more severe defect of ROP5 deficient parasites vs . ROP18 deficient parasites in wild type mice , suggest that in addition to regulating ROP18 , it has other functions , perhaps serving as scaffold for regulating other important kinases in T . gondii . ROP kinases are highly polymorphic and have expanded in the T . gondii genome under strong selective pressure [8] . Similarly , the IRG pathway is highly amplified in rodents where it plays a major role in resistance to intracellular pathogens such as T . gondii . Placing the ROP5 pseudokinase at the center of this pathway may be an evolutionary strategy to divert attention from the active kinases , in which diversity is constrained to preserve catalytic activity .
All animal experiments were conducted according to the U . S . A . Public Health Service Policy on Humane Care and Use of Laboratory Animals . Animals were maintained in an AAALAC-approved facility and all protocols were approved by the Institutional Animal Care and Use Committee ( School of Medicine , Washington University in St . Louis ) . CD-1 and C57BL/6 mice were purchased from Charles River Laboratories . Ifnγr1−/− , Rag1−/− , and Nos2−/− mice , , all on a C57/BL6 background were obtained from Dr . Herbert Virgin ( Washington University ) . Mice deficient in the superoxide-generating NADPH-oxidase gp91phox subunit ( NOX2 ) , referred to as X-CGD mice , on a C57BL/6 background [58] were obtained from Dr . Mary C . Dinauer ( Washington University ) . Irgm3−/− mice [17] from Dr . Greg Taylor ( Duke University ) were bred locally at Washington University . Age and sex- matched mice were challenged by i . p . or s . c . injection with parasites and survival followed for 30 days post injection , as described previously [27] . Mice inoculated with luciferase expressing parasites were weighed at intervals after infection and imaged by bioluminescence as described previously [59] . For cytokine measurements , blood was obtained from the saphenous vein . Sera were obtained using microtainer serum separators ( BD Bioscience ) by centrifugation at 3 , 000 g and stored at −20°C . IL-12p40 was measured using a mouse OptEIA ELISA kit ( BD Biosciences ) . The other cytokines were quantified using the Cytometrix Bead Array Mouse Inflammation Kit ( BD Biosciences ) , detected using a FACS Canto II flow cytometer ( BD Biosciences ) , and analyzed using FCAP ARRAY ( Soft Flow , Inc . ) . Fully virulent , type I RH strain parasites that are deficient in Ku80 ( RHΔku80 ) , attenuated Δrop18 ( RHΔku80Δrop18 ) , avirulent Δrop5 ( RHΔku80Δrop5 ) , and virulent ROP5 complemented ( RHΔku80Δrop5Complement ) parasite strains were described previously [24] , [27] . Generation of the ROP18 complement ( RHΔku80Δrop18Complement ) is described in Figure S4 . Parasites were serially passaged in human foreskin fibroblasts ( HFF ) monolayers , as described previously [24] . Cultures were negative for mycoplasma contamination using the e-Myco plus mycoplasma PCR detection kit ( Boca Scientific ) . Peritoneal macrophages harvested from naïve CD1 mice , bone marrow derived macrophages , and RAW 264 . 7 cells ( American Type Culture Collection ( ATCC ) ) were cultured as described previously [24] . Cells were activated by treatment overnight with either 10 or 50 units/mL murine IFN-γ ( R&D Systems ) and 0 . 1 ng/mL LPS from E . coli O55:B5 ( Sigma-Aldrich ) . Gr1+ monocytes were harvested at day 4 from the peritoneal cavity of mice that have been primed 4 days previously by inoculation with 200 CTG strain parasites , as described previously [24] . Macrophages were characterized by expression of the cell surface markers using mAb RB6-8C5 against Gr1 that was directly conjugated to Alexa Fluor 594 and/or mAb HB-198 ( ATCC ) against F4/80 that was directly conjugated to Alexa Fluor 488 , using commercially available coupling kits ( Invitrogen ) . The pDestR4R3-UPRTKO-Clickluc plasmid was constructed with the MultiSite Gateway 3-fragment pDest-R4R3 system ( Invitrogen ) . Flanking fragments ( 1 kb 5′ pDONR-P41r and 3′ pDONR-P2rP3 ) for the T . gondii uracil phosphoribosyl transferase gene ( UPRT ) gene were amplified from RH strain lysate with iProof High Fidelity DNA polymerase ( Bio-Rad ) . The middle fragment ( pDONR-P1P2 ) contained the Click Beetle luciferase ( Clickluc ) gene driven by the dihydrofolate reductase ( DHFR ) promoter as described [60] ( Table S1 ) . RHΔku80 , RH Δku80Δrop18 , or RHΔku80Δrop5 parasites ( 1 ) , were electroporated with pDestR4R3-UPRTKO-Clickluc linearized with BsiWI , selected with fluorodeoxyuridine ( FUDR ) ( 1×10−5 M ) , and luciferase positive single cell clones were identified by positive bioluminescence . A 5-fragment Gateway clone was generated to integrate the ROP18 gene under control of the IMC1 promoter into the UPRT locus ( Figure S5 ) . The targeting construct was amplified by PCR as described previously [61] , and electroporated into RHΔku80Δrop18 strain parasites , stable transformant clones isolated and verified by immunofluorescence staining and western blotting . Recruitment of cells to the peritoneal cavity was analyzed by FACS , using protocols described previously [59] . In brief , mice were sacrificed at intervals after infection , peritoneal cells were isolated , and stained and analyzed by FACS . Cells were incubated at 4°C in Fc block ( clone 2 . 4G2 , BD Bioscience ) in MACS buffer ( PBS , 0 . 5% BSA , 2 mM EDTA , pH 7 . 2 ) and negative cells excluded by staining with AmCyan-Aqua Fixable Dead Cell Stain ( Invitrogen ) . Labeled antibodies V450 anti-Gr1 ( RB6-8C5 , BD Bioscience ) , APC-anti-F4/80 ( Invitrogen ) , PE-Cy7 anti-CD11b ( M1/70 , BD Bioscience ) , and APC-eFluor 780 anti-B220 ( RA3-6B2 , eBioscience ) were incubated for 15 min at room-temp , washed and re-suspended in MACS buffer . Samples were detected using a FACS Canto II flow cytometer ( BD Biosciences ) , and analyzed with FlowJo software ( Tree Star , Inc ) . Absolute cell numbers were calculated using the total cell count multiplied successively by the percentages for the appropriate gates obtained through flow cytometry . A total of 100 , 000 cells were analyzed for each sample . Infected HFF or macrophage monolayers cultured on coverslips were fixed in 4% formaldehyde and permeablized in 0 . 05% Triton X-100 in PBS or 0 . 05% saponin for 10 min . Samples were blocked with 10% FBS , incubated with primary antibodies for ∼20 min , washed 3 times with PBS , and incubated with species-specific secondary antibodies conjugated to Alexa Fluors ( Invitrogen ) for ∼20 min . Samples were rinsed in PBS , mounted in ProLong Gold with DAPI ( Invitrogen ) and examined with a Zeiss Axioskop 2 MOT Plus microscope ( Carl Zeiss , Inc . ) . Images were acquired with an AxioCam MRm camera ( Carl Zeiss , Inc . ) and processed with Photoshop CS4 . Macrophage monolayers were challenged with parasites that were propagated in HFF cells as described above . Clearance was assessed by comparing the percentage of cells infected following a 30 min infection pulse vs . those remaining after 20 h , as described previously [24] . The numbers of infected cells were determined by counting of 10 fields using a 40× objective lens from 3 replicates per condition . Two or three replicate experiments were performed for each assay . Gr1+ monocytes or IFN-γ treated bone marrow derived macrophages were challenged with parasites for 30 min , fixed in formalin buffered saline , and processed for immunofluorescence , as described previously [24] . T . gondii containing vacuoles were stained with mAb Tg17–113 , which recognizes dense granule protein 5 ( GRA5 ) [62] . Irgb6 was localized with rabbit anti-Irgb6 [34] , and the numbers of positive vacuoles were determined by counting of 10 fields using a 40× objective lens from 3 replicates per condition . Two or three replicate experiments were performed for each assay . Parasite lysates were resuspended in denaturing Laemmli sample buffer , resolved in 10% acrylamide gels , transferred to nitrocellulose and probed with rabbit anti-ROP18 [24] , rabbit anti-ROP5 [27] , or rabbit anti-actin [63] . Blots were washed , incubated with goat anti-rabbit IgG conjugated to HRP ( Jackson ImmunoResearch ) , and detected using the ECL Plus western blotting system ( GE Healthcare ) and FLA5000 phosphorimager analysis ( Fuji Life Sciences ) . Immunoprecipitations were performed as described previously [24] . In brief , cells were lysed 1% NP-40 , 50 mM Tris–HCl , 150 mM NaCl , pH 8 . 8 plus protease inhibitors , centrifuged at 1 , 000 g 4°C , and pre-cleared by incubation with protein G sepharose ( Pierce Biotechnology Inc . ) for 1 h at 4°C . ROP18 was immunoprecipitated using either the mAb BB2 against the Ty-1 tag , or polyclonal rabbit anti-ROP18 . Irgb6 was immunoprecipitated from IFN-γ activated bone marrow derived macrophages with rabbit anti-Irgb6 , and purity was confirmed by MS/MS , as described previously [24] . Protein G sepharose was charged with antibodies for 1 h at room temperature , washed with PBS , incubated with cell lysates overnight at 4°C , followed by washing with PBS . The efficiency of ROP18 precipitation was assessed by probing with rabbit anti-ROP18 that was directly conjugated to NHS-biotin ( Pierce , Thermo Scientific ) , washed in PBS , incubated with streptavidin conjugated to HRP , and detected as described above for western blotting . ROP18-kinase domain ( ROP18-KD ) was expressed and purified as described previously [10] . Genomic DNA from the type I RH strain of T . gondii was used to amplify the genes encoding full length ROP18 ( ROP18-FL ) ( starting from Glu83 based on the second ATG of GenBank protein CAJ27113 ) or ROP5 ( starting from Val25 of GenBank protein AAZ73240 . 1 ) using iProof high-fidelity polymerase ( Bio-Rad ) and primers listed in Table S1 . Amplicons were cloned into pGEX-6P-1 using primers that introduced a C-terminal His6 tag . ROP18-FL was expressed in BL21 ( DE3 ) -V2R-pACYC LamP , as described previously [10] . ROP5-FL was expressed in Rosetta ( DE3 ) pLysS ( Novagen ) . Cells were induced with 1 mM IPTG , grown overnight at 15°C , and soluble proteins were purified using Glutathione Sepharose 4B ( GE Healthcare ) according to the manufacturer's recommendations . Protein purity and concentration were assessed by SDS-PAGE and SYPRO Ruby staining . Kinase activity of recombinant or immunoprecipitated ROP18 was tested on the heterologous substrate dephosphorylated myelin basic protein ( dMBP ) ( Millipore ) ( 0 . 5 µg/reaction ) , or separately on Irgb6 that was immunoprecipitated from IFN-γ activated RAW cells using rabbit anti-Irgb6 . Kinase reactions were conducted in 25 mM Tris–HCl pH 7 . 5 , 15 mM MgCl2 and 2 mM MnCl2 . containing 10 µCi of 32P γ-ATP ( specific activity: 3 , 000 Ci/mmol ) ( Perkin Elmer , Inc ) in addition to 33 µM unlabelled ATP ( Sigma-Aldrich ) . Reactions were allowed to proceed at 30°C for 30 min , samples were heated to 95°C in Laemmli sample buffer , resolved on 12% or 10% SDS–PAGE gels , dried , and imaged using a FLA5000 phosphorimager . Statistical calculations were performed in Excel . Student's t tests were performed under the assumption of equal variance and using a two-tailed test , , where P≤0 . 05 was considered significant . Confluent HFF monolayers grown in T75 flasks were infected with 12×106 wild type ( RHΔku80 ) or ROP5 deficient ( RHΔku80Δrop5 ) parasites ( MOI of 4 ) , or mock infected and allowed to grow for 24 hours . Cells were washed in PBS lacking divalent cations , removed by trypsinization , washed in DMEM containing 10% FBS , pelleted by centrifugation at 400 g for 10 min , and the pellets stored at −80°C . To extract RNA , pellets were thawed and processed with the Qiagen RNeasy kit supplemented with β-mercaptoethanol and DNase I treatment ( Qiagen ) . Total RNA was processed and labeled into cRNA using the Ambion Message Amp Premier ( Ambion ) according to the manufacturer's protocol using 500 ng starting RNA . A total of 10 µg cRNA was hybridized to the HG-U113A_2 Affymetrix Human Genome Array ( Affymetrix ) using standard manufacturer's hybridization and scanning protocols . Data was processed using GeneSpring 7 . 2 ( Agilent Technologies ) with the following normalizations: Robust Multi-array Average ( RMA ) , Data transformation: Set measurements less than 0 . 01 to 0 . 01 , Per Chip and Per Gene: Median Polishing . There was only one gene >2 . 5 fold different between RHΔku80 and RHΔku80Δrop5 infected host samples , a serine protease inhibitor - BC005224 ( Genbank id for the probe set specific for this gene; http://www . ncbi . nlm . nih . gov/nuccore/BC005224 ) and the variance in the RHΔku80Δrop5 data for this gene resulted in the difference being non-significant . Data was submitted to NCBI GEO record GSE32104 .
|
The ability of microorganisms to cause disease in their hosts is often mediated by proteins that are secreted by the pathogen into the host cell as a means of disarming host signaling . Previous studies with the protozoan parasite Toxoplasma gondii have revealed that secretion of parasite protein kinases into the host cell mediates virulence in mouse , a natural host for transmission . Curiously , some of these virulence factors are active protein kinases , while other related pseudokinases lack enzymatic activity; hence , it was unclear how they functioned in promoting virulence . In the present work we demonstrate that ROP5 , an inactive member of this protein kinase family , regulates the active protein kinase ROP18 , which normally prevents clearance of the parasite in interferon-activated macrophages . Allosteric regulation of enzymes is a common theme in biology , but this is the first example of such a mechanism regulating a pathogen virulence factor . The potential advantage of such a layered process is that it might allow greater temporal or spatial control and perhaps protect the parasite from disabling strategies by the host .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"immunology",
"biology",
"microbiology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] |
2012
|
The Polymorphic Pseudokinase ROP5 Controls Virulence in Toxoplasma gondii by Regulating the Active Kinase ROP18
|
High-throughput methods such as EST sequencing , microarrays and deep sequencing have identified large numbers of alternative splicing ( AS ) events , but studies have shown that only a subset of these may be functional . Here we report a sensitive bioinformatics approach that identifies exons with evidence of a strong RNA selection pressure ratio ( RSPR ) —i . e . , evolutionary selection against mutations that change only the mRNA sequence while leaving the protein sequence unchanged—measured across an entire evolutionary family , which greatly amplifies its predictive power . Using the UCSC 28 vertebrate genome alignment , this approach correctly predicted half to three-quarters of AS exons that are known binding targets of the NOVA splicing regulatory factor , and predicted 345 strongly selected alternative splicing events in human , and 262 in mouse . These predictions were strongly validated by several experimental criteria of functional AS such as independent detection of the same AS event in other species , reading frame-preservation , and experimental evidence of tissue-specific regulation: 75% ( 15/20 ) of a sample of high-RSPR exons displayed tissue specific regulation in a panel of ten tissues , vs . only 20% ( 4/20 ) among a sample of low-RSPR exons . These data suggest that RSPR can identify exons with functionally important splicing regulation , and provides biologists with a dataset of over 600 such exons . We present several case studies , including both well-studied examples ( GRIN1 ) and novel examples ( EXOC7 ) . These data also show that RSPR strongly outperforms other approaches such as standard sequence conservation ( which fails to distinguish amino acid selection pressure from RNA selection pressure ) , or pairwise genome comparison ( which lacks adequate statistical power for predicting individual exons ) .
Global analyses of alternative splicing ( AS ) have established its importance in protein diversity and gene regulation in higher eukaryotes [1] , [2] . Alternative splicing can regulate biological function by altering the sequence of protein products and modulating transcript expression levels [3] . Alternative splicing can modify binding properties , intracellular localization , enzymatic activity , protein stability or post-translational modifications[4] . Alternative splicing is often regulated in a tissue-specific manner [5] and can undergo important changes in disease states such as cancer [6] , [7] . All of these illustrate that it is necessary to study the functional effects of alternative splicing to understand the complexity of biological system and human disease . One major challenge is the identification of functional alternative splicing events . In general , experimental methods that can directly demonstrate a biological function for an AS event are time-consuming , ad hoc , and impractical on a genome-wide scale . By contrast , high-throughput methods for surveying the transcriptome , such as EST sequencing [8] , microarrays [9] or deep sequencing [10] , mainly enable detection of whether a given splicing event is “present” or “absent” in a sample . Genome-wide analysis of such datasets has produced large databases of detected alternative splicing events , but with limited guidance for biologists about which ones are likely to be functional . More importantly , a number of studies have shown that a significant fraction of these detected events are probably not functional [2] , [11] , [12] . In this context , biologists need improved ways of distinguishing AS events that are likely to have important biological functions , before initiating costly experiments , such as high-throughput studies of regulation [13] , [14] . There are multiple aspects of function that can be assessed using bioinformatics . Many studies have used the independent observation of the same alternative splicing event in multiple species as evidence that it is functional [15]–[18] . By contrast , introduction of a STOP codon more than 50 nt from the last exon-exon junction is predicted to cause nonsense-mediated decay; such a splice form will not produce a functional protein product ( although the AS event itself might still play an important role in regulating function by down-regulating the transcript level ) [19]– . Mapping of the AS exon to known protein domains or structures has also been suggested as an indicator of useful biological function [23] , [24] . Many studies have indicated that an AS sequence segment consisting of an exact multiple of 3 nt length is more likely to be functional , since it can be alternatively spliced without affecting the protein reading frame [25] , [26] . Evidence of tissue-specific regulation ( from EST or microarray data ) has also been taken as evidence of function [27]–[30] . While all of these criteria have been shown to be useful indicators of “function” , it should be emphasized that no one method can adequately capture this ill-defined concept , precisely because it has many different aspects . Functions that are important for reproductive success are subject to selection pressure , which can be defined as a reduction in the frequency of observed mutations relative to that expected under a neutral model . For example , sequence conservation both within alternatively spliced exons and in flanking introns has been cited as evidence of important regulatory motifs . One useful extension of this principle is to separate total conservation into non-synonymous sites ( i . e . where mutations will change the amino acid sequence ) vs . synonymous sites ( where mutations leave the amino acid sequence unchanged ) . Whereas alternative exons show poorer conservation than constitutive exons by total conservation metrics like phastCons [31] , several studies have reported that measures of synonymous mutations ( Ks , or ds ) drop dramatically in certain types of alternative exons [26] , [32] , [33] , particularly those that show tissue-specific regulation . Unlike standard conservation , such Ks effects cannot be attributed to protein function , and have thus been used as a measure of “RNA selection pressure” for features such as splicing factor binding sites , RNA secondary structure etc . [26] , [34]–[40] . With the rapid growth in complete genome sequences for animals and plants [41] , the strategy of seeking to detect RNA selection pressure gains increasing power as a way of predicting strongly selected AS regions [42] , [43] . Past applications of Ks to this problem typically relied upon comparing a single pair of related genomes ( e . g . human vs . mouse ) . Given the high level of identity seen in such exon comparisons ( around 87% for human vs . mouse ) , the number of synonymous mutations expected in an alternative exon ( just based on size , with no RNA selection pressure ) is low , perhaps ten or fewer . Even if the observed number of synonymous mutations were three-fold lower , i . e . three or fewer , implying strong RNA selection pressure , the result would not be statistically significant , due to the small number of counts being compared . For this reason , such studies have typically not tried to predict which individual AS exons are strongly selected , but instead to compare entire groups of exons , e . g . all “minor-form exons” vs . all constitutive exons . However , large multigenome alignments such as the UCSC 28 vertebrate genome alignment could improve predictive power , by measuring Ks simultaneously in many separate species . This has two benefits . First , the dataset of mutation counts for any given exon is greatly increased ( naively , by a factor of up to 20-fold compared with a single pair of species ) , increasing the statistical power for detecting real selection pressure cases . Second , the ability to discriminate spurious cases is enhanced by utilizing a much more diverse set of genomes: various types of artifacts that might occur in one genome ( e . g . a mutation “cold-spot” in mouse evolution ) are unlikely to be conserved over 28 genomes spanning 300 million years of vertebrate evolution ( such conservation would in fact indicate consistently strong selection ) . In this paper we present a robust method for applying this approach , combining multigenome alignment data and RNA selection pressure calculations . This approach enables the detection of statistically significant RNA selection pressure for each individual exon , providing a direct prediction of whether that AS event has been strongly selected during vertebrate evolution . We have tested these prediction using a wide variety of data including known sets of regulatory targets , large-scale EST and microarray data , and RT-PCR analysis of tissue-specific splicing regulation .
As an initial test of the RNA selection pressure ratio ( RSPR ) metric , we applied it to GRIN1 , a gene whose alternative splicing is well understood [44] . Of the 21 coding region exons , two exons ( exons 5 and 21 ) show evidence of dramatically increased RSPR ( RSPR values of 7 and 10 respectively ) compared with the remaining exons ( Figure 1 ) . These two exons correspond to the well-studied N1 and C1 alternative exons , which have been shown to be regulated by PTB , NOVA2 , hnRNP H and hnRNP A1 [45] , and in turn control receptor desensitization [46] and NMDA receptor interactions [47] . The RSPR data indicate that these two exons have synonymous mutation rate approximately ten-fold lower than the surrounding exons ( p-value = 7 . 7×10−25 ) . This very strong signal suggests that it should be possible to detect regulated alternative exons using the RSPR metric . These data suggest that more than half of the synonymous sites in the N1 and C1 exons are under some kind of negative selection pressure . This seems compatible with existing splicing factor motif databases . RESCUE-ESE [48] and FAS-ESS [49] predicted 43% of the N1 exon sequence and 50% of the C1 exon sequence as splicing factor binding sites ( Figure 1D ) . The GRIN1 data provide evidence that using multigenome alignments ( in this case GRIN1 sequences from 16 species ) can make RSPR more sensitive than simply comparing a pair of genomes . Figure 1C shows the synonymous substitution data ( Ks ) used to compute RSPR , annotated on the lineages leading to each species . The Ks values for the GRIN1 N1 and C1 alternative exons are about ten-fold lower than the Ks values for the other exons , on each lineage . For example , within the primate lineage , these alternative exons showed a Ks level of 0 . 027 vs . 0 . 26 in the neighboring exons . Independent data for other mammal lineages ( mouse , rat , dog ) show a similar pattern , with total Ks values of 0 . 05 vs . 0 . 70 respectively . Moreover , frog , and zebrafish also show big decreases in Ks for the two alternative exons . These data indicate that the GRIN1 N1 and C1 exons have been under a consistently strong negative RNA selection pressure for over 450 My . They also show why this computational approach can be more sensitive than simply comparing a pair of species , since each additional genome contributes further statistical evidence for the significance of this pattern . Next , to estimate the sensitivity of RSPR and compare it with existing methods , we applied it to a test set of alternative exons that are known targets of NOVA , generated by Jelen et al [50] . Using a conservative criterion ( RSPR>3 , P_RSPR<0 . 001 ) , RSPR detected 55% ( 25/45 ) to 73% ( 11/15 ) of these alternative exons as having strong RNA selection pressure ( Figure 2; Table S1 ) , depending on whether RSPR was computed for the target exon vs . all other exons in the gene , or vs . constitutive exons in the gene . An additional 24% ( 11/45 ) showed evidence of substantial selection pressure ( RSPR>1 . 4 ) and had significant p-values , but fell below our cutoff of RSPR>3 . To assess the specificity of RSPR , we also tested it on a dataset of alternative exons not known to be regulated . In the absence of a gold standard dataset of “true negative” exons that are known not to be targets of splicing regulation , we used a test set of 295 mouse alternative exons whose inclusion level remained relatively constant across a set of 10 different mouse tissues [51] . We detected RSPR>3 in only 9 . 5% ( 28/295 ) of these exons . This suggests that RSPR is relatively specific to exons that are targets of regulation . We also tested two standard existing methods: pairwise genome comparison [52]; and standard measures of sequence conservation computed over the same multigenome alignments as the RSPR . RSPR differs from previous methods in combining two distinct approaches: first , instead of calculating simple sequence conservation , it separates evidence of nucleotide selection pressure from amino acid selection pressure ( and ignores the latter ) ; second , it measures this pressure across an entire family of aligned genomes . Ordinarily , studies of RNA selection pressure have compared a pair of genomes ( e . g . human vs . mouse ) , but the pairwise method detected only 4% of the NOVA targets ( Figure 2 ) . Furthermore , even the successful pairwise predictions had much weaker p-values than from the multigenome RSPR calculation , by about 4 to 32 orders of magnitude . Secondly , we tested two standard measures of sequence conservation that are computed from all aligned genomes: baseml [53] , and phastCons [54] . Baseml is calculated almost identically as RSPR , using the same software package ( PAML ) , but unlike RSPR does not specifically estimate RNA selection pressure . With optimized signal-to-noise cutoffs , standard sequence conservation ( baseml ) detected 38% ( 17/45 ) of the NOVA targets , and phastCons detected 40% ( 18/45 ) of the NOVA targets . We next sought to compare RSPR's performance on a much larger dataset ( Table 1 ) . Numerous studies have reported that an alternative splicing event can be characterized as functional if it is independently observed in expression data from divergent species such as human and mouse [15] , [16] , [18] . We therefore tested each human exon in our dataset to see whether the orthologous mouse exon is observed to be alternative spliced in mouse EST data , and graphed the results as a function of the RSPR value ( Figure 3A ) . For values of RSPR<1 , only a small fraction of exons ( approximately 10% ) were also observed to be alternatively spliced in mouse , but the rate increased rapidly to 60–70% for values of RSPR of 3 or higher . These data suggest again that RSPR can help distinguish AS events that are likely to be functional , and that a large fraction of the exons with RSPR>3 are actually functional in both human and mouse . Furthermore , this pattern of validation can be seen not only in mouse EST data , but holds true in independent EST data from all of the vertebrate species in our study ( Figure 3A ) . Although the total percentage of validation depends on the level of EST coverage for each organism ( highest in mouse; lowest in zebrafish ) , the pattern is consistent in every EST dataset from mammals , chicken , frog , or zebrafish: high RSPR values predict which exons' alternative splicing will be conserved in other species . These results are consistent with the analysis by Plass and Eyras [39] . Throughout all species , an RSPR value of 3 appears to be a threshold for predicting strongly selected alternative splicing events . We have also used the AS validation rate in mouse to estimate the true positive vs . false positive rates for RSPR ( Figure 3B ) . We classified exons that were independently observed to be alternatively spliced in mouse EST data as “true positives” ( validated ) , and exons that were not observed to be alternatively spliced in the mouse EST data as “false positives” ( not validated ) . We graphed the rate of true positives vs . false positives both as a function of increasing RSPR ( Figure 3B ) . These data show that RSPR outperforms standard measures of conservation ( e . g . 60% true positive rate for RSPR vs . 50% true positive rate for baseml vs . 40% true positive rate for phastCons , at the 20% false positive point ) over the whole range of error rates . It should be emphasized that the actual false positive rate is likely to be significantly less , because mouse EST coverage is poor , resulting in many AS events that fail to be observed simply due to insufficient EST sampling . To further assess the robustness of RSPR predictions , we subdivided the exons based on whether they showed standard conservation or not ( measured using baseml ) , and evaluated the predictive power of RSPR in both cases ( Figure 4A ) . RSPR showed a high level of predictive power both for exons with strong standard conservation ( baseml mutation ratio>3 ) and weak conservation ( mutation ratio<3 ) . Thus RSPR is robust in predicting functional alternative splicing ( regardless of the exact baseml value; Figure 4B ) . We have also found that it is possible to combine RSPR ( measured from the exon sequence ) with conservation measured in the flanking introns ( Figure 4C ) . We computed the intron mutation ratio in the 50 nt on each side , using baseml . Exons with low RSPR showed flanking intron mutation ratios around one ( i . e . neutral selection ) . By contrast , exons with high RSPR ( RSPR>3 ) nearly all display higher levels of conservation in their flanking introns . This difference was statistically significant ( p-value <10−300 ) . On average , the amount of RNA selection pressure measured within the exon ( RSPR ) is matched by a proportional amount of selection pressure in the flanking 50 nt ( intron mutation ratio ) . However , the intron mutation ratio displays a large variance ( as shown by the error bars ) . These data suggest that exon-based RSPR can be complemented by the independent signal supplied by intron conservation , to improve prediction of strongly selected AS . We also analyzed the effect of unusually small trees ( only 5–9 species; Figure S1B ) and of trees containing only placental mammal species ( Figure S1C ) compared with trees spanning both placental mammals and more distantly related vertebrates ( Figure S1D ) . These data showed that the RSPR calculated using only 5–9 species successfully predicts whether the exon will be observed to be alternatively spliced in another species ( in this case , mouse ) , just as we previously showed for our complete dataset ( Figure S1B ) . Similarly , the RSPR calculations that only used placental mammal species showed the same direct correlation between RSPR and the AS validation rate . It should be noted that these are both unusual cases in our dataset: 86% of the RSPR calculations used 10 or more species , and 78% of the RSPR calculations spanned both placental mammals and more distant vertebrate species ( Figure S1A ) . Finally , it should be emphasized that the highest RSPR values ( those significantly above RSPR = 3 ) come almost exclusively from the cases with broad species representation , as one can see by comparing the RSPR range for the Placental Mammals set ( Figure S1C ) vs . the set that spans both Placental Mammals + Other Vertebrates ( Figure S1D ) . We calculated RSPR for a set of 4626 alternative exons in human and 1935 exons in mouse , and applied a threshold of RSPR>3 and P_RSPR<0 . 001 , yielding 345 exons in human ( vs . 4 . 5 expected by random chance ) and 262 exons in mouse ( Table 1 and Table S2 ) . Some of the alternative splicing events detected by RSPR have already been demonstrated to be functional by experimental studies , such as LRP8 [55] , SYK [56] , DGKH [57] and WT1 [58] ( Table S3 ) . For example , the alternative exon in LRP8 ( ApoER2 ) encodes a 13 amino acid insertion containing a furin cleavage site important for LRP8 function . SYK alternative splicing appears to play an important role in cancer; the exon-skip form ( Syk ( S ) , deleting 23 amino acid residues ) occurs frequently in primary tumors but never in matched normal mammary tissues . DGKH has been shown to produce two splice forms , DGKH1 and DGKH2 which differ in the inclusion of a serile alpha motif ( SAM ) domain [57] . DGKH2 was detected only in testis , kidney and colon while DGKH1 was ubiquitously distributed in various tissues . WT1 is a transcriptional regulator with an alternative exon encoding a 17 amino acids insertion , which appears to play a role in regulating cell survival and proliferation . First , we performed RT-PCR experiments to test a subset of the RSPR predictions ( Figure 5; Table S4 ) . Tissue-specific splicing is one type of functional regulation that can be assayed easily in a panel of different tissue samples . We therefore extracted a random sample of 20 high-RSPR exons and 20 low-RSPR exons , and assayed their splicing by RT-PCR in cerebellum , heart , kidney and seven other human tissues . We classified an alternative exon as tissue specific if it was observed to be expressed as the major-form in at least one tissue , and as a minor-form in at least one other tissue . 75% ( 15/20 ) of the high-RSPR exons were found to be regulated in a tissue specific manner in this panel , vs . only 20% ( 4/20 ) among the low-RSPR exons ( p-value 6 . 1×10−4 ) . Thus a large fraction of high-RSPR exons were validated by experimental evidence of regulation . Low RSPR may indicate a lower probability of functional AS regulation; nearly half of the low-RSPR exons ( 8/20 ) did not even show evidence of alternative splicing in this panel ( whereas the expected AS pattern was detected in this panel in 100% of high-RSPR exons ( 20/20 ) . We tested all the RSPR predictions using large-scale EST and microarray data . If RNA selection pressure is associated with exons whose splicing is tightly regulated ( e . g . tissue-specific ) , then high-RSPR exons should not be expressed ubiquitously , but only in a subset of tissues or cells , and thus would be classified in EST datasets as minor-form ( included in less than a third of a gene's transcripts ) rather than major-form ( included in more than 2/3 of transcripts ) . To test this hypothesis , we compared the ratio of major-form vs . minor-form exons in low-RSPR vs . high-RSPR groups ( Table 1 ) . Whereas the minor/major ratio was 0 . 0795 in the low-RSPR set , it rose to 1 . 79 in the high-RSPR set , more than a 20-fold increase . The ratio of intermediate-form over major-form also increased , from 0 . 158 in the low-RSPR set , to 1 . 27 in the high-RSPR set . Overall , only 24 . 6% of the high-RSPR exons were classified as major-form , vs . 74 . 7% in the low-RSPR set . A large fraction of exons in the high-RSPR dataset ( 44% ) are annotated to be minor form . Thus , high levels of RSPR appear to be associated with exons that are included in a non-ubiquitous fashion , i . e . in a minority of a gene's transcripts . To assess whether some of this pattern can be traced to tissue-specific splicing , we compared our RSPR results with a mouse microarray study that measured the inclusion level of mouse exons in 10 tissues [51] . Exons with a tissue-switched splicing pattern ( i . e . minor-form in some tissue ( s ) , but major-form in other tissue ( s ) [59] were three times as abundant in the high-RSPR set ( 28 . 2% , 11/39 ) than in the low-RSPR set ( 8 . 6% , 25/292 ) . This difference was statistically significant , p-value = 0 . 0011 ( Fisher exact test ) . We also saw a significant enrichment of tissue-specific exons detected by a previous EST analysis [5] . In the high-RSPR set , 9 . 3% ( 30/324 ) had strong evidence of tissue-specificity in the EST data ( LOD>3 ) , approximately double that observed in the low-RSPR set ( 5 . 2% , 204/3958 ) . The p-value for this difference was 0 . 0026 . To verify the EST analyses of tissue-specificity , we tested a sample of 10 high-RSPR exons with putative brain-specific splicing in the EST data , using RT-PCR in ten human tissues . 80% ( 8/10 ) showed tissue-specific splicing in this experiment ( Table S5 ) . Numerous studies have examined the evolutionary patterns associated with different AS inclusion levels , such as minor-form , major-form , and intermediate form [60]–[63] . We have analyzed the RSPR distributions of these different forms of alternative splicing . Exons with different inclusion levels show strikingly different distributions of RNA selection pressure as measured by RSPR ( Figure 6A ) . Major-form exons form a tight , symmetric distribution centered on RSPR = 1 ( neutral selection ) . By contrast , minor-form exons display a broader distribution with a peak at RSPR = 3 . At least within the constraints of this study ( which was limited to exons that have been retained in mammalian genomes long enough to be found in multiple species , so that we can measure an RSPR value ) , 36% of conserved minor-form exons display strong RNA selection pressure , whereas only 2 . 3% of major-form exons had strong RSPR , which is in agreement with Xing et . al [26] . Intriguingly , intermediate-form exons revealed a bimodal distribution , with a main peak that closely follows the profile of the major-form distribution , and a small peak extending above RSPR = 3 . Validation by mouse EST data , and by frame-preservation , also highlight interesting differences between major- vs . minor-form exons . First , it is striking that high values of RSPR are predictive of functional AS ( as measured by observation of alternative splicing of the orthologous mouse exon , in mouse EST data ) , at all three inclusion levels , even in major-form exons ( Figure 6B ) . Thus , it appears that a small fraction of major-form exons are under RNA selection pressure for maintaining an important alternative splicing function . However , the total level of AS validation ( by mouse EST data ) was approximately two-fold higher for minor-form exons ( over 80% for RSPR>3 , vs . about 40% for major-form exons ) . Second , RSPR was predictive of functional AS at much lower values of RSPR for minor-form exons than for major-form exons . Third , frame-preservation revealed another interesting difference ( Figure 6C ) . Whereas both minor- and intermediate-form exons displayed strong increases in frame-preservation with increasing RSPR , major-form exons remained at background frame-preservation levels ( around 40% ) across the whole range of RSPR values . This implies that RNA selection pressure is associated with “modular” protein sequence insertions for minor- and intermediate-form exons ( i . e . insertions that do not alter the reading frame of the downstream protein sequence ) , but not for major-form exons . Our RSPR results predict hundreds of alternative exons as strongly selected alternative splicing events . As one example , EXOC7 is a component of the exocyst , an evolutionarily conserved octameric protein complex essential for exocytosis . The protein structure of mouse Exoc7 [64] , published recently , includes 19 α-helices linked with loops . EST data reveal several alternative exons , including two groups that map to disordered regions in the protein structure , one between helix 6 and helix 7 , and the other between helix 12 and helix 13 ( see Table S6 ) . The latter ( exon_id 25261 in ASAP II ) contains 39 nucleotides , and is alternatively spliced within independent EST data not only for human , but also for mouse , dog , cow and frog ( Figure 7 ) . Our RSPR analysis detected this exon as having strong RNA selection pressure: RSPR = 6 . 24 , measured over Human , Chimpanzee , Rhesus , Rat , Mouse , Hedgehog , Dog , Cat , Horse , Cow , Opossum , Platypus , Chicken , Lizard , Frog , Tetraodon , Fugu , Medaka and Zebrafish , with a P_RSPR of 1 . 67×10−10 . ASAP2 reports this splicing event as tissue-specific to brain_nerve ( with LOD 2 . 5 ) and retina ( LOD 2 . 2 ) [5] , and RT-PCR experiments confirmed its brain-specific splicing pattern ( Figure 7D ) . NetPhos analysis [65] indicates a set of phosphorylation sites nearby ( See Figure 7B ) . FAS-ESS ( http://genes . mit . edu/fas-ess/ ) [49] identifies four ESS motifs in this exon ( Figure 7E ) . Although the function of this exon is unknown , all these clues suggest that its alternative splicing has played an important functional role over a very long period of vertebrate evolution .
We have presented an effective method for estimating RNA selection pressure within an individual exon , and have tested its predictions against a variety of empirical measures of functional alternative splicing , such as known NOVA-regulated exons , conserved alternative splicing , frame preservation , and tissue-specific splicing patterns . We have also predicted a large dataset of strongly selected AS exons that can be useful targets for biologists to study the regulation of alternative splicing . Not only can the high-RSPR dataset furnish biologists with new insights into well-studied genes , but also identifies many new targets worthy of experimental study , in the form of strongly selected alternative splicing events . These data suggest several possible benefits of RSPR . It provides a general way for distinguishing selection pressures that operate at the nucleotide level rather than protein level . RNA selection pressure may reflect many possible functional mechanisms , such as binding sites of splicing regulators including exon splicing enhancers ( ESE ) and exon splicing silencers ( ESS ) [48] , [49] . As an example , ESE/ESS analyses for GRIN1 and EXOC7 annotated slightly under one-half of sites in these high-RSPR exons as ESEs or ESSs . Finally , RSPR integrates several powerful tools in comparative genomics , such as MULTIZ multiple genome alignments and PAML evolutionary model inference , and can in principle be applied to any genome . Previous studies have reported that minor-form exons were associated with increased values of Ka/Ks and Ka compared with neighboring constitutive exons [26] , [35]–[37] ( for a review see [42] ) . Our results are consistent with this pattern; the distributions of Ka/Ks and Ka for minor-form exons showed significant increases relative to their control regions ( neighboring consitutive exons; Figure S2 AB ) . Moreover , this pattern was also observed for minor-form exons with high RSPR values ( Figure S2 C ) . Our approach has important limitations . First , it is important to emphasize that it seeks to detect RNA selection pressure , but gives no suggestion of what specific functional mechanism might cause it . Other possible patterns of selection that might be detected by RSPR include RNA secondary structure present within the pre-mRNA [42] , miRNA binding sites , or binding sites in the parent DNA sequence . While some RSPR may be associated with ESE and ESS motifs , we cannot assume that they fully explain the RSPR in alternative exons . Consistent with several previous studies [35] , [66] ( see [43] for a review ) , we did not observe an overall correlation between the RESCUE-ESE density and RSPR , or between FAS-ESS density and RSPR in alternative exons ( data not shown ) . Second , in this paper we have focused on demonstrating the predictive value of calculating RSPR within exonic sequence , without taking into account other useful information such as the flanking intron sequence , frame preservation , expression data from multiple species etc . An integrated prediction method would presumably make use of all available information [18] . A naive initial approach , based on simply multiplying the p-values from baseml ( for the flanking intron conservation ) and from codeml ( for the exon RSPR conservation ) did not appear to give major improvements of prediction accuracy in our preliminary tests , compared with simply using the codeml p-value . Since the two calculations use different programs ( baseml vs . codeml ) and different mutation models ( single-nucleotide based vs . codon based ) , combining them in a single integrated calculation is not trivial . Third , RSPR is calculated based on the multiple genome alignment , and thus requires that an exon be sufficiently conserved among several genomes to be aligned .
We obtained data for alternative exons ( exon skipping ) and constitutive exons in the same gene , from the ASAP II database [67] . Based on EST data , ASAPII classified each alternative exon as Major form ( exon inclusion level greater than 2/3 ) , Minor-form ( exon inclusion level less than 1/3 ) and Intermediate-form ( exon inclusion level between 1/3 and 2/3 ) . We defined a constitutive exon as an exon that is included in all transcript isoforms of the gene ( inclusion level 100% ) . We used several additional kinds of data to validate the prediction of functional alternative splicing , such as conserved alternative splicing based on independent EST data from multiple species assembled in the ASAPII database [67] , tissue-switched alternative exons identified in the mouse microarray data of Pan et al . [51] , [59] . We obtained all genome alignment information used in this study from the UCSC 28 vertebrate genome alignment hg18_multiz28way [68] , available from ftp://hgdownload . cse . ucsc . edu/ . This alignment includes the following complete genomes: Human , Armadillo , Bushbaby , Cat , Chicken , Chimpanzee , Cow , Dog , Elephant , Frog , Fugu , Guinea Pig , Hedgehog , Horse , Lizard , Medaka , Mouse , Opossum , Platypus , Rabbit , Rat , Rhesus , Shrew , Stickleback , Tenrec , Tetraodon , Tree shrew and Zebrafish . We used the list of NOVA target exons of Jelen et al . [50] as a validation testset for our RSPR predictions . We were able to obtain genomic coordinates and genome alignments for 45 of NOVA targets published in Jelen et al . , which were used for the validation tests presented in Figure 1D and Table S1 . We calculated the RNA Selection Pressure Ratio ( RSPR ) for each alternative exon compared with the constitutive exons within the same transcript isoform . Here we briefly summarize each step ( See Figure S3 ) : 1 ) each exon was mapped to orthologous exons in the 28 aligned genomes using the NLMSA alignment query tool [69] in the Pygr software package ( http://code . google . com/p/pygr/ ) . 2 ) each orthologous exon was required to retain the aligned splice sites and maintain a minimum of 70% amino acid identity ( calculated by needle in EMBOSS [70] . For each exon , a minimum of 5 species was required . We ranked constitutive exons in a given gene by the number of species with orthologous exons , and identified the top third ( i . e . most widely conserved exons ) , or a minimum of four constitutive exons , to represent that gene . 3 ) We then found the subset of species that were each aligned to all of these exons as well as to the alternative exon . That is , we computed the intersection of the sets of species that are aligned to each of these exons . This yielded the subset of species that we used for the RSPR calculation . 4 ) Next , we generated the list of constitutive exons that were aligned to all of these species , and used these as the control region for the RSPR calculation . 5 ) We extracted the alignment consisting of just this subset of species , for the control region + the alternative exon . 6 ) Prior to calculating RSPR , gaps were removed from the alignment . Specifically , only codons that were present in each of the species in the subset were retained in the alignment: amino acid sequences for the orthologous exons were aligned using clustalw [71] , columns containing gaps were removed . This procedure ensures that RSPR is calculated using the exact same tree of species for the control region as for the alternative exon . We used the multi-partition model D [72] of the PAML program codeml ( http://abacus . gene . ucl . ac . uk/software/paml . html ) to calculate maximum likelihood estimates of the RSPR , and a p-value for the null hypothesis of neutral RNA selection ( i . e . RSPR = 1 ) . Codeml uses a codon substitution model that is similar to the HKY85 nucleotide substitution model . Although codeml does not directly compute RSPR , its output parameters can be used to calculate RSPR . The gap-trimmed nucleotide sequences and a tree file including the phylogenetic tree for the subset of aligned species[68] were submitted to the codeml program , which estimates a set of evolutionary parameters from the whole tree , by maximum likelihood . We defined the RSPR as the ratio of synonymous mutation rate Ks for the alternative exon vs . the constitutive exons among all branches in the phylogenetic tree: ( 1 ) where subscript 0 indicates the constitutive exons , and subscript 1 indicates the alternative exon . Based on this definition , a high RSPR value implies the alternative exon is under stronger negative RNA selection pressure than the corresponding constitutive exons . Ordinarily , for each branch in the tree , codeml estimates the total branch length t [73] , which is related to the synonymous and non-synonymous substitution densities Ks and Ka via the ratio ( 2 ) where S and N are the number of synonymous and non-synonymous sites respectively . In multiple partition mode , tb for each branch b is replaced by t0b ( for the constitutive exons ) and t1b ( for the alternative exon ) , related by the partition ratio ( 3 ) which has a single value over the entire tree ( i . e . for every branch b , t0b is constrained to be equal to t0b = r t1b ) . Based on the input phylogenetic tree and sequence alignment , codeml simultaneously estimates t1b ( for each branch ) , r ( for the whole tree ) , as well as the πj , κ0 , κ1 , and ω0 , ω1 values ( for the whole tree ) . Finally , combining equations 1 , 2 , and 3 , we computed the RNA selection pressure ratio ρ: ( 4 ) where S0 , S1 and N0 , N1 are the number of synonymous and non-synonymous sites in the constitutive exons vs . the alternative exon , respectively . We also modified codeml to be able to compute the likelihood under the constraint ρ = 1 . We computed the p-value P_RSPR for the null hypothesis RSPR = 1 based on the log-odds ratio 2log ( L ( ρML ) /L ( ρ = 1 ) ) , which follows a χ2 distribution with one degree of freedom , where ρML is the original maximum likelihood estimate of ρ obtained above . We used two different standard methods for computing sequence conservation , baseml [53] and phastCons [54] . The baseml calculation used the HKY85 nucleotide substitution model ( model = 4 ) and the Mgene = 3 multiple partition mode , similar to our codeml calculation . RSPR and baseml are calculated almost identically using the PAML package; the only difference is that whereas RSPR is calculated from the Ks ratio ( eq . 1 , above ) , for baseml we simply used the total nucleotide substitution ratio r ( eq . 3 , above ) . We calculated this ratio both for the alternative exon , and its flanking intronic regions ( 50 nt flanking each exon on each side ) . PhastCons [54] is a widely used method for measuring sequence conservation in multiple genome alignments , used for example in the UCSC genome browser . We used phastCons to compute the ratio η for a region of interest compared with a control region ( analogous to the RSPR ) , as follows: ( 5 ) where is the average probability of the phastCons non-conserved state in the control region , vs . in the region of interest , and is the average phastCons score in the control region , vs . in the region of interest . We applied this both to exons ( constitutive exons vs . alternative exon ) , and their flanking introns ( 50 nt flanking each exon on each side ) . We also calculated a P value based on the phastCons score for the null hypothesis that the mutation density is equal in the control region vs . region of interest . Specifically , we performed the Wilcoxon rank sum test on the phastCons scores for each nucleotide from the control region , vs . for each nucleotide from the region of interest . For performing the NOVA analysis ( Figure 2 ) , we first determined cutoffs for the baseml ratio and phastCons ratio that yielded the same false positive rate in our ROC analysis ( Figure 3B ) as our RSPR cutoff ( RSPR = 3 ) : baseml ratio = 2 . 9; phastCons ratio = 45 . Thus the NOVA analysis compares the sensitivity of these different methods , when calibrated to the same level of specificity . We generated a random sample of 20 high-RSPR exons ( RSPR>3 and p<0 . 001 ) , and a random sample of 20 low-RSPR exons ( RSPR<1 . 0 and p<0 . 001 ) . We then designed primers and performed RT-PCR as described below . As a separate test to confirm putative brain-specific splicing identified from EST data , we performed a join of the high-RSPR exon set and a previous database of EST evidence of brain-specific splicing [5] . We selected ten exons from this group , and performed RT-PCR as described below . Total RNA samples from 10 human tissues were purchased from Clontech ( Mountain View , CA ) . Single-pass cDNA was synthesized using High-Capacity cDNA Reverse Transcription Kit ( Applied Biosystems , Foster City , CA ) according to manufacturer's instructions . For each tested exon , we designed a pair of forward and reverse PCR primers at flanking constitutive exons using PRIMER3 . Two µg of total RNA were used for each 20 ul cDNA synthesis reaction . For each exon , 15 ng total RNA equivalent of cDNA were used for the amplification in a 10 µl PCR reaction . For each exon tested , three DNA polymerase systems were used to optimize RT-PCR reaction: Herculase® II Fusion DNA Polymerase ( Stratagene , La Jolla , CA ) , HotStarTaq DNA Polymerase ( Qiagen , Valencia CA ) and Phire® Hot Start DNA Polymerase ( NEB , Ipswich , MA ) . PCR reactions were run between 25 to 35 cycles ( optimized for each exon ) in a Bio-Rad thermocycler with an annealing temperature of 62 to 66°C ( optimized for each exon ) . The reaction products were resolved on 2% TAE/agarose gels or 5% TBE polyacrylamide gels . Each result was a representation of 3–6 RT-PCR replications . DNA fragments with ambiguous sizes were cloned for sequencing using Zero Blunt® TOPO® PCR Cloning Kit ( Invitrogen , Carlsbad , CA ) . Gel images ( Table S4 and Table S5 ) were visually assessed as tissue specific if the alternative exon's splicing fraction ( percentage of the exon-inclusion isoform vs . the exon-skip isoform ) changed by a factor of two or greater in different tissues .
|
Alternative splicing is an important mechanism for regulating gene function in complex organisms , and has been shown to play a key role in human diseases such as cancer . Recently , high-throughput technologies have been used in an effort to detect alternative splicing events throughout the human genome . However , validating the results of these automated detection methods , and showing that the minor splice forms they detected play an important role in regulating biological functions , have traditionally required time-consuming experiments . In this study we show that such regulatory functions can very often be detected by a distinctive pattern of strong selection on RNA sequence motifs within the alternatively spliced region . We have measured this “RNA selection pressure ratio” ( RSPR ) across 28 animal species representing 400 million years of evolution , and show that this metric successfully predicts known patterns of alternative splicing , and also have validated its predictions experimentally . For example , whereas high-RSPR alternative splices were found experimentally to undergo tissue-specific regulation in 75% of cases , only 20% of low-RSPR cases were found to be tissue-specific . Using RSPR , we have predicted over 600 human and mouse alternative splicing events that appear to be under strong selection . These data should be valuable for biologists seeking to understand the functional effects and underlying mechanisms of splicing regulation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"biology/alternative",
"splicing",
"molecular",
"biology/rna",
"splicing",
"genetics",
"and",
"genomics/comparative",
"genomics",
"computational",
"biology/genomics"
] |
2009
|
Predicting Functional Alternative Splicing by Measuring RNA Selection Pressure from Multigenome Alignments
|
Many genome-wide datasets are routinely generated to study different aspects of biological systems , but integrating them to obtain a coherent view of the underlying biology remains a challenge . We propose simultaneous clustering of multiple networks as a framework to integrate large-scale datasets on the interactions among and activities of cellular components . Specifically , we develop an algorithm JointCluster that finds sets of genes that cluster well in multiple networks of interest , such as coexpression networks summarizing correlations among the expression profiles of genes and physical networks describing protein-protein and protein-DNA interactions among genes or gene-products . Our algorithm provides an efficient solution to a well-defined problem of jointly clustering networks , using techniques that permit certain theoretical guarantees on the quality of the detected clustering relative to the optimal clustering . These guarantees coupled with an effective scaling heuristic and the flexibility to handle multiple heterogeneous networks make our method JointCluster an advance over earlier approaches . Simulation results showed JointCluster to be more robust than alternate methods in recovering clusters implanted in networks with high false positive rates . In systematic evaluation of JointCluster and some earlier approaches for combined analysis of the yeast physical network and two gene expression datasets under glucose and ethanol growth conditions , JointCluster discovers clusters that are more consistently enriched for various reference classes capturing different aspects of yeast biology or yield better coverage of the analysed genes . These robust clusters , which are supported across multiple genomic datasets and diverse reference classes , agree with known biology of yeast under these growth conditions , elucidate the genetic control of coordinated transcription , and enable functional predictions for a number of uncharacterized genes .
Heterogeneous genome-wide datasets provide different views of the biology of a cell , and their rapid accumulation demands integrative approaches that exploit the diversity of views . For instance , data on physical interactions such as interactions between two proteins ( protein-protein ) , or regulatory interactions between a protein and a gene via binding to upstream regions of the gene ( protein-DNA ) inform how various molecules within a cell interact with each other to maintain and regulate the processes of a living cell . On the other hand , data on the abundances or expression of molecules such as proteins or transcripts of genes provide a snapshot of the state of a cell under a particular condition . These two data sources on physical interaction and molecular abundance provide complementary views , as the former captures the wiring diagram or static logic of the cell , and the latter the state of the cell at a timepoint in a condition-dependent , dynamic execution of this logic [1] . Researchers have fruitfully exploited this complementarity by studying the topological patterns of physical interaction among genes with expression profiles that are condition-specific [2] , periodic [3] , or correlated [4]; and similarity of the expression profiles of genes with regulatory , physical , or metabolic interactions among them [5] . Another line of research focuses on integrating the physical and expression datasets to chart out clusters or modules of genes involved in a specific cellular pathway . Methods were developed to search for physically interacting genes that have condition-specific expression ( i . e . , differential expression when comparing two or more conditions , as in “active subnetworks” [6] ) , or correlated expression ( eg . subnetworks in the network of physical interactions that are coherently expressed in a given expression dataset [7]–[9] ) . A challenge in expanding the scope of this research is to enable a flexible integration of any number of heterogeneous networks . The heterogeneity in the connectivity structures or edge density of networks could arise from the different data sources used to construct the networks . For instance , a network of coexpression relations between gene pairs is typically built using expression data of a population of samples ( extracted from genetically varying individuals , or individuals subject to varying conditions/treatments ) . Whereas a network of physical interactions between protein or gene pairs is typically built by testing each interaction in a specific individual or in-vitro condition . Towards addressing this challenge , we propose an efficient solution to a well-defined computational framework for combined analysis of multiple networks , each describing pairwise interactions or coexpression relationships among genes . The problem is to find common clusters of genes supported by all of the networks of interest , using quality measures that are normalized and comparable across heterogeneous networks . Our algorithm solves this problem using techniques that permit certain theoretical guarantees ( approximation guarantees ) on the quality of the output clustering relative to the optimal clustering . That is , we prove these guarantees to show that the clustering found by the algorithm on any set of networks reasonably approximates the optimal clustering , finding which is computationally intractable for large networks . Our approach is hence an advance over earlier approaches that either overlap clusters arising from separate clustering of each graph , or use the clustering structure of one arbitrarily chosen reference graph to explore the preserved clusters in other graphs ( see references in survey [10] ) . JointCluster , an implementation of our algorithm , is more robust than the earlier approaches in recovering clusters implanted in simulated networks with high false positive rates . JointCluster enables integration of multiple expression datasets with one or more physical networks , and hence more flexible than other approaches that integrate a single coexpression or similarity network with a physical network [7]–[9] , or multiple , possibly cross-species , expression datasets without a physical network [11]–[13] . JointCluster seeks clusters preserved in multiple networks so that the genes in such a cluster are more likely to participate in the same biological process . We find such coherent clusters by simultaneously clustering the expression data of several yeast segregants in two growth conditions [14] with a physical network of protein-protein and protein-DNA interactions . In systematic evaluation of clusters detected by different methods , JointCluster shows more consistent enrichment across reference classes reflecting various aspects of yeast biology , or yields clusters with better coverage of the analysed genes . The enriched clusters enable function predictions for uncharacterized genes , and highlight the genetic factors and physical interactions coordinating their transcription across growth conditions .
To integrate the information in multiple physical interaction and gene expression datasets , we first represent each dataset as a network or graph whose nodes are the genes of interest and edges indicate relations between gene pairs such as physical interaction between genes or gene products in physical networks , or transcriptional correlation between genes in coexpression networks . Given multiple graphs defined over the same set of nodes , a simultaneous clustering is a clustering or partition of the nodes such that nodes within each set or cluster in the partition are well connected in each graph , and the total cost of inter-cluster edges ( edges with endpoints in different clusters ) is low . We use a normalized measure to define the connectedness of a cluster in a graph , and take the cost of a set of edges to be the ratio of their weight to the total edge weight in the graph . These normalized measures on clustering quality , described in detail in Methods , enable integration of heterogeneous graphs such as graphs with varying edge densities , and are beneficial over simpler formulations as described in detail in a previous study on clustering a single graph [15] . Our work extends the framework used in the single graph clustering study to jointly cluster multiple graphs , such that the information in all graphs is used throughout the algorithm . The algorithm we designed , JointCluster , simultaneously clusters multiple graphs using techniques that permit theoretical guarantees on the quality of the output clustering relative to the optimal clustering . Since finding the optimal clustering is a computationally hard problem , we prove certain approximation guarantees that show how the cluster connectedness and inter-cluster edge cost measures of the clustering output by our algorithm are reasonably close to that of the optimal clustering ( as formalized in Methods , Theorem 2 ) . The basic algorithm , to which these guarantees apply , works with sparse cuts in graphs . A cut refers to a partition of nodes in a graph into two sets , and is called sparse-enough in a graph if the ratio of edges crossing the cut in the graph to the edges incident at the smaller side of the cut is smaller than a threshold specific to the graph . Graph-specific thresholds enable search for clusters that have varying connectedness in different graphs . The main steps in the basic JointCluster algorithm are: approximate the sparsest cut in each input graph using a spectral method , choose among them any cut that is sparse-enough in the corresponding graph yielding the cut , and recurse on the two node sets of the chosen cut , until well connected node sets with no sparse-enough cuts are obtained . JointCluster implementation employs a novel scaling heuristic to reduce the inter-cluster edge cost even further in practice . Instead of finding sparsest cuts in input graphs separately as in the basic algorithm , the heuristic finds sparsest cuts in mixture graphs that are obtained from adding each input graph to a downscaled sum of the other input graphs . The mixture graph with unit downscaling is the sum graph whose edge weights are the sum of weights of the corresponding edges in all input graphs , and the mixture graphs with very large downscaling approaches the original input graphs . The heuristic starts with mixture graphs with small downscaling to help control inter-cluster edges lost in all graphs . But the resulting clusters are coarse ( eg . clusters well connected in some graphs but split into smaller clusters in the rest are not resolved further ) . The heuristic then refines such coarse clusters at the expense of more inter-cluster edges by increasing the downscaling factor ( see Figure 1 ) . The scaling heuristic works best when combined with a cut selection heuristic: if for a particular downscaling , more than one mixture graph yields a sparse-enough cut , choose among them the cut that is sparse-enough in the most number of input graphs ( breaking ties toward the cut with the least cost of edges crossing the cut in all graphs ) . A rigorous description of the algorithm with heuristics for advancing the downscaling factor and selection of cuts is provided in Methods . Our method runs in an unsupervised fashion since algorithm parameters such as graph-specific thresholds are learnt automatically . The recursive cuts made by our algorithm naturally lead to a hierarchical clustering tree , which is then parsed objectively to produce the final clusters [16] using a modularity score function used in other biological contexts [17] , [18] . The modularity score of a cluster in a graph is the fraction of edges contained within the cluster minus the fraction expected by chance in a randomized graph obtained from degree-preserved shuffling of the edges in the original graph , as described in detail in Supplementary Methods in Text S1 . To aggregate the scores of a cluster across multiple graphs , we take their minimum and use this min-modularity score as the cluster score . The ( min-modularity ) score of a clustering is then the sum of the ( min-modularity ) scores of the constituent clusters . We used simulated datasets to benchmark JointCluster against other alternatives: ( a ) Tree: Choose one of the input graphs as a reference , cluster this single graph using an efficient spectral clustering method [16] to obtain a clustering tree , and parse this tree into clusters using the min-modularity score computed from all graphs; ( b ) Coassociation: Cluster each graph separately using the spectral method , combine the resulting clusters from different graphs into a coassociation graph [19] , and cluster this graph using the same method . Tree method resembles the marginal cluster analysis in [20] as it analyses multiple networks using the clustering tree of a single network . The simulated test data was generated as in an earlier study [18] , under the assumption that the true classification of genes into clusters is known . Specifically , one random instance involved generating two test graphs over nodes each , and implanting in each graph the same “true” clustering of equal-sized clusters . A parameter controlled the noise level in the simulated graphs by controlling the average number of inter-cluster edges incident at a node . The average number of total edges incident at a node was set at 16 , so measures the false positive rate in a simulated graph . We used the standard Jaccard index , which ranges from 0 to 1 , to measure the degree of overlap between the true clustering and the clustering detected by the methods . Please see Supplementary Text S1 for more details . Figure 2 A shows the performance of different methods in recovering common clusters in graphs with the same noise level , averaged over random instances of for each value of the noise level parameter . When the noise level is low ( or false positive rate at most 25% ) , the clusters output by all methods are close to the true set of clusters ( a Jaccard index close to ) . But when the noise level is high ( or false positive rate 25%–50% ) , the cluster structure becomes subtler , and JointCluster starts to outperform other methods and achieves the best improvement in Jaccard index over other methods at . Note that values where false positive rates are above 50% do not lead to a meaningful cluster structure , and are only shown for context . Thus , within the setting of this benchmark , JointCluster outperformed the alternatives in recovering clusters , especially ones with a weak presence in multiple graphs . To simulate real-world scenarios where the integrated networks could've different reliabilities , we benchmarked the methods on clustering graphs with different noise levels . Instead of varying the common value of the graphs as above , we fixed the noise level of at and varied the of the other graph from 0 to 16 . The relative performance of Tree and Tree methods ( see Figure 2 B ) showed that better clusters were obtained when clustering tree of the graph with the lower noise level was used . JointCluster integrated the information in the two graphs to produce a joint clustering tree , which when parsed yielded better clusters than Coassociation and single tree clusters for a larger range of the parameter values ( see Figure 2 B ) . The empirical evaluation of JointCluster and competing methods was done using large-scale yeast datasets , and described in detail next . Expression of transcripts were measured in segregants derived from a cross between the BY and RM strains of the yeast Saccharomyces cerevisiae ( denoted here as the BxR cross ) , grown under two conditions where glucose or ethanol was the predominant carbon source , by an earlier study [14] . From these expression data , we derived glucose and ethanol coexpression networks using all 4 , 482 profiled genes as nodes , and taking the weight of an undirected edge between two genes as the absolute value of the Pearson's correlation coefficient between their expression profiles . The network of physical interactions ( protein-protein indicating physical interaction between proteins and protein-DNA indicating regulatory interaction between a protein and the upstream region of a gene to which it binds ) among the same genes or their protein products , collected from various interaction databases ( eg . BioGRID [21] ) , was obtained from an earlier study [9] . The physical network was treated as an undirected graph after dropping interaction orientations , and contained 41 , 660 non-redundant interactions . We applied JointCluster and other clustering methods to integrate the yeast physical and glucose/ethanol coexpression networks , and assessed the biological significance of the detected clusters using reference sets of genes collected from various published sources . The reference sources fall into five diverse classes: We overlapped the detected clusters with the reference sets in these classes to differentiate clusters arising from spurious associations from those with genes coherently involved in a specific biological process , or coregulated due to the effect of a single gene , TF , or genetic factors . The results are summarized using standard performance measures , sensitivity ( fraction of reference sets significantly enriched for genes of some cluster output by a method ) and specificity ( fraction of clusters significantly enriched for genes of some reference set ) , both reported as percentages for each reference class . The significance cutoff for the enrichment P-value ( denoted hereafter ) is 0 . 005 , after Bonferroni correction for the number of sets tested . The sensitivity measures the “coverage” of different biological processes by the clusters , and the specificity the “accuracy” of the clusters . We compared JointCluster with Coassociation [19] , single graph [16] , and single tree ( Tree ) methods , and when applicable with competing methods , Matisse [9] and Co-clustering [7] , which integrate a single coexpression network with a physical network . All reported results focus only on clusters with at least 10 genes . To provide context , we present results from clustering each network separately using the single graph method ( Glucose/Ethanol/Physical Only ) in Figure 3 A . Physical Only performs better than the other two methods wrt ( with respect to ) GO Process and TF Binding Sites , and Glucose/Ethanol Only fare well wrt eQTL Hotspots . This relative performance is not surprising due to the varying levels of bias in the reference classes , and the different data sources used to construct the networks . Though physical interactions between genes or gene products are known to be predictive of shared GO annotations , certain GO annotations inferred from physical interactions introduce bias . The same ChIP binding data [26] was used to predict TF binding sites and protein-DNA interactions , so validation of clusters derived from the physical network using TF Binding Sites is biased . Finally , the same expression data underlying the coexpression networks was used with the independent genotype data to define the eQTL hotspots [14] . Hence the eQTL Hotspots class does not by itself provide a convincing validation of the coexpression clusters; however it can be used to understand the extent of genetic control of coordinated transcription and to validate clusters derived from networks comprising only physical interactions . The reference classes offering truly independent validation of clusters are TF Perturbations and Compendium of Perturbations , and the three single graph methods perform similarly in these perturbation classes . Integration of the yeast physical network with the glucose/ethanol coexpression networks was done to find sets of genes that clustered reasonably well in all three networks . JointCluster performed a better integration of these networks than Coassociation for all reference classes except eQTL Hotspots ( Figure 3 B ) . The enrichment results of single tree methods in Figure 3 B followed a trend similar to the single graph methods in Figure 3 A , reflecting the bias in the reference classes . In the two truly independent perturbation classes , JointCluster showed better sensitivity than the other methods at comparable or better specificity . In summary , though different single graph and single tree methods were best performers in different reference classes ( from Figures 3 A and 3 B ) , JointCluster was more robust and performed well across all reference classes characterizing diverse cellular processes in yeast ( Figure 3 B , first bar ) . The clusters identified by JointCluster that were consistently enriched for different reference classes are explored in depth next . The clusters in a clustering were ordered by their min-modularity scores , and identified by their rank in this ordering . We highlight the biology and multi-network connectivity of the top-ranked clusters detected by JointCluster in an integrated analysis of the yeast physical and glucose/ethanol coexpression networks . The member genes and enrichment results of all preserved clusters detected by JointCluster are provided as Supplementary Data in Text S1 ( see also Table 1 in Supplementary Text S1 for GO Process enrichment of many top-ranked clusters ) . The preserved cluster with the best min-modularity score , Cluster #1 , comprised genes with a min-modularity score of . The respective modularity scores in the physical , glucose , and ethanol networks were , , and , which were significantly higher than the modularity of a random set of genes of the same size in the respective networks ( see Figure 1 in Supplementary Text S1 for the cluster's connectivity in the three networks ) . This cluster was significantly enriched for genes involved in the GO Processes , translation ( 1e-20; see Table 1 in Supplementary Text S1 ) , mitochondrion organization ( 1e-20 ) , mitochondrial translation ( 1 . 8e-17 ) and cellular respiration ( 3 . 1e-8 ) . The enrichments noted for Cluster #1 is consistent with and even extend published results on this dataset . The shift in growth conditions from glucose to ethanol triggers large changes in the transcriptional and metabolic states of yeast [28] , with the primary state being fermentation in glucose and respiration in ethanol . The transcription of functionally related genes , measured across different timepoints during the shift , are highly coordinated [28] . The coregulation of related genes is also evident from the clusters of coexpressed genes found under the glucose condition , using expression profiles of genetically perturbed yeast segregants from the BxR cross [29] . Our results take this evidence a step further , because the coexpression of cluster genes are elucidated by genetic perturbations in both growth conditions ( regardless of the expression level changes of cluster genes between the conditions ) . We also note that the top-ranked cluster is significantly enriched for genes linking to the eQTL hotspot region glu11 in Chromosome 14 [14] ( 4 . 6e-25 ) , which highlights the role of genetic factors in the coregulation of genes involved in ( mitochondrial ) translation and cellular respiration . A different perspective on yeast biology in the glucose medium is offered by Cluster #2 consisting of genes ( with a significant min-modularity score 0 . 00021; see Figure 1 in Supplementary Text S1 ) . This cluster is significantly enriched for ribosome biogenesis ( 2 . 4e-37; see Table 1 in Supplementary Text S1 ) , and related GO Process terms such as ribonucleoprotein complex biogenesis and assembly ( 9 . 4e-37 ) , ribosomal large subunit biogenesis ( 8 . 8e-35 ) , and rRNA processing ( 3 . 8e-33 ) . Genes in this cluster significantly overlap with the perturbation signature of BUD21 , a component of small ribosomal subunit ( SSU ) processosome ( 4 . 1e-15 ) , and with genes whose expression links to genetic variations in the eQTL hotspot region glu12 in Chromosome 15 [14] ( 7 . 9e-16 ) . These results are consistent with the literature on the regulation of yeast growth rate in the glucose or ethanol medium , achieved by coregulation of genes involved in ribosome biogenesis and subsequent protein synthetic processes [28] . To further understand the biological significance of these preserved clusters in physical and coexpression networks , we used the reference yeast protein complexes in MIPS [30] ( comprising literature-based , small-scale complexes of at least five genes , at level at most two in the MIPS hierarchy ) . The enrichment of the joint clusters wrt this MIPS Complex class was % sensitivity and % specificity . Of the clusters not enriched for any MIPS complex , some were significantly enriched for other functionally coherent pathways ( eg . Cluster #13 was enriched for amino acid biosynthetic process; see Table 1 in Supplementary Text S1 ) . So the clusters detected by JointCluster overlapped with several known complexes or other functional pathways . One of the goals of jointly clustering multiple networks is to identify subtle clusters: sets of genes that cluster reasonably well , but not strongly , in all networks . We start with biologically significant clusters i . e . , clusters enriched for some reference set wrt all five reference classes , and test if any such cluster has a weak modularity score in some graph . We identified 5 biologically significant clusters using JointCluster: Clusters #4 , #13 , #15 , #19 , and #28 . Table 2 in Supplementary Text S1 shows the reference sets they were enriched for , and Figure 2 in Supplementary Text S1 the modularity scores of Clusters #4 and #28 . Cluster #28 , the biologically significant cluster with the lowest min-modularity score , had genes and was enriched for the GO Processes , multi-organism process ( 2 . 5e-12 ) and conjugation ( 2e-10 ) . This cluster's role in mating was further supported by its significant enrichment for perturbation signatures of STE12 ( 5 . 7e-9 ) and FUS3/KSS1 ( 8 . 1e-21 ) , because Ste12p is a TF regulating the expression of mating genes and is activated by the Fus3p/Kss1p kinases in the well-studied mitogen-activated protein kinase ( MAPK ) cascade [31] . Such a cluster of well-studied genes was recovered just by the single graph method Physical Only , but not by Glucose/Ethanol Only . Here we considered a cluster of genes to be recovered by a method if this cluster is significantly enriched for some cluster found by the given method ( as in reference set enrichment ) . JointCluster was able to detect this cluster due to its high modularity in the physical network combined with its significant , albeit weak , modularity in the coexpression networks ( see Figure 2 in Supplementary Text S1 ) . To explore more subtle clusters , we focused on the clusters identified by JointCluster that were enriched for at least four reference classes , instead of all five required above . Clusters #52 and #54 had the two lowest min-modularity scores among such clusters , and were each recovered just by the Physical Only method , but not by Glucose/Ethanol Only . Cluster #52 comprised of genes had a significant min-modularity score ( see Figure 3 in Supplementary Text S1 ) , and was enriched for the GO Process , ubiquitin-dependent protein catabolic process ( 4 . 4e-23 ) . RPN4 is a TF involved in regulation of the protein catabolic process [32] , and this cluster was significantly enriched for genes in the deletion signature of RPN4 ( 1 . 4e-9 ) and genes with predicted binding sites of RPN4 ( 1 . 8e-23; see Supplementary Data in Text S1 for other enrichments ) . These examples reiterate how a combined analysis of multiple networks by JointCluster detects meaningful clusters that would be missed by separate clustering of the networks . Despite the intense focus on elucidating yeast biology by many researchers , roughly 1 , 000 Open Reading Frames ( ORFs ) are still uncharacterized [33] . Therefore , predicting the function of these ORFs is important to guide future experiments towards strains and perturbations that likely elucidate these ORFs [33] . While there have been many network-based function prediction studies ( see survey [34] ) , our study provides a different perspective by using clusters preserved across multiple coexpression and physical networks . Our prediction strategy , based on a module-assisted guilt-by-association concept [34] , annotates the uncharacterized ORFs in a cluster detected by JointCluster to the GO Process reference set for which this cluster is most significantly enriched . To test the utility of these predictions for a well-studied process in yeast , we focused on clusters enriched for ribosome biogenesis ( Clusters #2 and #22; see Table 1 in Supplementary Text S1 ) . Two ORFs in Cluster #2 , a top-ranked cluster discussed above , were marked as uncharacterized by SGD [35] ( April 2009 version ) : YER067W and YLR455W . Our predictions for these ORFs have different types of support: YER067W is significantly correlated with 67 and 33 of the 76 genes in this cluster in glucose/ethanol expression datasets respectively ( Pearson's correlation test , Bonferroni corrected for the cluster size ) , and YLR455W has known protein-protein interactions with five other genes in the cluster , NOC2 , BRX1 , PWP1 , RRS1 , EBP2 , all of which were implicated in ribosome biogenesis . Cluster #22 had 9 uncharacterized ORFs , YIL096C , YOR021C , YIL091C , YBR269C , YCR087C-A , YDL199C , YKL171W , YMR148W , and YOR006C . Two of them ( YIL096C and YOR021C ) have predicted roles in ribosome biogenesis based on function predictions collected from the literature by SGD for some of the uncharacterized ORFs . This lends support to the two predictions and leaves the other novel predictions for further validation . All of the uncharacterized ORFs in Cluster #22 except YBR269C were significantly correlated with more than three-fourths of the 35 genes in the cluster in both glucose/ethanol expression datasets ( using the same criteria above based on Pearson's correlation test ) . The predictions here were based on either support from the physical network ( for YLR455W ) or from both coexpression networks ( for the rest ) , and hence illustrates the advantage of using multiple data sources . Of the 990 ORFs classified as uncharacterized by SGD ( April 2009 version ) , 524 overlapped with the genes used to build the yeast networks . We could predict the function for 194 of them , by virtue of their membership in preserved clusters significantly enriched for some GO Process term . Using single graph ( Glucose/Ethanol/Physical Only ) clusters in place of the preserved clusters detected by JointCluster yielded predictions for 143 , 148 and 247 uncharacterized ORFs respectively , reflecting the relative GO Process specificity of these methods ( Figures 3 A and 3 B ) . The relative number of predictions from different methods should be viewed in context of the systematic evaluations above , which showed that whereas Physical Only performed best wrt GO Process , JointCluster produced clusters that were more coherent across all reference classes . The predictions from JointCluster were also complementary to those from Physical Only , with the functions of only uncharacterized ORFs predicted by both methods . The functions predicted using the preserved clusters are available as Supplementary Data in Text S1 , and point to well-studied biological processes that have escaped complete characterization . To compare JointCluster against methods that integrate only a single coexpression network with a physical network , such as Matisse and Co-clustering , we considered joint clustering of a combined glucose+ethanol coexpression network and the physical network . The glucose+ethanol network refers to the single coexpression network built from expression data that is obtained by concatenating the normalized expression profiles of genes under the glucose and ethanol conditions . The results of different methods on this two-network clustering is in Figure 4 A . Since our results focus on clusters with at least 10 genes , we set the minimum cluster size parameter in Matisse to 10 ( from its default 5 ) . All other parameters of Matisse and other competing methods were set at the default values . The default size limit of 100 genes for Matisse clusters was used for JointCluster as well to enable a fair comparison . Co-clustering didn't have a parameter to directly limit cluster size . Despite setting its parameter for the number of clusters at 45 to get an expected cluster size of 100 , Co-clustering detected very few ( 26 ) clusters of size at least 10 genes , half of which were large with more than 100 genes ( including one coarse cluster with more than 800 genes ) . So Co-clustering achieves greater specificity than other methods ( Figure 4 A ) at the expense of a coarser clustering comprising few large clusters . JointCluster has sensitivity and specificity that is comparable or slightly lower than Matisse across all reference classes except TF Binding Sites . However , JointCluster produces clusters that cover significantly more genes than Matisse ( 4382 vs 2964 genes respectively; see also Figure 4 A ) . Matisse assumes that the physical network is of better quality , and searches for coexpression clusters that are each connected in the physical network . This connectivity constraint excludes genes whose physical interactions are poorly studied or untested . JointCluster does not use such a constraint when parsing the clustering tree into clusters , and hence identifies clusters supported to varying extents in the two networks , including ones with weak support in the physical network . This could be a huge advantage in organisms such as human and mouse where the knowledge of physical interactions is far less complete than in yeast , especially for interactions that are tissue-specific or condition-specific . The extreme examples among the roughly 1500 genes excluded by Matisse clusters were the physically isolated genes ( i . e . , genes that do not interact with any of the other profiled genes in the physical network ) . JointCluster used connectivity in the coexpression network to include physically isolated genes in its clusters , and of these genes were significantly correlated ( Pearson's correlation test , Bonferroni corrected for the cluster size ) with more than half of the genes in their assigned cluster . Figure 4 in Supplementary Text S1 shows example coexpression clusters identified by JointCluster despite the poor physical connectivity among the isolated/other genes within the clusters . The earlier study on Matisse extended physical connectivity within clusters by adding extra genes called “back” genes and their interactions to the physical network [9] . The physical interactions of back genes serve to better connect the ( expression ) profiled genes in the physical network , but the back genes' expression data is not used ( or not available ) for analysis . The results of integrating this extended yeast physical network , with 1774 extra back genes and 22 , 330 extra interactions , with the glucose+ethanol coexpression network is in Figure 4 B . The clusters of JointCluster covered a large fraction of genes , comprising back and profiled genes , but they showed poor specificity due to the inclusion of several back genes with no expression information . Matisse on the other hand was specially designed to exploit a few of these back genes as needed to enhance physical connectivity , so it showed better sensitivity and specificity at a coverage of 182 back and 3327 profiled genes . Though back genes helped increase Matisse's coverage of profiled genes , Matisse clusters still missed several of the 4482 profiled genes . Considering the results before and after extension of the physical network , we see that the inclusive criteria used in JointCluster is preferable when the integrated physical network is not comprehensive .
Heterogeneous large-scale datasets capturing diverse aspects of the biology of a cell are accumulating at a rapid pace , and efforts to integrate them into a coherent view of cell regulation are intensifying . This integration could greatly facilitate a genome-wide model of the cell that could predict cellular response to various environmental and genetic perturbations ( eg . [36] ) . The simultaneous clustering algorithm proposed here provides a versatile approach to integrating any number of heterogeneous datasets that could be represented as networks among genes , and summarizes the result as a collection of clusters supported by multiple networks . Since its early applications to classifying cancer subtypes [37] , clustering has rapidly become a standard analysis of expression datasets . We believe that simultaneous clustering is a natural progression in the application of clustering from single to multiple expression and interaction datasets . We demonstrated the utility of a combined analysis by applying our JointCluster algorithm on simulated and empirical datasets . We found the clusters produced by JointCluster on yeast physical and glucose/ethanol coexpression networks to be comparably or more consistently enriched for reference classes that reflect various aspects of yeast biology , in comparison to other methods of integrating the networks . Further , JointCluster can handle multiple heterogeneous networks , and hence more flexible than two-network clustering methods such as Matisse that search for coexpression clusters that are each connected in the physical network . This flexibility enables JointCluster to yield better coverage of genes , and to be broadly applicable in human or other organisms where the knowledge of physical interactions is less complete than in yeast . In the future , the framework could be extended to scale networks of different interaction types by different factors before integration . Simultaneous clustering offers an unsupervised and exploratory approach to data integration , and hence complementary to supervised approaches that train machine-learning models on multiple data types to make directed predictions . Such supervised approaches could integrate different data types to predict functional linkages between gene pairs ( see [38] , [39] and references therein ) , or protein complexes using a training set of known complexes [40] . Though our method is not directed to predict complexes , the joint physical and coexpression clusters we found were enriched for reference protein complexes in MIPS . More importantly , the unsupervised fashion in which we parsed the joint clustering tree recovered functionally coherent pathways other than complexes . Simultaneous clustering is also complementary to unsupervised approaches that identify spectral patterns ( not modules ) shared between similarity graphs based on gene expression or TF binding data [41] , or identify modules from paired datasets such as gene expression and drug responses profiled in the same cell lines [42] . The clusters detected by JointCluster from the yeast physical and expression datasets are consistent with known biology , and importantly extend our knowledge by highlighting biological processes , such as ribosome biogenesis , that may not have been completely characterized despite intense efforts to dissect them . The tangible value of a combined analysis is evident from the systematic evaluation of the clusters , and the case studies presented in this work . The intangible benefit of seeking support in the multiple networks considered in this study is the ready interpretation provided by the protein-protein and protein-DNA interactions within a cluster , in explaining the coordinate transcription of the cluster .
Consider a graph , where is the set of nodes and is a non-negative edge weight function . The weight for any node pair could for instance quantify the connection strength or similarity between the two nodes; note that a sparse graph would've many zero weight edges . For convenience , let us denote the total weight of any edge set by . Similarly for any node sets , let , and . Using these notations , the total edge weight in the graph is . Also for singletons , sums up the weights of edges incident at node . The conductance of a cut in a node set , measured using the function , is defined as ( with the convention that this ratio is zero if its denominator is zero , since the numerator is also zero then ) . By normalizing the sum of edge weights crossing the cut , the definition captures an intuitive notion of connectivity that is robust and invariant to scaling the edge weight function by any constant . To illustrate the intuitive notion , consider a cut separating a single node from other nodes in a cluster . If the conductance of this cut is high , then a large fraction of all edges incident at node ends at another node in the cluster . Extending this notion , if the conductance of all cuts in are high , the nodes in are robustly connected together . So the conductance of a cluster is defined as the minimum conductance of any cut in the cluster , and the conductance of a clustering or partition of is the minimum conductance of any cluster in the partition . When maximizing the conductance of the partition , it is desirable to control the cost of the inter-cluster edges as well . Let denote the inter-cluster edges , i . e . , unordered node pairs where and belong to different clusters in the partition . An clustering of is a partition of its nodes into clusters such that We outline the approximate-cluster algorithm and its guarantees presented in [15] . The algorithm finds a cut approximating the sparsest cut ( cut of minimum conductance ) in the graph and recurses on both the pieces induced by this cut . Since finding the sparsest cut in a graph is a NP-hard problem , an approximation algorithm for the problem is used . Note that clustering a graph by minimizing for a given is also NP-hard by a reduction from the sparsest cut problem . The repeated removal of sparsest cuts is done until the pieces or clusters become well connected with no sparse cuts left in them . Making repeated cuts to partition a graph is strategically similar to a method on clustering an expression dataset [43] , but that method works with minimum cuts rather than sparsest cuts . Sparsest cut is preferable in our context of heterogeneous datasets , since it minimizes the normalized measure of conductance . To formalize the guarantees on the approximate-cluster algorithm , let denote the number of nodes in the graph , and let the sparsest cut of conductance be approximated by a cut of conductance at most ( where is independent of , and is a constant between and ) . For instance , there are algorithms to find a cut of conductance at most using metric embedding techniques [44] ( all logarithms in this paper are to base two ) , or using efficient spectral techniques [15] , [45] ( our implementation uses spectral techniques ) . Consider graphs over the same nodes and different non-negative edge weight functions . An simultaneous clustering of the graphs is a partition of the nodes such that The conductance thresholds are graph-specific to enable search for clusters of varying quality in heterogeneous graphs . A natural approach to the inter-cluster edge cost is the sum of graph-specific costs , however it's a special case of the above aggregated cost when each edge weight function is scaled by a constant . Note how these scalings , which our implementation employs , set the total edge weight of each graph to the same value without changing the conductance value of cuts . This scale-invariance of conductance comes from the normalization factor in its definition as mentioned before . The problem of minimizing the inter-cluster edge cost given a set of conductance thresholds is NP-hard by reduction from the single graph case . We now present our basic algorithm , JointCluster , to simultaneously cluster multiple graphs , along with certain approximation guarantees on the quality of the clustering produced . The algorithm starts with as the current node set . For each graph , the algorithm finds an approximate sparsest cut in the current node set , using the graph-specific edge weight function to measure conductance . The algorithm chooses among them any cut that is sparse enough as defined below , and recurses on the two pieces ( node sets ) induced by this cut . If no cuts get chosen for the current node set , the node set is output as a well connected cluster in all graphs . The cut approximating the sparsest cut in the current node set in is sparse enough if the conductance of the cut , measured using the edge weight function , is at most . To provide formal guarantees on the clustering produced by this algorithm , let , denote the respective edge weight functions of a sum and a min graph obtained from the multiple graphs . That is , for every edge , and . As before , let be the number of nodes , the approximation guarantee of the sparsest cut algorithm , and the inter-cluster edges of a given partition . We analysed our JointCluster algorithm ( see Supplementary Methods in Text S1 ) to prove this theorem: The theoretical guarantees of JointCluster algorithm are further augmented by effective heuristics and efficient implementation in practice .
|
The generation of high-dimensional datasets in the biological sciences has become routine ( protein interaction , gene expression , and DNA/RNA sequence data , to name a few ) , stretching our ability to derive novel biological insights from them , with even less effort focused on integrating these disparate datasets available in the public domain . Hence a most pressing problem in the life sciences today is the development of algorithms to combine large-scale data on different biological dimensions to maximize our understanding of living systems . We present an algorithm for simultaneously clustering multiple biological networks to identify coherent sets of genes ( clusters ) underlying cellular processes . The algorithm allows theoretical guarantees on the quality of the detected clusters relative to the optimal clusters that are computationally infeasible to find , and could be applied to coexpression , protein interaction , protein-DNA networks , and other network types . When combining multiple physical and gene expression based networks in yeast , the clusters we identify are consistently enriched for reference classes capturing diverse aspects of biology , yield good coverage of the analysed genes , and highlight novel members in well-studied cellular processes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"genetics",
"and",
"genomics/bioinformatics",
"computational",
"biology/systems",
"biology",
"computational",
"biology/genomics",
"computational",
"biology/transcriptional",
"regulation"
] |
2010
|
Simultaneous Clustering of Multiple Gene Expression and Physical Interaction Datasets
|
In RNA silencing , small RNAs produced by the RNase-III Dicer guide Argonaute-like proteins as part of RNA-induced silencing complexes ( RISC ) to regulate gene expression transcriptionally or post-transcriptionally . Here , we have characterized the RNA silencing machinery and exhaustive small RNAome of Toxoplasma gondii , member of the Apicomplexa , a phylum of animal- and human-infecting parasites that cause extensive health and economic damages to human populations worldwide . Remarkably , the small RNA-generating machinery of Toxoplasma is phylogenetically and functionally related to that of plants and fungi , and accounts for an exceptionally diverse array of small RNAs . This array includes conspicuous populations of repeat-associated small interfering RNA ( siRNA ) , which , as in plants , likely generate and maintain heterochromatin at DNA repeats and satellites . Toxoplasma small RNAs also include many microRNAs with clear metazoan-like features whose accumulation is sometimes extremely high and dynamic , an unexpected finding given that Toxoplasma is a unicellular protist . Both plant-like heterochromatic small RNAs and metazoan-like microRNAs bind to a single Argonaute protein , Tg-AGO . Toxoplasma miRNAs co-sediment with polyribosomes , and thus , are likely to act as translational regulators , consistent with the lack of catalytic residues in Tg-AGO . Mass spectrometric analyses of the Tg-AGO protein complex revealed a common set of virtually all known RISC components so far characterized in human and Drosophila , as well as novel proteins involved in RNA metabolism . In agreement with its loading with heterochromatic small RNAs , Tg-AGO also associates substoichiometrically with components of known chromatin-repressing complexes . Thus , a puzzling patchwork of silencing processor and effector proteins from plant , fungal and metazoan origin accounts for the production and action of an unsuspected variety of small RNAs in the single-cell parasite Toxoplasma and possibly in other apicomplexans . This study establishes Toxoplasma as a unique model system for studying the evolution and molecular mechanisms of RNA silencing among eukaryotes .
Apicomplexa are unicellular eukaryotes that multiply intracellularly in their mammalian hosts . They include parasites of major medical importance like Plasmodium species , the causative agent of malaria , and Toxoplasma gondii , the most widespread apicomplexan parasite , present virtually everywhere on earth . Although usually causing only mild symptoms in the adult , Toxoplasma can cause severe and life-threatening diseases in developing fetuses and in immunocompromised individuals , especially AIDS and transplant patients [1] , [2] . Toxoplasma has a complex life cycle that includes infections of more than one host organism , differentiation through several morphologically distinct forms , and both sexual and asexual replication [3] . Changes in gene expression is expected as ( i ) parasites progress through the cell cycle , ( ii ) parasites differentiate in specific stages , and ( iii ) parasites are exposed to the host immune system during infection [4] . How these changes are regulated at the molecular level remains to a large extent unknown . A puzzling feature is the apparent lack , in apicomplexan parasites , of large families of recognizable specific transcription factors ( TFs ) operating in other eukaryotes [5] . Despite the paucity of recognizable TFs , apicomplexans are endowed with a rich repertoire of enzymes associated with epigenetics and chromatin remodeling , and this observation has fueled the idea that epigenetics could play an important role in the control of gene expression [6] , [7] . Small regulatory RNAs are linked to epigenetic regulation of gene expression in several organisms but these are presently understudied in the Apicomplexa . The defining features of small silencing RNAs are their short length ( ∼20–30 nucleotides ) and their association with members of the Piwi/Argonaute ( AGO ) family of proteins , which they guide to their regulatory targets [8] , [9] . Many , albeit not all , small RNAs ( sRNA ) are produced by the RNase III-related enzyme Dicer . Small interfering RNAs ( siRNA ) are generated as populations from multiple Dicer cleavages along long dsRNA precursors , whereas microRNAs ( miRNA ) are discrete species generated from a single Dicer cleavage event of noncoding primary precursor transcripts containing small , imperfect stem–loop structures [10] . These distinct small RNA pathways compete and collaborate as they regulate genes and protect genome integrity from invading nucleic acids including viruses and transposons . They function as guides for effector complexes ( RNA-induced silencing complexes , RISCs ) that regulate gene expression by degrading mRNA , repressing its translation , or modifying chromatin . RNA silencing is an evolutionary ancient regulatory mechanism , and small RNA pathways in unicellular organisms appear , so far , to be relatively simple . In fission yeast , a single class of endogenous siRNAs has demonstrated roles in epigenetic silencing at centromeres and the initiation of heterochromatin assembly at the mat locus [11] . In the ciliated protozoan Tetrahymena thermophila , small RNAs are involved in developmentally regulated DNA elimination [12] , [13] and post-transcriptional gene regulation [14] . Particularly surprising is the recent finding that the unicellular green alga Chlamydomonas produces microRNAs that had been previously associated with developmental regulation and multi-cellularity [15] , [16] . Here , we show that the T . gondii genome , unlike in Plasmodium species [17] , encodes all core components of an elaborate RNA silencing machinery that has been evolutionary shaped as a patchwork of factors of plant and fungal origin . We establish a comprehensive sRNA landscape of T . gondii through deep sequencing , and unravel that the most abundant sRNA classes are formed by metazoan-like miRNAs as well as plant-like repeat-and-satellite-associated sRNAs coined rdsRNA and satRNA , respectively . Beyond the surprising complexity of the small RNAome , we provide a thorough biochemical characterization of the proteins that associate with the single T . gondii Argonaute protein , Tg-AGO . Unexpectedly these proteins constitute the near-entire cohort of previously identified human and fly miRNA-RISC components . Our data indicate that miRNA-loaded T . gondii argonaute associates with polysome probably to regulate translation of Tg-miRNA predicted targets , many of which include mRNA with perfect or near perfect complementarity . Tg-AGO also co-purifies with chromatin-repressing complexes , suggesting a role in transcriptional silencing , most likely through it demonstrated association with rdsRNAs or satRNAs .
Previous analyses have suggested a monophyletic origin for plant and animal Dicer proteins [18] . Sequence analyses show that the Toxoplasma genome ( TOXODB release v6 . 0 ) [19] encodes only one Dicer-like protein , Tg-Dicer , which displays significant variability in primary sequence and domain organization compared to the Dicer consensus ( Figure 1A ) : Tg-Dicer possesses an RNA helicase domain and two RNaseIII catalytic domains ( RNaseIIIa and RNaseIIIb ) , but it lacks recognizable domains for dsRNA binding ( DSRM ) and PIWI-ARGONAUTE-ZWILLE ( PAZ ) functions . This organization is strikingly reminiscent of the DCL1 protein of the single cell algae C . reinhardtii ( Figure 1A ) . Toxoplasma and Chlamydomonas Dicer-like sequences seem , indeed , orthologous , as they form a specific clade supported by a strong bootstrap score ( Figure 1B ) , a consequence of a Drosha-like signature polypeptide that is more related ( albeit weakly ) to eubacterial RNaseIII enzymes and known to form an out-group with respect to higher plant and animal Dicers [18] . Ago-related proteins are divided into the Ago-like and Piwi-like subfamilies ( Figure 2A ) ; a third clade , termed ‘Group 3 Argonautes’ , is worm-specific [20] . Toxoplasma Argonaute ( Tg-AGO ) is represented at a single genomic locus; there is no evidence for Toxoplasma-encoded Piwi proteins . Tg-AGO belongs to the Ago-like family but only with weak bootstrap support , suggesting that the protein diverges significantly from its metazoan and plant counterparts . Nonetheless , the two key signature domains of the AGO family —the PAZ domain and the C-terminal PIWI domain— are conserved in Tg-AGO ( Figure 2B ) . Overall , the PIWI domain shows the highest degree of conservation , but the Asp-Asp-Glu/Asp catalytic triad required for slicer activity is not found , suggesting that the protein lacks endonucleolytic cleavage capacity ( Figure 2B ) . The second signature motif , the PAZ domain , contains only a few residues that are strictly conserved , while the middle ( MID ) domain of Tg-AGO harbors the residues ( Y596 , K600 , Q610 and K644; Figure 2C ) required to bind the characteristic 5′ phosphate group of guide small RNA strands [21] , [22] . We also noted the presence , in the amino terminal part of Tg-AGO , of a stretch of repeated RGG residues ( amino acids 1–68 ) , in which the arginines have the potential to undergo methylation ( Figure 2B ) . This feature is found in metazoan and plant AGO-related proteins and was shown to alter their stability and/or sub-cellular distribution , ultimately impacting their function [23] , [24] , [25] . A phylogenetic tree constructed by aligning the stereotypical RNA-dependent RNA polymerase ( RDR ) domain supports the monophyletic origin of the proteins found in C . elegans , fungi and Arabidopsis [26] . Inspection of the Toxoplasma genome showed the presence of a single RDR-like gene , suggesting the existence of an amplified RNA silencing machinery in this organism . Tg-RDR is closely related to Neurospora crassa RDR , QDE1 , and forms a specific clade with plant RDRs , which itself constitutes an out-group from the RDRs of metazoans and from the fission yeast S . pombe ( Figure 1C ) . We conclude from this analysis that a patchwork of factors of plant and fungal origin form the core processor components of the Toxoplasma RNA silencing machinery . This finding can be rationalized partly by the fact that the apicomplexa ancestor is a presumed endosymbiont of red algae [27] . We note , however , the moderate or poor phylogenetic relationship observed between Tg-Dicer , Tg-AGO and the corresponding paralogous proteins of its mammalian hosts . Having established that the Toxoplasma genome encodes all core components of an elaborate RNA silencing machinery , we sought to determine the small RNA landscape of this organism . To this aim , we prepared total RNA from freshly released , filtered parasites . Ethidium bromide staining revealed a relatively abundant class of small ( s ) RNAs , ∼30 nucleotides ( nt ) in length ( Figure S1A ) . The 20–40nt sRNA fraction was recovered by gel excision , cloned and subjected to deep sequencing using the Illumina technology . The sRNA library was constructed so as to represent only those sRNAs with a 5′ monophosphate and a 3′ hydroxyl group , the termini expected of miRNAs and siRNAs [28] . About 75% of a total of 5 , 701 , 506 reads , represented tRNA and rRNA turnover products , as previously reported for other organisms ( Figure S1B ) [29] . After filtering low quality reads , 3′ , 5′ adapters , and reads shorter than 17 nucleotides , a remaining 1 , 555 , 290 reads ( ∼30% of total reads ) matched the Toxoplasma genome ( ToxoDB , version 4 . 3 ) ( Figure S1B ) . Comparing mRNA and sRNA data for highly expressed genes suggested that the sRNA fraction contained only a very low level of degradation products from longer mRNAs ( data not shown ) . Most sRNA sequences were found to be 25–27nt in length , with 25nt representing the dominant size class ( Figure 3A ) . Of the 1 , 222 , 203 total reads , 94 , 170 corresponded to non-redundant sRNAs , and 92% of these were single reads , thus unraveling a highly complex sRNA population . Plotting Toxoplasma sRNAs ( Tg-sRNAs ) species with 100% match on the reference genome ( in 10-kb sliding windows ) showed that non-redundant Tg-sRNA with high read numbers ( >1000 ) originate predominantly from non-coding intergenic regions or are embedded within introns of protein-coding transcriptional units ( TUs ) ( Figure 3B ) . Other , medium-to-low abundance Tg-sRNA , by contrast , mapped to protein-coding TUs and a variety of DNA repeats and satellites ( Figure 3B ) . These two classes of Tg-sRNA were detailed further , as described in the following sections . In a search for putative Toxoplasma miRNA candidates , we evaluated sRNA reads exhibiting the following features: ( 1 ) high abundance of sequence reads sharing the same 5′ terminus; ( 2 ) exact match to one or several genomic loci displaying a characteristic fold-back structure typical of MIRNA precursors; and ( 3 ) , when applicable , low abundant sequence reads corresponding to the labile ( miRNA* ) passenger strand of miRNA/miRNA* duplexes , as predicted within the fold back structures . Fourteen sRNA families cloned at a high frequency met these MIRNA features and were thus annotated with high confidence as T . gondii miRNAs ( Tg-miRNAs ) ( Figures 3 , S2-S13 and Table S1 ) . Genome browser views of some of these Tg-MIRNA indeed indicated the existence of a low frequency , single miRNA* ( passenger strand ) corresponding to the opposite strand of the duplex within the fold back-structure ( e . g . Tg-miR-60b; Figure 3C ) . Moreover , the reconstituted duplexes sometimes had small 3′ overhangs characteristic of Dicer processing ( Figures 3C , S8-S10 ) . Of the 14 annotated Tg-miRNA families , 7 gave detectable hybridization signals in Northern analysis carried out with 5′ end-labeled antisense oligonucleotides ( Figures 4A , 4C and S14 ) . No signal was detected with RNA extracted from host cells , confirming that these sRNAs are Toxoplasma-specific . Hybridization often unraveled discrete sRNA species that were heterogeneous in size , reminiscent of the sRNA signals observed in S . pombe [30] and other organisms in which Dicer lacks a PAZ domain , required for the precise sizing of processed sRNAs [31] , [32] , [33] . In all cases , hybridization with antisense probes from precursor sequences flanking the mature miRNA gave no signal ( data not shown ) , confirming the excision , by Tg-Dicer , of a single sRNA species , a landmark of plant and metazoan miRNA biogenesis . The members of the remaining 7 miRNA families were below detection levels , in agreement with their much lower cloning/sequencing frequencies ( Figures S8-S13 ) . In all cases -and as expected from their poor read counts due to their intrinsic instability- cloned miRNA* were also below detection levels of Northern analysis . Sequencing and Northern analyses showed that the miR-60 family largely dominates the Toxoplasma miRNA landscape , accounting for 335 , 014 reads , of which 61% ( 280 , 723 reads ) were contributed by miR-60a alone ( Figure S2B ) . MIR-60 , together with MIR-4 , also constitute the two most diversified Tg-MIRNA gene families ( with 8 distinct members in each ) among the 14 families identified with high confidence ( Figure S2B and S3B ) . In most cases of Tg-miRNAs with multiple precursors ( 6 families , Table S1 ) , the mature miRNAs were not located on the same fold-back arms , which , furthermore , were also found to vary in sequence , suggesting that these genes do not share a common ancestor and , thus , have evolved separately . The 14 high-confidence Tg-miRNAs showed no significant homology to any of the known miRNAs of plants and metazoans , as assessed in the central miRBase depositary ( release 14 ) . Nearly all Tg-miRNAs and Tg-miRNA* ( when available ) had , however , directly identifiable orthologs in the genome of the apicomplexan Neospora caninum ( dog parasite ) , when up to 3-nucleotide polymorphism was tolerated ( Figures 3 , S2-S8 and S13 ) . Moreover , the size and abundance of these orthologous N . caninum ( Nc ) -miRNAs was confirmed by Northern blot analysis using the same antisense oligonucleotide probes employed for detection of Tg-miRNAs ( Figures 4A and S14 ) . The notable exception to sequence conservation in N . caninum was observed with Tg-miR-62 , -64 , -65 or -66 , although this could be attributed to the incomplete N . caninum genome annotation ( Figures S9-S12 ) . Taking into account recent observations made in the single cell algae Chlamydomonas , apicomplexans thus provide the second reported example of unicellular organisms that produce miRNAs . Unlike in Chlamydomonas , nonetheless , and despite the relatedness of the Dicer proteins found in both organisms ( Figure 1A and 1B ) , Tg-MIRNA fold-backs have length and thermodynamic features that are much closer to those of mammalian hosts than those of Chlamydomonas or higher plants [10] . Consistent with this idea , most Tg-miRNAs display a clear 5′- nucleotide bias towards A , as also observed for most mammalian miRNAs [34] . Unlike many mammalian MIRNAs , however , Tg-MIRNAs were not found to form genomic clusters . These findings further emphasize the surprising mosaic nature of the Toxoplasma RNA silencing machinery and small RNA loci . The above 14 miRNA families were identified through deep-sequencing of small RNAs isolated from freshly egressed parasites , and so other miRNAs might exist that were simply too low in abundance to be cloned under these specific growth conditions . In addition , several Tg-sRNAs cloned at moderate to low frequency mapped to imperfect fold-backs scattered along the genome , with relatively low free energy ( Figure S15 ) . These hairpins are much more heterogeneous in size and structure than cognate Tg-MIRNA precursors , yet their processing produces discrete sRNA species . Although their relatively modest cloning frequencies precludes their detection by Northern analysis , including in N . caninum , the corresponding sRNA might represent recently-evolved miRNAs that may engage into miRNA-like regulatory activities . In plants , a model for MIRNA gene evolution , termed “spontaneous evolution” stems from the high density of small-to-medium sized fold-back sequences scattered throughout the Arabidopsis genome . It has been proposed that following the capture of transcriptional regulatory sequences , some of these random fold-backs could occasionally give rise to new MIR genes . Stabilization through co-evolution with targets initially found by chance could then lead to the fixation of these genes in the genome [35] . Unlike metazoan miRNAs , plant miRNAs are methylated at the 2′ hydroxy positions of their 3′-last nucleotides [36] . This modification , mediated by the methyl-transferase HEN1 , protects miRNA from 3′ end uridylation and subsequent degradation [37] . However , Tg-miRNA species were found sensitive to b-elimination by periodate , which causes a diagnostic shift in sRNA mobility ( Figure 4F ) . Thus , unlike their plant counterparts , but similar to metazoan miRNAs , Tg-miRNAs do not carry 3′-end modifications , a result also consistent with our failure to identify a HEN1 homolog in the Toxoplasma genome ( TOXODB , release v5 . 2 ) . Nonetheless , the 3′ end of several cloned Tg-miRNAs was often found to contain untemplated adenine residues , which must be added , therefore , after processing by an as yet unidentified terminal adenyl-transferase ( Figure S3B ) . It was shown recently that addition of adenylic acid residues on the 3′-end apparently slows down miRNA turnover in Populus trichocarpa [38] . In plants and animals , miRNA are recruited by AGO proteins to enhance the turnover , or inhibit the translation of cognate mRNA targets . Consequently fractions of most plant and metazoan miRNAs associate with polysomes , the sites of active translation . The existence of miRNAs in Toxoplasma together with the absence of detectable slicer residues in the Tg-AGO predicted that Tg-miRNA would also associate , at least partly , with polysomes . To test this idea , protein extracts from freshly egressed parasites ( E ) ready to invade , or from fast-growing intracellular parasites ( I ) , were fractionated and resolved on sucrose density gradients ( see Methods ) . For the former ( I ) , the absorbance profiles at 254 nM reflected the ribosome pattern expected from rapidly growing cells: there were few monosomes ( 80S ) and the bulk of the ribosomes sedimented in the polysomal fractions ( Figure 4D ) . By contrast , the amount of polyribosomes in invading parasites ( E ) was substantially reduced , and this was accompanied by a concomitant increase in 80S monosomes ( Figure 4D ) . As previously observed in plants and metazoans , Tg-miRNAs distribution was found to span a wide range of molecular weights across the gradient ( Figure 4E ) [39] , [40] , [41] . Nonetheless , a fraction of several Tg-miRNAs co-sedimented with polysomes ( Figure 4E , fractions 13–18 ) . Moreover , the association with translating ribosomes was more pronounced in exponentially growing parasites ( I ) , as expected . Not all Tg-miRNA , however , were found associated to polysomes , and this was notably the case of Tg-miR-4 ( Figure 4E ) . Tg-miR-4 and other non-polysomal miRNA may regulate target mRNAs at later stages of parasite differentiation ( e . g . bradyzoite ) ; alternatively , they might not be involved in translation control ( data not shown ) , as has been recently shown for a class of Arabidopsis miRNA [41] that use cleavage-competent and/or cleavage-resistant target sites found in specific non-coding RNAs to initiate the production of trans-acting ( tasi ) RNAs via the action of RDR6 [42] . Whether tasiRNA exist in Toxoplasma is an interesting question for future experiments . In any case , our findings demonstrate that several Tg-miRNAs are present in the form of miRNPs in polyribosome-containing fractions where they are likely to negatively regulate translation of target transcripts . miRNAs orchestrate many biological functions and are notably involved in cell fate determination and/or integration of developmental or external stimuli . Plant and metazoan miRNA expression may thus vary greatly depending on growth conditions , changes in developmental stages or , in the case of parasites , changes in virulence . To investigate if , similarly , Tg-miRNA accumulation/processing is regulated differentially in Toxoplasma , we sampled miRNA from freshly egressed ( E ) or intracellular ( I ) parasites , as well as from three classical Toxoplasma isotypes . These isotypes are representative of the European and North American parasite population and correspond to three clonal lineages , designated type I , II and III , corresponding to reference strains RH , PRU and CTG , respectively [43] . These genotypes display contrasted virulence in mice: type I strains are lethal , whereas type II and III strains are hypo-virulent and typically establish chronic infections . There are additional phenotypic differences in migration , growth rate , and ability to convert from tachyzoite to the cyst-forming bradyzoite stage , notably [43] . For several of the Tg-miR detected by Northern analysis , we observed differential abundances between the Toxoplasma strains ( Figure 4A and 4C ) . Thus , normalized to the Tg-tRNAAla signal , the miR-4 signal was six fold greater in type I than it was in types II and III ( Figure 4A ) . Likewise Tg-miR-4 , -49 and -60 were more abundant in type I strain , whereas Tg-miR-40 and -56 were clearly more abundant in type II . Further investigation of these variations in miR-56 levels showed that they were attributable to differences in miRNA processing rather than transcription , because similar levels of pre-miR-56 were observed among the three isotypes , in Northern analyses ( Figure 4C ) . This result reinforces the growing view that MIRNA genes can undergo extensive post-transcriptional regulation through mechanisms that selectively affect pri-miRNA processing and/or pre-miRNA stabilization [44] , as uncovered recently with interactions involving murine pre-Let-7 , lin-28 [45] , [46] and the RNA-binding protein KSRP [47] , a homolog of which was indeed found associated with Tg-AGO ( see following sections ) . These observations thus extend this concept to a single cell parasite; given the overall low genetic diversity among Toxoplasma isotypes [43] , they further suggest that differential regulations of pathogen's miRNA repertoires might , indeed , influence virulence . Analyses of Tg-miR accumulation between freshly egressed ( E ) and intracellular ( I ) Toxoplasma revealed additional scope for modulation of mature miRNA levels between the two parasitic states . For instance , there was a clear mobility shift with miR-43 , which is unlikely explained by changes in pre-miR-43 steady levels , but rather , by alternative Dicer-mediated processing events producing small RNA length variants or with modified termini ( Figure 4B ) . Collectively , these observations unravel highly complex regulations of Tg-MIRNA gene expression , which might be used to refine the amplitude or regulatory outputs of target gene regulation during the parasite's multiple biological states . We then attempted to identify putative targets for representative members of the 14 unambiguous Tg-MIRNA families retrieved in this study . A bioinformatics approach was used to scan Toxoplasma transcripts for Tg-miRNA complementarity sites . Despite the resemblance of Tg-MIRNA and mammalian MIRNA genes , and the association of both types of molecules to polysomes , the absence of slicer residues in the Tg-AGO protein strongly suggests that the parasites' miRNA engage into a distinct type of pairing to their targets ( Figure 2B ) . Hence , in the mammalian host , most AGO2-bound miRNAs exhibit only moderate pairing to their targets , notably through a stretch of 6–7 contiguous 5′ nucleotides known as ‘seed’ , which is usually followed by several central mismatches that sterically hinder the RNAseH activity of AGO2 [48] , [49] . This loose miRNA:target pairing , which is thought to favor translational repression over slicing , makes it difficult to predict mammalian miRNA targets using computer algorithms . These algorithms , moreover , are often biased towards 3′ UTRs , because these regions evolve much more rapidly than coding regions , and are , therefore , more prone to the identification of contiguous , 6–7nt seed-complementary sequences [50] . We found that most of the 14 Tg-miRNA analyzed have readily identifiable target sites in a variety of cellular transcripts ( Table S2 and data not shown ) . Interestingly , these sites exhibit complete to near-complete complementarity to miRNAs -a feature of plant but not of metazoan miRNAs- and they are found in 5′-UTR , coding region and 3′-UTR , although there is a clear bias towards the latter region for most miRNA analyzed ( Figure S16A and Table S2 ) . Allowing up to 3 mismatches , more than 80 putative target transcripts were identified for miR-60a alone , the most abundantly sequenced Tg-miRNA . Using the same stringent parameters , an average of 25 cellular targets could be retrieved for each of the 14 Tg-miRNA ( Figure S16A and Table S2 ) . GO-term analysis of the putative Tg-miRNA target transcripts showed that they encompass virtually all known biological functions , with a somewhat stronger emphasis on translational control and cell cycle regulation , which might be expected for a single-celled , highly dividing parasite ( Figure S16B ) . These predicted Toxoplasma miRNA:target interactions thus constitute an unprecedented situation in all eukaryotes studied so far , whereby a miRNA-loaded , slicer-deficient Ago ( see later in the text ) might regulate target gene expression , presumably at the translational level , through perfect or near-perfect binding sites that are predominantly –albeit not exclusively- located in 3′-UTRs . To test the possibility of target cleavage and degradation mediated by Tg-AGO , we examined the levels of mRNA predicted as strong targets of isotype-specific Tg-miRNAs ( Table S2 ) . Real time PCR analyses revealed little , if any changes in mRNA levels between type I and II isotypes , contrasting with the differential abundances of the corresponding Tg-miRNAs . This result corroborates partially the suggestion that Tg-AGO acts mainly as a translational regulator , which is also in agreement with the cellular factors found in association with Tg-AGO ( see later in the text ) . Owing to the lack of available antibodies for predicted targets and our current inability to generate Tg-Dicer or Tg-AGO mutants , experimental validation of the above hypothesis will be part of future experiments . To date , the use of RNAi for specific gene silencing has remained largely inconclusive in Toxoplasma . Many laboratories have attempted to use this tool to down-regulate gene expression but very few reports showed successful double-stranded RNA induced gene silencing and there is currently no evidence for the production of specific siRNA [51] . We note that the use of RNAi is normally expected to result in mRNA turnover , as in metazoans or plants . The nature of the Tg-AGO ( slicer deficient ) and its possible mode of operation through translational repression ( with a usually modest output on gene expression in metazoans ) is obviously one parameter that could explain the lack of significant levels of mRNA degradation upon RNAi treatments in this organism . The largest bulk of medium-to-low abundance Tg-sRNAs does not meet the criteria of miRNA annotation and appears to match repetitive elements REP1 , REP2 and REP3 ( Figure 5A ) [52] . REP elements are mitochondrial-like sequences dispersed throughout the nuclear genome of Toxoplasma . They are typically composed of mitochondrial-like genes , including COX1 ( cytochrome oxidase subunit 1 ) and COB ( apocytochrome b ) that are flanked by a 91 bp short-dispersed repetitive sequence ( SDR ) organized as a direct or inverted repeat ( Figure 5A ) that might play roles in generation or dispersal of the REP elements . Nonetheless , there is no sequence similarity between SDRs and other terminal repeats such as those of retroviral LTRs . Moreover , REP elements do not seem to be highly mobile [52] . Genome mapping showed that Toxoplasma REP-derived sRNAs ( rdsRNAs ) form discrete species that are exclusively generated from regions located downstream of the COX1 and COB sequences , ( Figure 5A ) , with read counts typically ranging from >10 , 000 ( rdsRNA-17 ) to a few hundred reads ( rdsRNA-28 ) . This fairly high abundance might be explained by the fact that the estimated number of REP elements is >500 copies per genome [52] . While their size range ( 21–27nt , Table S3 ) and sensitivity to periodate ( not shown ) was similar to that of Tg-miRNAs , about half of the rdsRNAs had a 5′ terminal U instead of the prevalent A found in miRNAs . All tested Tg-rdsRNAs were readily detected by Northern blotting using 5′end-labeled antisense oligonucleotides but , unlike Tg-miRNAs , they were consistently much more abundant in the highly virulent Toxoplasma isotype-I ( Figure 5B ) . Interestingly , sense probes generated no signal for the Tg-rdsRNAs tested , consistent with the absence of Illumina reads corresponding to opposite-strand sRNA species ( Figure 5C and data not shown ) . A sense probe for Tg-rdsRNA-20 , however , gave a high molecular weight signal potentially resulting from hybridization of a double-stranded RNA precursor , although RNA folding algorithms did not reveal any significant secondary structure at , or in the vicinity of Tg-rdsRNA sequence matches ( Figure 5C ) . Nonetheless , the detection of identical rdsRNA species in the related apicomplexan N . caninum and their isotype-specific accumulation ( Figure 5B ) together with their abundant loading into Tg-AGO ( see below ) make it unlikely that these species are simply random degradation products . This rather suggests the existence of a conserved mechanism that accounts for REP-dependent production of precursor molecules required for rdsRNA synthesis . A second class of low abundant Tg-rdsRNAs mapped directly to a long , imperfect stem-loop structure resulting from annealing of an individual ‘solo SDR’ unit . This structure is depicted in Figure 5D , together with the cloned sequences of contiguous or overlapping sRNAs that are likely produced via stepwise processing by the Tg-Dicer . Such imperfect structures might well represent the equivalent of the plant proto-MIRNA genes that arise from DNA-type non-autonomous elements known as miniature inverted-repeat transposable elements ( MITEs ) . MITEs readily fold into imperfect stem-loops typical of miRNA precursors [53] , [54] and often generate multiple sRNA species , including heterochromatic siRNA that dampen MITE expression transcriptionally , as well as recently-evolved ( or young ) miRNAs that may not have yet undergone positive selection for host transcript targeting , and tend to accumulate at low levels , as seen here with the SDR-derived Tg-rdsRNAs . Sequence analysis also revealed the existence of a third class of repeat-associated sRNAs in Toxoplasma , which map perfectly to high-copy-number ( >800 copies per genome ) satellite DNA Sat350 ( ABGTg/TGR family , Figure 6A ) and Sat529a ( Figure 6B ) [55] . Although these satellite-associated ( Tg-sat ) RNAs had very low read numbers , they formed near-contiguous stretches of sequence along the corresponding SAT loci . These two features ( low read-number , accumulation as populations rather than discrete species ) are highly reminiscent of plant heterochromatic siRNAs found at DNA repeats and transposon loci with no intrinsic potential to form fold-back structures . In Arabidopsis , heterochromatic siRNAs are typically synthesized thought the conversion of aberrant RNA molecules into long-dsRNA , via the action of RDR2 [56] . Upon its processing by DCL3 , the resulting siRNA population engages into AGO4 or AGO6 to mediate cytosine methylation and histone modifications at the sites of its production , resulting in heterochromatin formation [57] . We speculate that , similarly , Tg-satRNAs originate from the action of Tg-RDR using SAT-derived aberrant transcripts as templates , and contribute to maintain the heterochromatic state found at both SAT350 and SAT529 , which , indeed , are enriched in silent chromatin marks including H4K20 and/or H3K9 monomethylations , as assessed by chromatin immunoprecipitation ( Figure 6C ) . Similarly these silencing marks are also poorly but clearly enriched at REP- and MITE-derived sRNA loci ( data not shown ) . We acknowledge that ChIP experiments for histone modifications only provide merely correlative evidence for a functional link between heterochromatin formation/spread and small RNA in Toxoplasma although some of the Tg-AGO-associated factors also support this idea ( see later in the text ) . Assessing the formal contribution of Tg-AGO in DNA-based heterochromatic processes will require further experiments . In the absence of obvious , additional Ago-like proteins ( including PIWI proteins ) in the Toxoplasma genome ( TOXODB , release v6 . 0 ) , both transcriptional and post-transcriptional gene silencing events must , therefore , be operated via the same and unique Tg-AGO . To address this issue , we generated transgenic parasites expressing ectopically HAFlag-tagged , full-length Tg-AGO . RNP complexes were immuno-affinity purified ( see next section ) , co-precipitated RNAs were extracted from the beads and analyzed by Northern using oligonucleotide probes specific to some of the highly abundant Tg-miRNAs and Tg-rdsRNAs studied above . Tg-miR-4 and -43 , and as well Tg-rdsRNA-17 and -28 were indeed detected in the HaFlag-Tg-AGO immuno-precipitates but not in control immuno-precipitates ( Figure 7A ) , indicating that Tg-AGO is a common effector of both types of sRNAs . This likely entails both cytoplasmic and nuclear distribution of the protein . Immunofluorescence and confocal microscopy revealed that Tg-AGO accumulates in tachyzoites mostly as granules of unidentified nature , but this labeling was superimposed over a diffuse cytoplasmic signal ( Figure 7B and data not shown ) . Using acetylated histone H4 as a marker , confocal analyses also revealed a faint nuclear staining indicating that a minor portion of Tg-AGO localizes to the nucleus . Nuclear localization of Tg-AGO could be transient or highly dynamic , and under steady-state conditions . Alternatively , nuclear Tg-AGO could be incorporated into large protein complexes that prevent its optimal accessibility to antibodies . The fact that a fraction of several Tg-miRNAs co-sediments with polysomes ( Figure 4E , fractions 13–18 ) suggests that a portion of Tg-AGO should also be associated with polysomes , as has been shown for plant and metazoan miRNA-loaded AGOs [41] , [58]–[62] . We thus examined the distribution of HaFlag-Tg-AGO using polysome gradients: cytoplasmic extracts from intracellular parasites were prepared and fractionated on sucrose gradients ( Figure 7C ) . The absorbance profiles at 254 nm showed a pattern of ribosomes with ribosomal subunits , monosomes , and polysomes . Consistent with previous findings in metazoans [61] , [62] , most HAFlag-Tg-AGO was found near the top of the gradient , where soluble material and small ribonucleoprotein particles sediment . Some HAFlag-Tg-AGO was also heterodispersed throughout the gradient fractions , where polyribosomes and Tg-miRNAs co-sediment ( Figures 4E and 7C ) . Treatments of cellular extracts with 30 mM EDTA or RNase T1 , known to dissociate polysomes into ribosomal subunits and monosomes , caused a shift in HAFlag-Tg-AGO distribution from the denser fractions to the lighter fractions of the gradient ( Figure 7C ) . This result suggests that a portion of miRNA-loaded Tg-AGO associates with polysomes to regulate translation of Tg-miRNA target mRNAs , perhaps in the cytoplasmic granules observed by immunofluorescence . As noted previously , a characteristic feature of Tg-AGO is the presence , at the amino terminus , of a repeated RGG-rich region ( amino acids 1–68 ) , in which the arginine residues have the potential to undergo methylation ( Figure 2B ) . This post-translational modification is known to influence the stability , activity and/or sub-cellular distribution of some metazoan AGO-like proteins [23]–[25] . Tudor-domain proteins specifically recognize symmetrically dimethylated arginines ( sDMA ) such as those found in AGO-like proteins [63] , [64] . Accordingly , the immunopurified HAFlag-Tg-AGO complex ( see below; Tables 1 and S4 ) was found to contain the Tudor-SN ( tudor staphylococcal nuclease ) /p100 homolog . Tudor-SN has five staphylococcal/micrococcal nuclease domains as well as Tudor domain , and it was described as a component of RISC in Caenorhabditis elegans , Drosophila and mammals [65]; more recently , the Tudor domain of the fly Tudor-SN was characterized as a specific sDMA-binding protein [64] . Reciprocal immunoprecipitation experiments further confirmed the specific binding of Tg-AGO to Tg-Tudor/SN ( Figure 7D ) . Furthermore , HAFlag-Tg-AGO was also found to co-purify with Tg-PRMT1 , which belongs to the family of arginine methyltransferases that use RGG motifs as substrates ( Table 1 ) . These results suggest that Tg-AGO is arginine-methylated , and that this modification might be specifically read by Tg-Tudor/SN , possibly to engage Tg-AGO into distinct modes of RNA silencing . In particular , the RGG-rich region of the Trypanosoma brucei Tb-AGO1 was found critical to its association with polysomes [66] . To test if the same was true of Tg-AGO , we engineered HAFlag-Tg-AGODRGG , which carries a deletion of the RGG domain ( amino acids 1 to 68 ) . While HAFlag-Tg-AGODRGG was loaded normally with Tg-miRNA and Tg-rdsRNA ( Figure 7A ) , the majority of the mutant protein was found near the top of polysome gradients , and was notably absent in fractions where polyribosomes sediment ( Figure 7C ) . In addition , mass spectrometry analysis of the HAFlag-Tg-AGODRGG complex showed that it was no longer associated to Tg-Tudor/SN ( data not shown ) . Thus , the proposed Tg-AGO arginine-methylation and association with Tg-Tudor/SN might allow post-loading sorting of distinct Tg-AGO-containing RNP complexes towards specific silencing modes . We note that its association to Tg-Tudor/SN through an RGG domain together with its predominant cytoplasmic localization evoke the as yet unexplored possibility that Tg-AGO may serve as a PIWI protein . Thus , in addition to its possible role in heterochromatin formation , Tg-AGO might contribute to post-transcriptional gene silencing of repeats and transposons via Tg-rdsRNAs , as is seen with metazoan PIWI proteins . To characterize the molecular composition of Tg-AGO-containing complexes , HAFlag-Tg-AGO was affinity purified from total cell extracts of intracellular tachyzoite by incubation with anti-FLAG agarose beads . The immunoprecipitated FLAG–protein complexes were eluted using the FLAG peptide . The immunoprecipitated proteins were then separated by SDS-PAGE , excised , and identified by mass spectrometry ( Figure 8A ) . Tables 1 and S4 list the names of the identified proteins , which , remarkably , were direct orthologs of nearly all of the previously identified components of human and Drosophila miRNA-RISC [39] , [40] , [67]–[69] . These co-purified proteins fell within several functional groups . The largest group encompasses mRNA-binding proteins , in particular the heterologous nuclear ribonucleoproteins: HNRNPA3 , HNRNPH1 , HNRNPL , and HNRNPM [39] , [40] . Several mRNA-binding proteins with putative functions in mRNA transport , stabilization and translation were also identified , including homologs of FUBP2/KSRP , nucleolin and FXR-related proteins , which are well known human and Drosophila Argonaute interactors [39] , [40] , [68] . Among the DEAD/DEAH box helicases , we found DDX17/DDX5 , an ortholog of Drosophila p68 , which has been shown to associate with Drosophila Ago2 [9] , and DDX3X/Belle or DDX6/p54 , which are all required for miRNA function ( Table 1 ) [67] , [68] . Consistent with the hypothesis that Tg-AGO associates with mRNPs , a homolog of the polyadenylation binding protein PABPC [70] was identified in the immuno-precipitate , indicating that mRNAs were present in the purifications . Accordingly , treatment of the lysate with RNase T1 prior to immuno-precipitation abolished the integrity of the Tg-AGO1 complex , indicating that the interactions between HAFlag-Tg-AGO and several proteins were RNA-mediated ( Figure 8B ) . Identification of translation initiation and elongation factors , together with various 40S and 60S ribosomal proteins ( Table S4 ) provides further support to the idea that the miRNA-loaded Tg-AGO , which associates with polysomes , might prevent translation of target mRNAs . Another noticeable partner of Tg-AGO was a ortholog of human FUBP2 , also known as KHSRP/KSRP , which binds with high affinity to the terminal loop of some miRNA precursors and promotes their maturation [47] . Tg-KSRP might account , at least party , for the post-transcriptional regulation of some Toxoplasma MIRNA genes , as uncovered in this study ( Figure 4B and 4C ) . Consistent with an effect of Tg-KSRP on miRNA maturation rather than activity , Tg-KSRP was found to bind Tg-AGO in an RNA-independent manner ( Figure 8C ) . Accordingly , Tg-KSRP did not associate to polyribosomes ( Figure 8D ) and was also found in HAFlag-Tg-AGODRGG immuno-precipitates ( Figure 8C ) . Additional immuno-purified proteins are not obviously related to translational control but have been previously implicated as RISC-associated factors including Tg-Tudor/SN and Tg-PRMT1 , already evoked above . Notably , Tg-AGO also co-purified with a conserved 14-3-3 protein ( Table 1 ) : 14-3-3 proteins that bind S . pombe Ago1 and human Ago2 are probably required for AGO protein functions in cell cycle and/or gene silencing pathways [71] . 14-3-3 proteins may also act as major regulators for the sorting of AGOs between distinct classes of RNA granules [69] , which may include the Tg-AGO foci detected by immunofluorescence in the present study ( Figure 7B ) . Collectively , these results provide compelling evidence that Tg-AGO is part of a functional RISC whose core components are nearly all orthologous to factors required for post-transcriptional gene silencing and its regulation in metazoans . Consistent with additional , DNA-level silencing functions of Tg-AGO , the second largest subunit of Tg-RNA polymerase II ( Rpb2 ) also co-purified in the HAFlag-Tg-AGOFL immuno-precipitates ( Table 1 ) . Interestingly , a mutation of Rpb2 in fission yeast , rpb2-m203 , disrupts coupling between transcription and siRNA processing in RNAi-dependent heterochromatin formation [72] . Also reminiscent of the fission yeast heterochromatic RNAi pathway , HAFlag-Tg-AGO was associated with the histone deacetylase TgHDAC3 , a protein that may play similar roles to S . pombe Clr3 in the spread of heterochromatin [73] , [74] . Remarkably , HAFlag-Tg-AGO was associated with all known components of the major transcriptional co-repressor complex Tg-CRC [75] , [76] , which contains the two repressor proteins Tg-CRC230 and Tg-TBL1 , the catalytic subunit Tg-HDAC3 and a new plant-like AP2-domain transcription factor ( Table 1 ) . Moreover , peptide sequencing by tandem mass spectrometry indicated that the subunits of the complex are sub-stoichiometrically represented . This finding is consistent with the as yet unconfirmed idea that Tg-AGO-bound rdsRNA and possibly satRNAs may guide transcriptional gene silencing processes by recruiting histone deacetylases , and subsequently histone methylases ( i . e . Tg-SET8 and Tg-SET3 , [77] ) , to heterochromatic regions of the genome . The present analysis thus uncovers an unsuspected level of complexity in the RNA silencing pathways of the single cell parasite T . gondii . This complexity not only lies in the mere diversity of the sRNAs identified , but also in the apparent mix-and-matched nature of the silencing components found in this organism , both in terms of their evolution and function . In this respect , the T . gondii RNA silencing machinery and its usage by the parasite bewilder many accepted notions in the field . For instance , in no organism studied so far has a single Ago protein evolved to mediate both repeat-associated and miRNA-mediated gene silencing , two pathways usually considered drastically different . Likewise , the metazoan-like Tg-miRNAs have readily identifiable mRNA targets displaying perfect to near-perfect complementarity in both CDS and UTRs , which is unprecedented in animals . Further studies of the Toxoplasma RNA silencing pathways will undoubtedly reveal other surprises and , more importantly , might shed light on the molecular bases of virulence in this important Human parasite .
The parasite strains used in this study are the following: the T . gondii type I RH strain that has lost the ability to complete the two-host life cycle , the T . gondii type II Prugniaud strain that is capable of robust bradyzoite differentiation , and the T . gondii type III CTG and C56 strains . All T . gondii and N . caninum strains were maintained by serial passage in HFF monolayer under tachyzoite conditions in DMEM ( Invitrogen ) supplemented with 10% ( vol/vol ) FBS ( Invitrogen ) . T . gondii type II Prugniaud strain was maintained under tachyzoite conditions in DMEM supplemented with 10% ( vol/vol ) FBS and 25 mM Hepes buffer , pH 7 . 2 . To induce in vitro bradyzoite differentiation , extracellular tachyzoites were allowed to invade HFF cells for 16 hours , and the culture medium was removed and replaced by RPMI-1640 supplemented with 1% FBS and 50 mM Hepes buffer , pH 8 . 2 . After 2–3 days of culture in alkaline medium , bradyzoite induction was assessed for P36 and SUMO expression by IFA as described previously [78] . The RHhxgprt- strain used in these studies contains a deleted or defective HXGPRT gene , which allows for the selection of transfected tachyzoites using mycophenolic acid . Antisera against Tg-FUBP2/KSRP ( 35 . m00901 gene ) were produced by Eurogentec using the ‘Super Speedy immunization’ protocol and the following peptides Tg-FUBP2-1 ( H2N-MARKKRGSAATPEEGC-CONH2 ) and Tg-FUBP2-2 ( H2N-GTDKREDRGVTPEE DC-CONH2 ) . Specific antibodies were affinity purified against both peptides . For immunoblot analysis purified antibodies were used at 1∶1000 dilutions . Primary antibodies for IFA , ChIP and Western blot included antibodies against haemagglutinin epitope tag ( HA , Roche Diagnostic , dilution at 1∶1000 ) , Polyclonal anti-H4-K20-1me ( Abcam ab9051 ) , Polyclonal anti-H4-K20-3me ( Abcam ab9053 ) , Polyclonal anti-H4-K20-1me ( gift from Rice JC , Sims et al . , 2006 ) , Polyclonal anti-H4 Acetylated ( K5-K8-K12-K16 ) ( upstate 06-866 ) , Anti-H3-K9-1me ( upstate 07-450 ) , Anti-H3-K9-2me ( upstate 07-441 ) , Anti-H3-K9-3me ( upstate 07-442 ) , Monoclonal anti-Myc ( 9E10 - sc40X , Santa-Cruz Bio . ) . Infected HFFs grown on coverslips were washed in PBS and fixed/permeabilized for 20 min at room temperature with PBS containing 3% ( vol/vol ) formaldehyde and 0 . 2% Triton X-100 ( vol/vol ) . Blocking was performed with PBS containing 5% FBS and 5% goat serum for 1 h at room temperature . Samples were incubated in PBS containing 1% FBS with the primary antibodies , followed by the secondary antibodies goat anti–mouse IgG coupled with Alexa Fluor 488 and goat anti–rabbit IgG coupled with Alexa Fluor 568 ( Invitrogen ) at a 1∶1 , 000 dilution each in PBS–1% FBS . Nuclei of host cells and parasites were stained for 10 min at room temperature with Hoechst 33258 at 2 µg/ml in PBS . After four washes in PBS , coverslips were mounted on a glass slide with Mowiol mounting medium ( 48 mM Tris-HCl [pH 8 . 5] , 4 . 8% Mowiol 4–88 [wt/vol] , 12% glycerol [vol/vol] ) , and images were acquired with a fluorescence microscope ( Axioplan 2; Carl Zeiss , Inc . ) . QChIP assays were performed based on a modification of previously published methods [77] , [79] . Immuno-precipitated DNA were purified through PCR Purification Kit columns ( QIAGEN ) and used as a template in semiquantitative QPCRs to detect specific targets . Specific primer pairs ( melting temperature , 55 to 65°C ) amplifying 200- to 450-bp fragments were used ( supplemental Table S1 ) . PCR was performed with 1 µL of DNA and 500 nM primers diluted to a final volume of 20 µL in SYBR Green Reaction Mix ( Roche ) . Accumulation of fluorescent products was monitored by real-time PCR using a LightCycler 2 . 0 ( Roche ) . Each PCR reaction generated only the expected specific amplicon , as shown by the melting-temperature profiles of final products ( dissociation curve , automatically measured by the LightCycler 2 . 0 ) and by gel electrophoresis of test PCR reactions . No PCR products were observed in the absence of template . The fold difference of a given target sequence precipitated by a specific antibody was determined by dividing the amount of target sequence in the immunoprecipitate fraction by the amount of target sequence in input DNA ( S8 , S13 ) . Real-time PCR was carried out in triplicate on 2 ng of DNA at 50°C for 2 min and 95°C for 10 min , followed by 40 cycles of 95°C for 15 s and 60°C for 1 min . Data were collected at 60°C . The concentration of primers and Taqman probes used was determined by following the optimization procedure described in PE Applied Biosystem's protocol . For each experiment , the threshold was set to cross a point at which real-time PCR amplification was linear . For the majority of the experiments , data were analyzed with a threshold of 0 . 05 . Data collected was analyzed and plotted using Microsoft Excel . Satellite 350B: OL17 ( CGACTCGGACGTCAGGCCATGCAGAG ) and OL18 ( GCGCCTGAACAATACGCCCAACC ) . Satellite 529A: OL19 ( CTGCAGGGAGGAAGACGAAAGTTG ) and OL20 ( CTGCAGACACAGTGCATCTGGATT ) . Whole-cell extract ( WCE ) from transgenic intracellular tachyzoites expressing ectopically HAFlag-TgAGOFL and HAFlag-TgAGODRGG was incubated with 500 ml of anti-FLAG M2 affinity gel ( Sigma ) for 1 h at 4°C . Beads were washed with 10 column volumes of BC500 buffer [20 mM Tris ( pH 8 ) , 0 . 5 M KCl , 10% glycerol , 1 mM EDTA , 1 mM DTT , 0 . 1% NP40 , 0 . 5 mM PMSF , aprotinin , leupeptide , pepstatin , 1 ug ml−1 each] . Bound peptides were eluted stepwise with 250 ug ml−1 FLAG peptide ( Sigma ) diluted in BC500 buffer . Each preparation was sufficiently clean such that individual peptide bands could be excised and sequenced by mass spectrometry . Protein bands were excised from colloidal blue-stained gels ( Invitrogen ) , oxidized with 7% H2O2 and subjected to in-gel tryptic digestion . Peptides were extracted with 5% [v/v] formic acid solution and acetonitrile , and injected into an Ultimate 3000 ( Dionex ) nanoLC system that was directly coupled to a LTQ-Orbitrap mass spectrometer ( Thermo Fisher Scientific ) . MS and MS/MS data were acquired using Xcalibur ( Thermo Fischer Scientific ) and processed automatically using Mascot Daemon software ( Matrix Science ) . Tandem mass spectra were searched against a compiled T . gondii database using the MASCOT program ( Matrix Sciences , London ) available via intranet . RNA was extracted into TRIzol ( Invitrogen ) , and aliquots of total RNA from RH strain were subjected to small RNA library construction as follows . To avoid any contamination with the host cell small RNAs , freshly released parasites were harvested from the culture supernatant , washed by centrifugation , and filtered through a 3-µm filter before use . For each library , 50 µg of total RNA was size fractionated on a 15% tris-borate-EDTA ( TBE ) urea polyacrylamide gel ( Invitrogen ) and a 19–40 base pair fraction was excised . RNA was eluted from the polyacrylamide gel slice in 300 µL of 0 . 3 M NaCl overnight at 4°C . The resulting gel slurry was passed through a Spin-X cellulose acetate filter column ( Corning Inc . ) and precipitated by the addition of 750 µL of ethanol and 3 µL of glycogen ( 5 mg/mL; Ambion ) . After washing with 75% ethanol , the pellets were allowed to air dry at 25°C and pooled in diethylpyrocarbonate ( DEPC ) -treated water . The 5′ RNA adapter ( 5′-GUUCAGAGUUCUACAGUCCGACGAUC-3′ ) was ligated to the RNA pool with T4 RNA ligase ( Promega ) in the presence of RNase Out ( Invitrogen ) 6 hours at 20°C . The ligation reaction was stopped by the addition of 2× Gel Loading Buffer II ( Ambion ) . The ligated RNA was size fractionated on a Novex 15% TBE urea polyacrylamide gel ( Invitrogen ) , and a 40–70 base pair fraction was excised . RNA was eluted from the polyacrylamide gel slice in 300 µL of 0 . 3 M NaCl overnight at 4°C . The RNA was eluted from the gel and precipitated as described above followed by resuspension in DEPC-treated water . The 3′ RNA adapter ( 5′-pUCGUAUGCCGUCUUCUGCUUGUidT-3′; p , phosphate; idT , inverted deoxythymidine ) was subsequently ligated to the precipitated RNA with T4 RNA ligase ( Ambion ) in the presence of RNase Out ( Invitrogen ) 6 hours at 20°C . The ligation reaction was stopped by the addition of 2× Gel Loading Buffer II ( Ambion ) . Ligated RNA was size fractionated on a Novex 10% TBE urea polyacrylamide gel ( Invitrogen ) , and the 70–100 base pair fraction was excised . The RNA was eluted from the polyacrylamide gel and precipitated from the gel as described above and resuspended in 4 . 5 µL of DEPC-treated water . The RNA was converted to single-stranded cDNA using Superscript II reverse transcriptase ( Invitrogen ) and Illumina's small RNA RT-Primer ( 5′-CAAGCAGAAGACGGCATACGA-3′ ) following the manufacturer's instructions . The resulting cDNA was PCR-amplified with Phusion™ High Fidelity DNA Polymerase ( NEB ) in 15 cycles using Illumina's small RNA primer set ( 5′-CAAGCAGAAGACGGCATACGA-3′; 5′- AATGATACGGCGACCACCGACAGGTTCAGAGTTCTACAGTCCGA-3′ ) . PCR products were purified on a Novex 6% TBE PAGE gel ( Invitrogen ) and the 100 base pair fraction was excised . The DNA was eluted into 100 µL of 1x NEBuffer 2 at room temperature for 2 hours . The resulting gel slurry was passed through a Spin-X filter ( Corning ) and precipitated by the addition of 325 µL of ethanol , 10 µL of 3 M sodium acetate , and 3 µL of glycogen ( 5 mg/mL; Ambion ) . After washing with 75% ethanol , the pellet was allowed to air dry at 25°C and dissolved in 10 µL of resuspension buffer ( 10 mM Tris-HCl , pH 8 . 5 ) . The purified PCR products were quantified on the Agilent DNA 1000 chip and diluted to 10 nM for sequencing on the Illumina 1G ( GATC BIOTECH , Konstanz , Germany ) . RNA from Toxoplasma and Neospora strains was extracted into TRIzol ( Invitrogen ) , deproteinized with phenol chloroform/isoamyl alcohol , and RNA was recovered by ethanol precipitation . For small RNA analyses , 30 ug of purified RNA were separated on a 15% polyacrylamide ( w/v ) 8 M urea gel and transferred to GeneScreen nylon membranes . DNA oligonucleotides complementary to tg-miRNAs or tg-rasiRNAs were labeled with [g-32P]ATP using T4 PNK ( Promega ) . Hybridizations were performed at 37°C overnight . Hybridized membranes were exposed to imaging plates that were recorded after 5 h ( PhosphoImager , FLA-8000 , Fuji ) . Whole cell extracts were prepared as described previously [59] but using a modified polysome buffer containing 100 mM NaCl , 40 mM Tris-Hcl pH 7 , 10 mM MgCl2 , 1 mM DTT , 1% Triton TX100 and protease inhibitor ( Complete ) . Cycloheximide ( 100 ug/ml ) was added to cells ( 10 min at 37°C ) prior to collecting the cells by centrifugation . The drug was present in all buffers throughout the entire procedure . For polysome fractionation experiment , approximately 1000 OD600 of whole cell extract were layered onto 12 ml 5%–40% sucrose gradients prepared in polysome buffer without Triton , and centrifuged at 4°C for 2h at 36 , 000 rpm in a Beckman SW41 rotor . After centrifugation , 500 ml fractions were collected from the top of the gradient and the 260-nm absorbance profile was recorded . For Northern-Blot analysis , RNAs from each fraction were precipitated by adding 0 , 1M NaCl and 3 volumes ethanol and extracted with the TRIzol method . For Western-Blot analysis , 15 ul of each gradient fraction were run on 12% SDS-PAGE gels . Total RNAs ( 30 ug ) from RH-infected fibroblasts were incubated in a solution containing 10 mM HEPES ( pH 7 . 0 ) and 250 mM sodium periodate for 30 min at 22°C . An equal volume of formamide loading dye was added to the samples , followed by incubation for 45 min at 99°C . The reaction mixture was then analysed by Northern-Blot . An equal amount of untreated RNA was also loaded onto the gel for comparison . RNA oligos of known sequence ( 19 , 21 and 23 nt synthetic RNA oligos ) were also treated with sodium periodate to check for the completion of b-elimination reaction , and the blots were probed with end-labeled oligos complementary to the synthetic oligos . We analysed a pool of 5 , 701 , 506 raw reads obtained by the sequencing-by-synthesis ( Illumina ) . Initially , all the sequences fully matching tRNAs or rRNAs were removed . The remaining sequences were used to build a local Mysql database and then trimmed in a 4 steps process: 1 ) We removed 3′ and 5′ adaptor sequences , using an iterative scheme and updated the database . 2 ) We removed from each sequence the nucleotides , in the 5′ and 3′ extremities , with a Phred Quality Score ( http://www . phrap . com/phred/ ) below 10 and updated the database . 3 ) We eliminated all the reads containing more than 6 stretches of C , T or G . 4 ) We screened the database to keep only reads having a length >19 nucleotides and an average Phred Quality Score >15 . The final pool of 1 , 555 , 290 reads was used to cluster small RNAs . The sequences without any variation were classified in the same cluster . A total of 275 , 888 distinct clusters were identified and used for further analysis . All these clusters were compared against the T . gondii genome ( http://www . ToxoDB . org , version 4 . 3 ) using Blast program with specific parameters ( Word size set to 4 and Penalty for a nucleotide mismatch set to -1 ) . This configuration allowed a more refined search of small versus large nucleotide sequences . All the results were saved in the database and used for mapping sRNAs on the chromosomes . Based on the chromosome position , we classify the clusters into families . First , we seeded our classification with results having 100% identity and an alignment length >19 and then recovered the clusters varying with less than three nucleotides . All further analyses were focused on the most abundant families . Images , multi-fastas and alignments against T . gondii genome were automatically generated and manually curated . All these treatments were made using an in-house API ( Genobrowser ) and functionalities ( tools unpublished ) written in PHP . Up to 10 sequence windows on both strands , spanning the locus and including variable lengths of flanking regions ( 5–200 bp on either side ) , were examined for their potential to form fold-back transcripts by using the RNAfold [80] and Mfold [81] programs . The predicted outcomes , including the minimal folding free energy ( MFE ) , at least 20 kcal/mole ( dG = −20 kcal/mole ) , the length of pre-miRNAs , and the number of nucleotides ( A , C , G , or U ) in each pre-miRNA were recorded and used for further analysis . Each predicted tg-microRNA was further checked manually to ensure that they were from good quality single-stranded hairpins and that a miRNA/miRNA* pair had 0–2 nt 3′ overhangs . The coding sequence of Tg-AGO ( genbank , GU046561 ) was amplified by RT-PCR to introduce BamHI and HindIII sites at the start and the stop codons respectively . Primers forward ( 5′-ggatccATGAACGGAGGAGGCAGAGGAAGAG-3′ ) and reverse ( 5′- aagcttCCATCAATGCTGTCTCAACAGAAC-3′ ) were used for PCR amplification . The PCR product allowed the cloning of Tg-AGOFL in frame with an N-terminal HAFlag tag into the T . gondii expression vector pMAH14 ( GRA1 promoter , [75] ) digest with BglII-HindIII . Tg-AGODRGG was amplified by PCR with forward ( 5′- ggatccCTGTACGATGGAGACCACCTTCTC-3′ ) and reverse ( 5′- aagcttCCATCAATGCTGTCTCAACAGAAC-3′ ) and cloned subsequently in pMAH14 ( BglII-HindIII ) .
|
Toxoplasma gondii is an important human parasite that causes life-threatening diseases in developing fetuses and in immunocompromised individuals , especially AIDS and transplant patients . Curiously , the Toxoplasma genome is deprived of most of the basic transcription factors that regulate gene expression in other eukaryotic cells . Therefore , alternative strategies must exist to modulate the many phases of the Toxoplasma complex life cycle that includes invasion of several hosts . Here , we investigate one of these strategies , by studying the repertoire of Toxoplasma silencing small RNAs ( sRNAs ) . In eukaryotes , most of these regulatory molecules , 20–30nt-long , are produced by members of the Dicer RNase-III family , and exert their various functions through ubiquitous proteins called Argonaute ( Ago ) . The surprising diversity of the Toxoplasma sRNAome uncovered in our study is consistent with those molecules exerting key functions during the parasite's life cycle , including , possibly , during virulent infection . The study also unravels an unsuspected level of complexity in the origin and mechanisms of action of the factors that generate and affect Toxoplasma sRNA , prompting a re-evaluation of our current views on RNA silencing in eukaryotes .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/post-translational",
"regulation",
"of",
"gene",
"expression",
"biochemistry/rna",
"structure",
"microbiology/parasitology",
"infectious",
"diseases/protozoal",
"infections",
"biochemistry/transcription",
"and",
"translation",
"genetics",
"and",
"genomics/epigenetics",
"molecular",
"biology/chromatin",
"structure"
] |
2010
|
A Complex Small RNA Repertoire Is Generated by a Plant/Fungal-Like Machinery and Effected by a Metazoan-Like Argonaute in the Single-Cell Human Parasite Toxoplasma gondii
|
Hybrid deterministic-stochastic methods provide an efficient alternative to a fully stochastic treatment of models which include components with disparate levels of stochasticity . However , general-purpose hybrid solvers for spatially resolved simulations of reaction-diffusion systems are not widely available . Here we describe fundamentals of a general-purpose spatial hybrid method . The method generates realizations of a spatially inhomogeneous hybrid system by appropriately integrating capabilities of a deterministic partial differential equation solver with a popular particle-based stochastic simulator , Smoldyn . Rigorous validation of the algorithm is detailed , using a simple model of calcium ‘sparks’ as a testbed . The solver is then applied to a deterministic-stochastic model of spontaneous emergence of cell polarity . The approach is general enough to be implemented within biologist-friendly software frameworks such as Virtual Cell .
It is not uncommon for a cell-biological model to include some components that might be stochastic in nature ( small copy numbers , rare events ) , whereas others , if uncoupled , would behave deterministically ( large copy numbers , fast reactions ) . Through their interaction , fluctuations in a stochastic subsystem may induce significant random perturbations in the ‘deterministic’ one , thus rendering the entire system stochastic . Calcium sparks in cardiomyocytes and other cells [1] is one such example , where calcium released from intracellular stores through calcium channels may , in turn , influence probabilities of opening and closing of those channels . Here , calcium concentration can often be regarded as a ‘deterministic’ component of the system , whereas the dynamics of calcium channels is inherently stochastic . Similarly , stochastic openings of voltage-sensitive ion channels depend on the ‘deterministic’ membrane potential , which , in turn , is affected by stochastic electric currents passing through the channels [2] . Stochasticity in otherwise deterministic cellular subsystems may also be brought about by their coupling to dynamics of cytoskeletal filaments , translation events , and other processes involving macromolecules and small organelles present in small numbers . Simulating such systems as fully stochastic can be prohibitively slow . Indeed , simulating calcium sparks stochastically with an account of every single calcium ion would be computationally expensive because their number is typically large . But in the limit of large copy numbers , the intrinsic fluctuations due to discreteness of molecules are insignificant , and one can design faster hybrid algorithms , in which deterministic and stochastic approaches are appropriately combined . While these efficient methods are approximate , the larger the copy numbers in the ‘deterministic’ subsystem , the more accurate their outcome . Numerical approaches to interacting systems with disparate levels of stochasticity are an area of active interdisciplinary research . In the context of cell-biological applications , various hybrid approaches were proposed for ‘well-mixed’ models of biochemical networks with fast and slow components [3 , 4] . In these models , a fast component , whose copy numbers are only moderately large , is often modeled as a Wiener stochastic process , rather than deterministically . The corresponding numerical techniques are a combination of methods of solving stochastic ordinary differential equations ( SDEs ) [5] , also termed Langevin equations in the physics literature , and Gillespie-type algorithms [6 , 7] that simulate stochastic reaction events in the slow component . Unlike stochastic hybrids , the deterministic-stochastic models are mathematically defined as piecewise deterministic Markov processes [8 , 9] , in which the system develops deterministically between consecutive stochastic events . Numerical approaches to such systems are based on a formulation that couples differential equations , which describe continuous variables , with equations that govern probability distributions of the stochastic components . The coupling occurs through ‘deterministic’ rates dependent on discrete stochastic variables and transition probability rates that are functions of continuous variables . Efficient numerical methods for solving deterministic-stochastic models rely on generating Monte Carlo realizations of a hybrid system . For this , a kinetic Monte Carlo algorithm advancing a stochastic subsystem in time must work in conjunction with a deterministic integrator that updates continuous variables by solving the corresponding differential equations . A variety of algorithms were proposed for spatially uniform , or well-mixed , deterministic-stochastic models . Fixed time step methods , applied to hybrid models of membrane potential [2] and calcium dynamics [10] , are conceptually straightforward but incur time-discretization errors in stochastic variables . In adaptive methods , which were first proposed for solving deterministic-stochastic models of biochemical networks [11] , the treatment of a stochastic subsystem is essentially free of time-discretization errors . In these algorithms , accurate sampling of stochastic reaction events coupled to continuous variables is achieved by adapting Gillespie’s methods for systems whose transition rates explicitly depend on time . A similar approach was used in a hybrid stochastic algorithm for well-mixed systems with fast and slow components [12] . It should be noted that in adaptive methods , special care is required for ensuring synchronous treatment of the ‘deterministic’ and stochastic subsystems . A rigorous convergence analysis of the hybrid adaptive methods was given in [13] . Numerical methods for spatially resolved deterministic-stochastic models are less common . A method described in [14] approximates a stochastic subsystem by a reaction-diffusion master equation [15–18] . In this approximation , a spatial domain is partitioned into subvolumes which are assumed to be well-mixed at any time , and a state of the stochastic subsystem is described in terms of copy numbers per subvolume . The master equation is then solved by an optimized variant of the Next Subvolume method [19] . Designed for models with relatively slow deterministic dynamics , the method of [14] is applicable only if the stochastic subsystems involve sufficiently large copy numbers per subvolume [20 , 21] . Stochastic subsystems with relatively low copy numbers can be described in terms of states and spatial locations of individual molecules . The particle-based approach was used to simulate a simple model of assembly of RNA granules in which RNA molecules bind to core complexes [22] . In the model , spatial distributions of RNA molecules were modeled deterministically by partial differential equations . The core complexes and RNA granules were treated stochastically as individual particles interacting with the deterministic subsystem while undergoing random walks . A similar approach was adopted in modeling actin bundles and asters [23 , 24] , where the stochastic subsystem was comprised of tips of actin filaments while ‘deterministic’ actin monomers were modeled as well-mixed because of their relatively fast diffusion . States and positions of individual channels were also used to define stochastic subsystems in spatial versions of the deterministic-stochastic models of membrane potential [25] and calcium release from inositol 1 , 4 , 5-trisphosphate ( InsP3 ) -receptor channels [26–28 , 2] . Algorithmically , the methods in these studies combined deterministic descriptions in terms of partial differential equations and the event-driven time stepping schemes [11] . Calcium-induced calcium release in cardiac muscle cells [29] was already mentioned above as a mechanism that naturally lends itself to a hybrid numerical treatment . Playing a key role in ensuring robustness of heart contractions in response to action potentials , it has been studied extensively by various methods [30] , including mathematical modeling [31] . The calcium release in cardiomyocytes occurs by way of clustered ryanodine receptor channels ( RyR ) and , in a healthy heart , takes the form of an avalanche of calcium ‘sparks’ , the localized spikes of calcium concentration [32] . Recent advances in experimental technologies have generated renewed interest in detailed predictive computational modeling of calcium dynamics in heart muscle cells for normal and pathological conditions [33 , 34] . Similar to calcium release from the ( InsP3 ) -receptor channels , the problem entails coupling of a spatial deterministic description of calcium and stochastic kinetics of RyR channels and can be solved efficiently by a hybrid numerical method . All of the above approaches were largely specific solutions to a specific modeling problem or a restricted domain of problems . In this article , we describe a general-purpose spatial deterministic-stochastic algorithm and discuss techniques used for its validation . The work was motivated by the need of providing tools for simulating spatial hybrid models to a wide range of cell scientists . The method is designed to be applicable to a broad spectrum of models , including those where continuous and discrete variables are defined both in volume and in the encompassing membranes . The current implementation of the method appropriately combines capabilities of one of the Virtual Cell ( VCell ) [35–39] spatial deterministic solvers and an efficient particle-based simulator called Smoldyn [40 , 41] . ( Note that Smoldyn has been recently adapted to accommodate a different type of hybrid stochastic models [42] , in which the subsystems with disparate levels of stochasticity are segregated in space but can interact in a ‘handshaking’ region [43–46] . ) The development of the VCell hybrid solver benefited from recent integration of Smoldyn into VCell as a method of solving spatial stochastic models [47] . A distinct feature of our hybrid solver is that the simulations of widespread fluctuations originating from point sources can be carried out in realistic geometries taken from experimental images , as both VCell and Smoldyn provide tools for simulating reaction-diffusion systems in arbitrary geometries [48 , 49 , 40] . This article is focused on physical underpinnings of the method and its algorithmic details , with special emphasis on rigorous validation of its key elements . Hybrid algorithms , often proposed heuristically , may appear intuitive , but their rigorous analysis and validation constitute a challenging task [12 , 25 , 7] . This is particularly true in the context of spatially resolved models . Tests against deterministic limits , while necessary , are insufficient because convergence to a correct deterministic limit does not yet guarantee correct behavior in the stochastic regime . Analytical solutions of stochastic models , required for convergence studies in the stochastic regime , are rare , particularly for spatial hybrid systems . In addition to truncation errors due to the time-space discretization , common to deterministic integrators , probability distributions and correlation functions obtained by Monte Carlo techniques include statistical errors due to finite numbers of realizations . Thus , the validation of a spatial hybrid solver entails analysis of multidimensional datasets representing multiple realizations of a hybrid system obtained with varying discretization parameters . The paper is organized as follows . The algorithm , along with its mathematical fundamentals , is described in Section Mathematical problem and algorithm using a simple model of calcium sparks as an example . It is then applied to two very different cell-biological phenomena . The calcium spark model introduced in Section Mathematical problem and algorithm is used in Section Validation of the method for validation of the method against analytical results and numerical solutions obtained by alternative methods . In Section Application to a hybrid model of spontaneous cell polarization , the method is applied to a hybrid model of spontaneous cell polarization; the actual VCell MathModel script for this application is included in S3 Text as an illustration of the software implementation . A summary of results and discussion of possible improvements conclude the paper .
Mathematically , the algorithm is based on a formulation of a deterministic-stochastic system , which is somewhat similar to how Wiener processes are described in terms of Langevin equations . To illustrate the approach and explain the workings of the algorithm , we employ a simple model of calcium sparks , whose ‘deterministic’ subsystem consists of a single variable , the calcium concentration U ( r , t ) , and its stochastic subsystem is comprised of calcium channels , through which calcium flows into the cell from intracellular stores . In muscle cells , calcium channels form small regularly distributed clusters . For simplicity , we will treat the calcium sources as single channels having two states , open and closed . The corresponding discrete stochastic variables are Ξi ( r , t ) ≡ δ ( r − ri ) ξi ( t ) , where the Dirac deltas δ ( r − ri ) define channel locations and the stochastic variables ξi ( t ) accept two values: 1 ( open state ) and 0 ( closed state ) . The index i enumerates the channels , and r , ri ∈ Ωcell , where Ωcell denotes the space of a cell . Dynamics of the continuous variable U ( r , t ) are affected by the following mechanisms: calcium release through channels , calcium diffusion , and removal of calcium from the cytosol via calcium pumps . The variable is therefore governed by a partial differential equation ( PDE ) with stochastic source terms , ∂tU=∇⋅ ( D∇U ) +J ( ∑i=1NchΞi ( r , t ) ) −Vp ( U−U0 ) , ( 1 ) where D is the calcium diffusion constant , J is the calcium flux through an open channel , Nch is the total number of channels in the cell , Vp is the calcium pump rate constant , and U0 is the steady-state calcium concentration in the absence of open channels . Eq ( 1 ) is subject to boundary conditions imposed at the cell membrane . For example , if calcium fluxes at the plasma membrane can be ignored , the corresponding no-flux boundary condition can be written as −D ( n⋅∇U ) |∂Ωcell=0 , ( 2 ) where n is an outward normal to the cell membrane ∂Ωcell . Dynamics of the stochastic subsystem are described by a two-component probability distribution function , {P0i ( t ) , P1i ( t ) } , given that in our simple model a channel has only two states . The differential Chapman-Kolmogorov equation that governs Markov processes [5] reduces in this case to P0i ( t+dt ) = ( 1−konidt ) P0i ( t ) +koffiP1i ( t ) dtP1i ( t+dt ) = ( 1−koffidt ) P1i ( t ) +koniP0i ( t ) dt ( i=1 , 2 , … , Nch ) , ( 3 ) where koni and koffi are the rate constants for channel openings and closings , respectively . ( Because P0i ( t ) +P1i ( t ) ≡1 , it is sufficient to solve only for one of the components , say , for P1i ( t ) . ) Importantly , parameters koni and koffi may depend on U ( r , t ) ; this would couple Eq ( 3 ) with Eqs ( 1 and 2 ) and also make the equations with different i , which otherwise would be independent , indirectly affect each other . Note that because of coupling with Ξi ( r , t ) , U ( r , t ) also becomes a stochastic variable . Eqs ( 1–3 ) fully determine the time-dependent behavior of the deterministic-stochastic system for given initial conditions {U ( r , 0 ) ;{P1i ( 0 ) }} . Their generalization to multivariate ( multistate ) models is straightforward , yielding descriptions that retain the structure and features of Eqs ( 1–3 ) . Specifically , a multivariate spatial piecewise-deterministic Markov process is defined in terms of random variables of two types [16 , 28] , the continuous ‘U-type’ and discrete ‘Ξ-type’ variables . Using vector notation for sets of these variables , all possible outcomes of the process , {U ( r ) , Ξ ( r ) } , form an infinite-dimensional function space [28] . The only practical approach to solving numerically for a time-dependent probability density functional p ( {U ( r ) , Ξ ( r ) } , t ) is by Monte Carlo simulations of individual realizations of a system based on generation of pseudorandom numbers . The description in the mold of Eqs ( 1–3 ) provides an intuitive script for an algorithm of this type . ( Alternatively , one can seek a direct numerical solution of a functional equation governing p ( {U ( r ) , Ξ ( r ) } , t ) [5] , which , however , quickly runs into memory constraints . Still , this approach can be used for testing purposes , see subsection Fully coupled systems with finite diffusion: validation against direct solutions of Fokker-Planck equation ) . Our spatial hybrid algorithm employs fixed time step integration due to its conceptual and logistical simplicity . The downside is that the stability constraints imposed on the time step , which should be sufficiently small to resolve fast ‘deterministic’ reactions , may result in slow performance . The inefficiency can be partially alleviated by applying an automatic pseudo-steady-state treatment [50] . A key element of a hybrid method is how the numerical treatments of the ‘deterministic’ and stochastic subsystems are merged . In our algorithm , the PDEs are discretized in space using a finite-volume scheme [51] , in which a computational domain Ω is partitioned into Nω subvolumes: Ω = {ωj} , j = 1 , … , Nω . The U-type variables are discretized respectively as U ( r ) → {Uj ≡ U ( rj ) } , where rj is the center of ωj and Uj has a meaning of a subvolume average: Uj=|ωj|−1∫ωjU ( r ) d3r , where |ωj| stands for the volume of ωj . Spatial histograms of stochastic variables that use the same subvolumes {ωj} as bins would have the similar meaning . Indeed , let NΞ be the number of particles of a given type Ξ; then the histogram Ξjs=|ωj|−1∫ωj∑i=1NΞΞi ( r ) |ξi=sd3r describes the density of particles of the molecular type Ξ in state s in the vicinity of rj or , more precisely , the number of particles Ξi|ξ = s in ωj divided by the volume of ωj ( s = 1 , … , Nst ( Ξ ) ; here , Nst ( Ξ ) is the number of states of a particle of type Ξ ) . For example , the spatial binning of the stochastic source term of Eq ( 1 ) yields |ωj|−1∫ωj∑i=1NchΞi ( r ) |ξi=1d3r=nj/|ωj| , where nj is the number of open calcium channels inside ωj . Then , as expected , Jnj/|ωj| is the rate of change of calcium concentration due to the influx through open channels located in the vicinity of rj . As a result , both the deterministic and stochastic rates can now be expressed in terms of sets {Uj , Ξjs} with components defined for the same spatial grid , which makes advancing the hybrid system in time conceptually straightforward . A realization of a piecewise deterministic Markov process at time t + Δt is generated on the basis of a known state at time t as follows . For sufficiently small time steps Δt , such that the sum of the O ( Δt ) terms in the expansion of the total transition probability for a particle is less than 1 , a particle may undergo at most one stochastic transition per Δt from its current state to a new one . ( For the example described by Eq ( 3 ) , this requirement yields a condition Δt<<1/max ( koni , koffi ) ) . Thus , without loss of generality , occurrences of the stochastic transitions can be assigned to the end points of the interval Δt . Therefore during the interval , the variables Ξ ( r , t ) remain unchanged and , upon the binning described above , the equations for U ( r , t ) become regular deterministic PDEs ( see , e . g . , Eq ( 1 ) of the simple calcium sparks model ) . The updated values U ( r , t + Δt ) are then found by integrating the PDEs over Δt with the corresponding boundary conditions ( exemplified by Eq ( 2 ) ) . In our method , this is done by employing a fixed time step PDE solver of VCell . The update of variables Ξ ( r , t ) is carried out by employing Smoldyn , a particle-based fixed time step Monte Carlo package [40 , 41] . Using again the simple calcium spark model as an example , the transitions of a channel between open and closed states can be interpreted as ‘unimolecular’ reactions , which are simulated by Smoldyn through acceptance-rejection sampling . First , those of the rate parameters koni and koffi in Eq ( 3 ) that depend on U ( r , t ) are updated accordingly . Next , given known states of the channels ξi ( t ) at time t , the probability of a transition to occur by the end of the time interval is computed . If , for example ξi ( t ) = 0 , i . e . the ith channel is in a closed state at time t , then P0i ( t ) =1 and P1i ( t ) =0 . As a result , the first of Eq ( 3 ) becomes dP0i/dt=−koniP0i , and because the rate constants stay fixed during the time interval , P0i ( t+Δt ) =exp ( −koniΔt ) and P1i ( t+Δt ) =1−exp ( −koniΔt ) . Finally , a random number r is generated and compared with P1i ( t+Δt ) . If r<P1i ( t+Δt ) , the transition to the open state with ξi ( t + Δt ) = 1 is accepted , otherwise it is rejected . The similar logic applies to channels that are open at time t . Note that bimolecular reactions , in which one of the participants is described by a U-type variable and the other is represented by a Ξ-variable , can be approximated in deterministic-stochastic models as unimolecular . Indeed , the copy numbers described by variables of U-type are assumed to be deterministically large even within ωj , so the changes due to binding to , or unbinding from , discrete particles can be ignored . In other words , the molecules described in terms of concentrations could be treated as ‘catalysts’ in this type of interactions . In summary , the algorithm includes the following steps: Initializing the system: Accuracy of our spatial deterministic-stochastic solver is affected by truncation errors , arising from discretization of space and time , and statistical errors due to finite numbers of Monte Carlo realizations . The algorithm was validated against analytical results and through comparison with alternative methods . The calcium spark model introduced in the previous section was used as a testbed for the tests described below . In this section , we formulate a deterministic-stochastic model of spontaneous emergence of cell polarity and simulate it with our method . The model is a hybrid version of a fully stochastic mechanism originally proposed by Altschuler et al . [54] . Division , differentiation , and proliferation of living cells rely on mechanisms of symmetry breaking . A key element of these mechanisms is emergence of asymmetric ( polar ) distributions of signaling molecules , often in form of molecular clusters . While clustering may be spurred by external cues , many cell types can polarize spontaneously ( see [54 , 55] and references therein ) . Positive feedback in cell signaling is thought to play a crucial role in establishing cell polarity . The model by Altschuler et al . demonstrates that the positive feedback combined with stochasticity is sufficient for the emergence of a unipolar distribution of membrane-bound molecules . In the model , molecules from a cytoplasmic pool randomly associate with , and dissociate from , the membrane . While in the membrane , they diffuse but also recruit more molecules from the pool . The positive feedback reinforces the clustering . Remarkably , stochasticity of the system is critical for self-polarization: the effect disappears if the copy number of molecules in the membrane exceeds a certain threshold , so that there are no asymmetric solutions in the deterministic limit . However , it is not uncommon for the membrane molecular clusters to involve large numbers of molecules . One such example is focal adhesions whose formation is initiated by membrane proteins called integrins . Activated by their binding to extracellular matrix , the integrins recruit many other molecules from the cytosol , which together form a focal adhesion . In our deterministic-stochastic model , the membrane receptor proteins that initiate clustering are distinguished from the cytosolic proteins recruited to the membrane . We assume that numbers of receptor proteins are sufficiently small to be represented by discrete variables , whereas copy numbers of cytosolic proteins , both recruited to the membrane and remaining in the cytoplasm , can be modeled continuously in terms of surface densities and volumetric concentrations . We then solve this hybrid model numerically using our method to determine if it retains the property of spontaneous polarization . The corresponding ‘Langevin-like’ formulation of the problem is as follows . Consider a cell Ω with the plasma membrane ∂Ω . Let U ( r , t ) ( r ∈ Ω ) be the volume density of the proteins in the cytoplasm and S ( r , t ) ( r ∈ ∂Ω ) be the surface density of the proteins recruited to the membrane . To describe receptor proteins residing in the membrane , we introduce discrete variables Γi ( r , t ) = δ ( r − ri ( t ) ) γi ( t ) with r ∈ ∂Ω and i = 1 , … , Nr , where Nr is the total number of receptors in the membrane . The discrete random variables γi ( t ) accept two values: 0 ( inactive receptor ) and 1 ( active receptor ) , whereas ri ( t ) are continuous random variables in ∂Ω ( see discussion in subsection capabilities and limitations of the method ) . Variables U ( r , t ) and S ( r , t ) form the ‘deterministic’ subsystem of the model and are governed by the following equations: ∂tU=DUΔU∂tS=DSΔsS+k1U∑i=1NrΓi−k2S , ( 8 ) where Δ is the Laplacian in Ω , whereas Δs is the Laplace-Beltrami operator describing diffusion in ∂Ω ( see , e . g . , [49] ) ; DU and DS are the corresponding diffusion constants . The two other terms in the equation for S are the rates with which the cytosolic proteins are recruited to , and dissociated from , the membrane; k1 , k2 are the corresponding on- and off- rate constants . The boundary condition for the equation describing U reflects the local mass conservation , −DU ( n∇U ) |∂Ωcell=−k1U∑i=1NrΓi+k2S , ( 9 ) where n is the outward normal . Realizations of γi ( t ) are governed by Poisson processes with the following transition probabilities: P ( Γi ( r , t+dt ) γi=1|Γi ( r , t ) γi=0 ) =k3S ( r , t ) dtP ( Γi ( r , t+dt ) γi=1|Γi ( r , t ) γi=1 ) =1−k4dtP ( Γi ( r , t+dt ) γi=0|Γi ( r , t ) γi=1 ) =k4dtP ( Γi ( r , t+dt ) γi=0|Γi ( r , t ) γi=0 ) =1−k3S ( r , t ) dt , ( 10 ) where k3 , k4 are the on- and off- rate constants for receptor activation . Stochastic variables ri ( t ) are modeled on an assumption that inactive receptors diffuse in the membrane , while active receptors are immobile . Accordingly , ri ( t+dt ) ={ri ( t ) +dr ( ri ( t ) , dt ) , ifγi ( t ) =0ri ( t ) , ifγi ( t ) =1 , ( 11 ) where dr ( ri ( t ) , dt ) is a realization of a Wiener-type stochastic process described by Green’s function for the diffusion operator ∂t − DΓΔs on ∂Ω; the function is centered at ri ( t ) . The initial positions of the receptors ri ( 0 ) are uniformly distributed in ∂Ω . Other initial conditions are discussed below . The model includes a positive feedback between Γi ( r , t ) and S ( r , t ) , given that the rate of recruitment of cytosolic proteins to the membrane depends on Γi ( r , t ) , while the receptor activation rate depends on S ( r , t ) . It is easy to see that the system described by Eqs ( 8–11 ) has an inactive steady state: γi ( t ) = 0 for all i , S ( r , t ) = 0 , and U ( r , t ) = U0 ( U0 is the initial uniform concentration of the cytosolic protein ) . For some parameter sets , however , the inactive steady state can become unstable or the model may exhibit multi-stability . These possibilities can be explored by solving the model with varying initial conditions . Alternatively , one can transiently perturb the inactive steady state used as an initial condition . The latter approach was implemented in the example below by adding a pre-activation pulse to the intrinsic activation rate P ( Γi ( r , t+dt ) γi=1|Γi ( r , t ) γi=0 ) = ( k0e−t/τ+k3S ( r , t ) ) dt and , correspondingly , P ( Γi ( r , t+dt ) γi=0|Γi ( r , t ) γi=0 ) =1− ( k0e−t/τ+k3S ( r , t ) ) dt; k0 and τ are the rate and time constants of the pulse . The model has been solved by the spatial hybrid method in a spherical cell with radius R = 4 μm for the following model parameters: U0 = 1 μM , Nr = 1000 , DU = 10 μm2/s , DS = DΓ = 0 . 1 μm2/s , k1 = 0 . 01 μM-1s-1 , k2 = 0 . 01 s-1 , k3 = 0 . 01 μm-2s-1 , k4 = 0 . 1 s-1 . For this parameter set , the inactive state is unstable: activation of a single receptor drives the system to its active state with an average of about 800 active receptors . Interestingly , spatial averages of all variables have reached their active steady-state regimes relatively quickly ( by t = 10 s , for the robust pre-activation characterized by k0 = 10 s-1 and τ = 1 s , and by t ≈ 350 s , when just ten receptors were initially activated ) , whereas the cluster structure evolves on a much longer time scale , see results in Fig 8 obtained for k0 = 10 s-1 and τ = 1 s . As in the original stochastic model [54] , the hybrid mechanism yields a spatially heterogeneous steady state with a single cluster of activated receptors and recruited proteins . But unlike the original model , the total number of proteins in clusters can be large , because the condition of small copy numbers applies in the hybrid model only to the receptors initiating the clustering . Note the increase of local densities in the surviving clusters ( see color scales in Fig 8 ) , which is consistent with the early stabilization of spatial averages . While the ‘attractive’ spatial correlations of active receptors originate from the positive feedback , a corresponding deterministic formulation does not yield a spatially heterogeneous steady state ( as was the case with the original model [54] ) , indicating that the discreteness and stochasticity of the receptors also play an essential role in establishing the polar distributions of membrane-bound molecules . Interestingly , the kinetics of cell polarization predicted by the model is reminiscent of glassy behavior , in which a system approaches a stable steady state by going through a long sequence of metastable states [56] .
The deterministic-stochastic algorithm described in this article integrates a spatial particle-based fixed time step Monte Carlo method ( Smoldyn ) and a conventional PDE solver with compatible time-stepping ( one of the VCell solvers ) . The PDE solver utilizes finite-volume spatial discretization of PDEs [48 , 49] , which ensures local mass conservation , and a semi-implicit time discretization scheme , in which the diffusion/ advection operator applies to variables at time t + Δt while the reaction and membrane flux terms are evaluated at time t [50 , 51] . To ensure consistency in handling geometry by the two methods , triangulation of surfaces is performed by applying Taubin smoothing [57] to watertight pixilated surfaces emerging from segmentation of space . The approach is applicable both to geometries defined analytically and to irregular realistic geometries based on experimental images . Implementation in VCell Math workspace of the hybrid model of spontaneous cell polarization described in Section Application to a hybrid model of spontaneous cell polarization is detailed in S3 Text . The corresponding VCell MathModel , ‘Hybrid_cell_polarity_public’ , along with simulation results , can be found by logging to VCell , http://vcell . org , and searching the database of public MathModels under username ‘boris’ .
Stochastic processes are ubiquitous in cellular systems . A deterministic-stochastic description of interacting components with disparate degrees of stochasticity provides an efficient alternative to a full stochastic treatment of the problem . In a hybrid numerical approach , an appropriate integration of deterministic and stochastic methods yields significant computational savings . In this paper , we describe a general-purpose hybrid method for solving spatial deterministic-stochastic models in realistic cell geometries . The emphasis is placed on the physical fundamentals of the method and its testing . The method is based on a formulation in terms of stochastic variables of two types: continuous variables , described by partial differential equations with stochastic source terms , and discrete variables governed by stochastic jump processes . Numerically , the algorithm is a Monte Carlo fixed time step integrator generating realizations of the hybrid system . The current implementation utilizes a VCell fixed time step PDE solver coupled with a particle-based stochastic simulator Smoldyn . Validating a hybrid deterministic-stochastic numerical scheme is conceptually nontrivial and logistically challenging . We tested our method against analytical results and numerical solutions obtained by alternative methods . The expected convergence of solution error was observed in tests with a separable stochastic subsystem . Testing of the method in conditions of full coupling was performed in the limit of fast diffusion against well-mixed solutions obtained with nonspatial Gibson-Bruck method and against a direct solution of a corresponding Fokker-Planck equation . The latter approach was also used for testing spatially heterogeneous solutions of fully coupled hybrid systems . The method has been applied to a hybrid model of spontaneous cell polarization based on the original idea of Altschuler et al . [54] . The solution recapitulates emergence of a stable asymmetric distribution of membrane-bound molecules , as a result of positive feedback and stochasticity . But in the hybrid version , the total number of membrane molecules is free from the small copy number requirement , which now applies only to the number of receptors that initiate clusters . The model predicts glassy-like kinetics of coalescence of the multi-cluster structure into a single cluster . While the VCell spatial hybrid solver is practical for many typical applications , its performance may become suboptimal for cases with disparate time scales ( ‘stiff’ problems ) , as the integration is done with a fixed time step . The handling of the discrete variables can be optimized by incorporating adaptive approaches , although potential savings should be weighed against costs associated with additional logistical complexity , particularly since the inefficiencies are often caused by stiffness of the deterministic subsystem . While the stiffness caused by fast reactions that persist throughout the time of interest can be addressed by applying the VCell automatic quasi-steady-state approximation ( see discussion in Section Mathematical problem and algorithm ) , the treatment of continuous variables would generally benefit from implementation of time-step control commonly employed in deterministic numerical algorithms .
|
Mechanisms of some cellular phenomena involve interactions of molecular systems of which one can be described deterministically , while the other is inherently stochastic . Calcium ‘sparks’ in cardiomyocytes is one such example , in which dynamics of calcium ions , which are usually present in large numbers , can be described deterministically , whereas the channels open and close stochastically . The calcium influx through the channels renders the entire system stochastic , but a fully stochastic treatment accounting for each calcium ion is computationally expensive . Fortunately , such systems can be efficiently solved by hybrid methods in which deterministic and stochastic algorithms are appropriately integrated . Here we describe fundamentals of a general-purpose deterministic-stochastic method for simulating spatially resolved systems . The internal workings of the method are explained and illustrated by applications to very different phenomena such as calcium ‘sparks’ , stochastically gated reactions , and spontaneous cell polarization .
|
[
"Abstract",
"Introduction",
"Results",
"Methods",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"applied",
"mathematics",
"electrophysiology",
"neuroscience",
"simulation",
"and",
"modeling",
"algorithms",
"membrane",
"proteins",
"ion",
"channels",
"mathematics",
"statistics",
"(mathematics)",
"membrane",
"receptor",
"signaling",
"cellular",
"structures",
"and",
"organelles",
"macromolecules",
"research",
"and",
"analysis",
"methods",
"polymer",
"chemistry",
"probability",
"density",
"proteins",
"mathematical",
"and",
"statistical",
"techniques",
"statistical",
"methods",
"chemistry",
"calcium",
"channels",
"monte",
"carlo",
"method",
"biophysics",
"cell",
"membranes",
"probability",
"theory",
"physics",
"biochemistry",
"signal",
"transduction",
"cell",
"biology",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"cell",
"signaling",
"neurophysiology"
] |
2016
|
Numerical Approach to Spatial Deterministic-Stochastic Models Arising in Cell Biology
|
Methylation of specific lysine residues in core histone proteins is essential for embryonic development and can impart active and inactive epigenetic marks on chromatin domains . The ubiquitous nuclear protein PTIP is encoded by the Paxip1 gene and is an essential component of a histone H3 lysine 4 ( H3K4 ) methyltransferase complex conserved in metazoans . In order to determine if PTIP and its associated complexes are necessary for maintaining stable gene expression patterns in a terminally differentiated , non-dividing cell , we conditionally deleted PTIP in glomerular podocytes in mice . Renal development and function were not impaired in young mice . However , older animals progressively exhibited proteinuria and podocyte ultra structural defects similar to chronic glomerular disease . Loss of PTIP resulted in subtle changes in gene expression patterns prior to the onset of a renal disease phenotype . Chromatin immunoprecipitation showed a loss of PTIP binding and lower H3K4 methylation at the Ntrk3 ( neurotrophic tyrosine kinase receptor , type 3 ) locus , whose expression was significantly reduced and whose function may be essential for podocyte foot process patterning . These data demonstrate that alterations or mutations in an epigenetic regulatory pathway can alter the phenotypes of differentiated cells and lead to a chronic disease state .
The process of embryonic development determines the differentiated state of all cells by establishing unique gene expression patterns , or signatures , for individual cell types that define their phenotypes . Once a differentiated state is established , it is difficult to erase that epigenetic imprint and reprogram the cell towards a different cell lineage or phenotype . Although reprogramming can be forced by nuclear transplantation [1] or by the expression of Oct4 and accessory factors [2] , [3] , the low efficiency of these processes speaks to the inherent stability of a differentiated cell . Gene expression patterns must be established and maintained by compartmentalizing the genome into active and inactive regions , which is thought to occur through the covalent modifications of DNA and its associated nucleosomes . Such modifications include DNA methylation of CpG islands and methylation , acetylation , and ubiquitination of histone tails , all of which are thought to determine chromatin structure and accessibility [4] , [5] . This epigenetic code is thus imprinted upon the primary genetic code during embryonic development to help establish cell lineages and restrict fate . The genetics and biochemistry of histone modifications have been well studied in a variety of model organisms and developmental contexts . Genes of the Polycomb and Trithorax families encode proteins that are required for methylation of different histone lysine residues and often correlate with gene silencing or activation , respectively [6]–[9] . Many Trithorax group proteins , such as Drosophila TRX and human KMT2A ( MLL ) , are histone H3 lysine 4 ( H3K4 ) methyltransferases ( KMTs ) and are essential for maintaining gene expression patterns in diverse organisms . Recently , we discovered a novel co-factor , PTIP ( Pax Transactivation-domain Interacting Protein ) , which is encoded by the Paxip1 gene . The PTIP protein co-purifies with the mammalian lysine methyltransferases KMT2B and KMT2C ( formerly ALR and MLL3 ) , is broadly expressed , and is essential for embryonic development [10]–[12] . At least in one case , PTIP is able to recruit the KMT2B complex to a developmental DNA binding protein in a locus specific manner [13] . Loss of PTIP function in the mouse results in gross developmental effects at gastrulation , with reduced levels of global H3K4 di- ( me2 ) and trimethylation ( me3 ) observed [13] , [14] . In cultured mouse embryonic stem cells , PTIP is needed to maintain pluripotency , Oct4 expression , and normal levels of H3K4 trimethylation [15] . Similarly , in neuronal stem cells , differentiation is abrogated and levels of H3K4 methylation are reduced in tissue specific PTIP knockouts [13] . In mouse embryo fibroblasts , loss of PTIP blocks differentiation by inhibiting PPARγ and C/EPBα activation and H3K4 methylation at their respective promoters [16] . Similarly , the Drosophila homologue of PTIP is also essential for development , epigenetic control of gene expression , and global histone H3K4 methylation [17] . During cell division , patterns of histone methylation must be inherited by daughter cells such that the cellular phenotype is maintained . For repressive histone methylation marks , such as histone H3 lysine 27 , the EED ( Embryonic Ectodermal Development ) protein is thought to bind and recruit the Polycomb Repressor Complex 2 to replicate and maintain gene silencing after mitotic cell division [18] , [19] . For highly expressed genes , the KMT2A ( MLL1 ) protein associates with promoter regions on condensed mitotic chromatin and is required to rapidly reactivate such genes after cell division [20] . These data suggest a model whereby histone methylation patterns are replicated during mitosis , but do not address the necessity for maintaining epigenetic modifications in terminally differentiated , non-dividing cells . Furthermore , changes in the expression of epigenetic regulatory genes have been reported in a variety of cancers [21] and disease states [22] , but whether these are the cause or the result of disease remains to be determined . To address the necessity of H3K4me3 in a stable non-dividing cell type , we utilized a Podocin-Cre transgenic driver to delete PTIP in the glomerular podocyte , a highly specialized and architecturally distinct cell that establishes the kidney filtration barrier . Podocytes are clinically relevant cells whose properties and expression profiles change in glomerular diseases and in older animals [23] . While the ubiquitous expression of PTIP , its role in H3K4 methylation , and its necessity in development and differentiation are all well established , whether PTIP deletion in terminally differentiated cells can induce changes in the pattern of H3K4me3 and gene expression has not been demonstrated . We show that loss of PTIP results in changes in the transcriptional profile of terminally differentiated podocyte cells , which ultimately leads to a chronic glomerular disease phenotype . Among the most affected is the neurotrophin receptor encoding gene Ntrk3 , whose function had not been previously studied in podocytes . Our results demonstrate a maintenance function for PTIP-mediated H3K4 methylation and identify a novel role for Ntrk3 in podocyte foot process patterning .
To specifically knockout PTIP protein in fully differentiated mouse podocytes , we utilized both floxed ( fl ) and conventional null ( - ) alleles of Paxip1 and a Cre driver strain specific for glomerular podocytes . The Paxip1fl/−:CreNPHS2 mice were crossed to Paxip1fl/fl animals to generate Paxip1fl/fl or Paxip1fl/− with or without CreNPHS2 . The CreNPHS2 mice utilize the NPHS2 promoter to express Cre recombinase only in late developing and mature podocytes [24] , [25] . The resulting progenies were born in the expected Mendelian ratios and did not show any gross kidney defects during the first 4 weeks of life ( data not shown ) . For simplicity , we will refer to the mice as either PTIP− ( Paxip1fl/−:CreNPHS2; Paxip1fl/fl:CreNPHS2 ) or PTIP+ ( Paxip1−/fl , or Paxip1fl/fl ) . PCR analysis indicated that recombination occurred at the Paxip1 locus in DNAs isolated from kidneys but not in DNAs from tails ( Figure 1A ) . Previous work established that the Paxip1fl allele produces normal levels of protein , but Cre-mediated excision of exon 1 and the promoter region results in complete absence of PTIP protein , essentially creating a null allele [13] , [15] . The specificity of the Cre driver strain was confirmed by crossing CreNPHS2 mice to the Rosa26-LacZ reporter mice ( Figure 1B ) . In 1 month old kidneys , lacZ expression was restricted to the glomerulus only , indicating efficient Cre mediated excision at this time . Immunostaining for PTIP and the podocyte marker WT1 also confirmed that PTIP protein levels were reduced only in the podocyte cells and not the mesangial or endothelial components of the glomerular tuft ( Figure 1C ) . Previous work showed that a loss of PTIP function results in reduced levels of total H3K4me3 levels in embryos and cultured cells [13]–[17] . To test whether podocytes showed reduced H3K4me3 , we stained kidney sections with antibodies specific for this modification ( Figure 1D ) . Many podocytes were observed with reduced signal intensities . To quantitate this effect , images were analyzed for signal intensity by integrating a fixed area over the nuclei of both podocytes and other cell types ( Figure 1E ) . Podocytes were co-stained with WT1 antibodies . The ratio of podocyte signal ( WT1+ ) to other cell types ( WT1− ) was calculated by counting at least 6 cells of each type per glomerulus . The ratios from at least 8 glomeruli were averaged for each genotype and shown to decrease by more than 20% in PTIP− kidneys compared to PTIP+ controls ( p<0 . 01 ) . These data confirmed that the specific deletion of PTIP in the podocytes correlates with a reduction in H3K4me3 in this cell type . Podocytes play a critical role in the establishment and maintenance of the glomerular filtration barrier . Interdigitated podocyte foot processes cover the glomerular basement membrane and form specialized junctions , called slit diaphragms , which create a highly selective barrier that filters small and negatively charged proteins and solutes from the blood to the urinary space . Damage to or loss of podocytes impairs the filtration barrier and results in increased rates of excretion of high molecular weight proteins , such as albumin , into the urine . Thus , we checked mice for proteinuria beginning at 1 month of age ( Figure 2A ) . At 1 month , low levels of albumin were detected in the urine but these were not significantly different between PTIP+ and PTIP− animals . However , by 3 months of age the PTIP− mice showed significantly higher levels of albumin in the urine and these levels increased further at 6 and 12 months . The urine albumin to creatinine ratio ( ACR ) provides a quantitative assay that correlates with filtration barrier integrity . No significant differences were observed at 1 month ( Figure 2B ) . However , by 3 and 12 months , ACR were 10 and 30 fold higher respectively in urines of PTIP− animals compared to PTIP+ mice . Mice that carried the CreNPHS2 transgene in a Paxip1+/+ or a Paxip1fl/+ genetic background did not show any renal abnormalities at 12 months ( data not shown ) , consistent with many published reports that have used this particular Cre driver strain [25]–[28] . Renal pathology was characterized by light microscopy at 1 , 3 , and 12 months of age . Standard Masson's Trichrome and Periodic-Acid-Shiff stainings revealed significant sclerosis and matrix deposition in 12 month old glomeruli from PTIP− animals ( Figure 2C ) . However , 3 month old kidneys did not show significant differences for most glomerular sections , at the light microscopy level , although evidence of limited matrix expansion could be observed in a small number of glomeruli of PTIP− kidneys . In 12 month old kidneys , significant interstitial fibrosis and protein filled cysts were also observed ( Figure 2D ) . These are likely to be secondary effects due to the glomerular pathology . Glomerular pathology and increased albuminuria can be the direct result of podocyte death [29] . Thus , we used a variety of markers to characterize the glomerular architecture and the numbers of podocyte cells at various ages to insure that the phenotype of the PTIP− mice was not just the result of early podocyte cell death . Immunostaining with WT1 , Nephrin , and Podocin antibodies enabled us to determine the podocyte numbers , as average per mid-cross section , and to indirectly assess the integrity of the slit diaphragm ( Figure 3 ) . The number of WT1 positive podocytes was not significantly different between PTIP+ and PTIP− glomeruli at 1 or 3 months of age . At 6 months , PTIP− glomeruli had slightly fewer podocytes and by 12 months , the number of podocytes was half that of the PTIP+ littermates . Immunostainings for podocyte markers such as WT1 , Nephrin , and Podocin did not reveal dramatic differences at 1 or 3 months , despite the increase in proteinuria , although some discontinuous staining could be seen with Podocin antibodies in PTIP− glomeruli ( Figure 3B ) . Consistent with this data , TUNEL staining for apoptosis did not reveal differences between PTIP+ and PTIP− kidneys at 1 or 3 months of age ( data not shown ) . Thus , the breakdown of the filtration barrier was not due to simple podocyte depletion at these early times . However by 12 months of age , the extensive network of Nephrin staining was partially depleted in PTIP− glomeruli ( Figure 3B ) . At the light microscopy level , the effects of PTIP loss on glomerular architecture seemed minimal at 3 months of age , yet the levels of albumin in the urine suggested significant functional defects . Thus , we utilized scanning and transmission electron microscopy to characterize the podocytes at the ultra structural level ( Figure 4 ) . Scanning electron micrographs revealed disorganized foot processes at 3 months . While PTIP+ podocytes had regularly arrayed tertiary foot-processes that were almost parallel ( Figure 4A ) , the PTIP− podocyte foot processes were much more irregular and flattened . The parallel pattern of interdigitation was clearly different and resembled a jigsaw puzzle with random patterning ( Figure 4B , 4C ) . Transmission electron micrographs at 3 months also revealed that the slit-diaphragms were not evenly spaced and fusion of foot processes was frequent ( Figure 4D–4F ) . By 12 months , the remaining podocytes in the PTIP− kidneys were broader , flatter and displayed significant fusion or effacement ( Figure 4G , 4H ) , consistent with the high levels of albumin detected in the urine . These data demonstrate that the initial glomerular phenotype in PTIP− kidneys is due primarily to differences in podocyte foot process morphology , which occurs prior to the loss of cell bodies . Alterations in cellular phenotypes could be the result of changes in the transcriptional program of PTIP− podocytes . Thus , we prepared RNA from glomeruli enriched fractions at 1 month of age , prior to the onset of any significant phenotype , and assayed for gene expression changes by Affymetrix microarrays . We compared glomerular RNA preps from 10 independent PTIP− animals and 8 PTIP+ littermates at 1 month of age . The data were highly consistent and indicated both gain and loss of gene expression in the PTIP− kidneys ( Table 1 and Table 2 ) . The entire dataset can be accessed at the Gene expression Omnibus ( GSE17709 ) . Expression changes were confirmed by quantitative RT-PCR for selected genes ( Figure 5 ) . Among the genes increased was Protamine1 ( Prm1 ) , which is not normally expressed in podocytes or other somatic cells but is found only in spermatids where it is essential for chromatin condensation and fertility [30] , [31] . The changes in RNA expression observed were surprising and did not correspond to any common pathways . In fact , the podocyte-specific genes that are known to function in cell viability and slit diaphragm integrity were largely unchanged ( Table S1 and Figure 5C ) . The data suggest that loss of PTIP in podocytes alters the transcriptional program to affect a limited number of genes whose functions in the podocytes have not been previously characterized . Among the most interesting genes whose expression was down regulated in PTIP− kidneys was the neurotrophic tyrosine kinase receptor type 3 ( Ntrk3 , formerly called TrkC ) , whose expression in podocytes had not been previously described . The Ntrk3 gene encodes two proteins that recognize neurotrophin 3 ( NT-3 ) and functions in axon guidance and innervation and in cardiac development [32]–[34] . Ntrk3 promotes axon outgrowth and guidance , presumably through actin based extension and retraction of cellular processes [35] . Given that podocyte foot processes are also actin based and may require some type of guidance , we examined the role of Ntrk3 further . Quantitative RT-PCR confirmed that Ntrk3 expression was down approximately 10 fold in glomerular preps from PTIP− compared to PTIP+ animals ( Figure 5A ) . We also examined Ntrk3 levels in kidneys by co-immunostaining kidney sections with Ntrk3 , WT1 and Nephrin antibodies ( Figure 6 ) . At 3 months of age , Ntrk3 could be seen in glomeruli of PTIP+ kidneys , however the staining intensity in PTIP− kidneys was severely reduced in almost every glomerulus examined ( Figure 6D , 6J ) . Some slight filamentous staining remained in the PTIP− glomeruli , but the overall intensity was markedly different . In PTIP+ glomeruli , Ntrk3 staining was remarkably similar to Nephrin ( Figure 6G–6I ) . However , Nephrin staining intensity was unaffected in PTIP− glomeruli even though Ntrk3 was much lower ( Figure 6J–6L ) . The Ntrk3 expression in glomerular preps and its decrease in the PTIP− kidneys suggested a function in foot process growth , guidance , and/or pattern formation . In order to more directly link PTIP to the Ntrk3 locus , we designed chromatin immunoprecipitation experiments to examine the presence of PTIP and the changes in histone methylation patterns around the transcription initiation site ( +1 ) of Ntrk3 ( Figure 7 ) . Chromatin was prepared from whole glomerular preps from PTIP+ and PTIP− kidneys , which also included mesangial and endothelial cells . Despite the presence of other cell types in the glomerular chromatin , we were able to detect a 5–6 fold decrease in PTIP localization to sequences around the start site of Ntrk3 transcription when comparing PTIP+ to PTIP− chromatin ( Figure 7B ) . No significant amount of PTIP was detected further upstream ( −1200 ) , nor did we see a significant difference , between PTIP+ and PTIP− chromatin , in PTIP localization within the 5′ UTR of exon 1 ( Figure 7B , P4 site ) . Clear differences in H3K4me2 were also measured , with an approximately 50–60% decrease in PTIP− chromatin with primer pairs P2–P4 , but not with P1 at −1200 ( Figure 7C ) . Similarly , H3K4me3 levels were also decreased in PTIP− chromatin at P2–P4 but not at P1 ( Figure 7D ) . We also examined changes in Polycomb mediated epigenetic silencing marks using an antibody against H3K27me3 ( Figure 7E ) , which appeared unchanged at all sites examined . These data demonstrate recruitment of PTIP to the promoter region of Ntrk3 in normal glomeruli . In order to determine if the loss of Ntrk3 alone would impact normal glomerular patterning , we examined homozygous Ntrk3 mutant mice . The Ntrk3 mutants die shortly after birth due to cardiac and neuromuscular defects; however their kidneys had not been studied previously . Therefore , we collected urine and kidney tissue for light and electron microscopy from 3–4 day old Ntrk3 mutants and littermates . At three days post partum , Ntrk3 mutants were small and sickly . Higher levels of albumin could be observed in the urines of Ntrk3−/− pups ( Figure 8A ) , compared to control littermates , although this could be due to delayed or arrested kidney development . Glomerular development was examined in kidney sections of 4 day old newborns ( Figure 8B ) . At this time , nephrons are still undergoing development and glomeruli at the periphery are just beginning to form whereas cortical glomeruli closer to the medulla are already fully functional . The tight junction protein Magi2 specifically localizes to podocyte cell junctions and exhibited altered patterning in Ntrk3 mutant kidneys , with discontinuous staining and excessive looping of the developing tuft . In mature glomeruli , Nephrin staining was reduced and patchy in the Ntrk3 mutants . The number of podocytes did not seem affected in the Ntrk3−/− mice at this time . Ultra structural analysis of Ntrk3 mutant kidneys revealed podocyte patterning defects both by scanning and transmission EM ( Figure 9 ) . At 4 days post-partum , we examined the most mature glomeruli , those located closest to the medullary zone . Podocyte foot processes from Ntrk3−/− mice exhibited disorganized secondary and tertiary processes that crisscrossed randomly over capillary vessels and were poorly interdigitated ( Figure 9A′ , 9B′ ) . Few sections showed the characteristic spacing indicative of the slit diaphragms at the glomerular basement membranes ( Figure 9D′ ) . These data suggest a critical role for Ntrk3 in the fine patterning events of secondary and tertiary foot process formation and interdigitation .
In this report , we utilized a conditional deletion to ask whether the PTIP dependent H3K4 methylation function is required in a terminally differentiated cell type , to maintain its differentiated state and its cell-type specific transcriptional program . Using the glomerular podocyte cell as a model , we show that deletion of PTIP results in subtle changes in gene expression patterns that ultimately lead to a slowly progressing disease state . These data support a model in which the gross stability of the differentiated state or podocyte cell survival , at least in the short term , does not depend on the PTIP/KMT complex , as many of the podocyte specific genes examined were unchanged in the absence of PTIP . Rather , the loss of PTIP was more subtle and revealed unexpected changes in a small number of genes and ultimately led to a chronic disease phenotype resembling glomerular sclerosis . Typical characteristics of chronic glomerular disease were present , including microalbuminuria , podocyte foot process fusion or effacement , remodeling of the filtration barrier , and increased extracellular matrix deposition . Methylation of histone H3 at lysine 4 correlates with gene expression and is thought to regulate cellular identity by establishing and maintaining a stable epigenetic state . The PTIP protein is part of an H3K4 methyltransferase complex that includes the mammalian Trithorax homologues KMT2B and/or KMT2C [10] , [11] , [13] , [16] . Previous studies in flies and mice demonstrated reduced H3K4 methylation in Paxip1 mutants and severe early lethal phenotypes . In the mouse , complete loss of PTIP protein results in developmental arrest just after gastrulation [14] , a phenotype more severe than any individual mouse KMT2 family gene mutation [12] , [36] , [37] , whereas a hypomorphic Paxip1 allele is lethal later in development [38] . In flies , maternal and zygotic ptip null embryos are embryonic lethal and fail to express many segmentation genes [17] . In mouse embryonic stem cells , PTIP protein is required for normal levels of H3K4 methylation and for maintaining pluripotency in cell culture [15] , whereas in embryonic fibroblasts PTIP is required for adipocyte differentiation [16] . All of these findings suggest that a PTIP H3K4 methyltransferase complex is needed for differentiation of stem cells and progenitor cells in development . However in terminally differentiated cells , the requirement for active H3K4 methylation may be different and the lack of cell division may abrogate the need for de novo methylation . Our results suggest that PTIP must still function in some non-dividing cells , perhaps as part of a maintenance complex , as overall levels of H3K4 methylation were reduced and activation and suppression of a small number of genes was affected . The mature podocyte is generally believed to be a non-dividing cell type , as classic cell BrdU labeling experiments do not mark this population over time [39] . However , more recent genetic lineage tracing experiments suggest that there is a population of parietal epithelial cells at the vascular pole of the Bowman's capsule that can replenish podocytes over time [40] , [41] . This replacement of podocytes appears slow under normal conditions , but may be especially critical in cases of glomerular injury . In our animal model , we would expect any podocyte replacement to also delete the Paxip1 gene once expression of the Cre driver is activated . Given that we do not see significant loss of podocytes until at least 6 months of age , it may be that alterations in the transcriptional profile are not lethal . Rather , loss of podocytes may be the result of the damaged filtration barrier , the increase in the mesangium , and the general environment of the glomerulus in older mice . Alternatively , if podocyte replacement is accelerated in our model , it may be that by 6 months the ability of parietal cells to replenish the podocyte population is exhausted . In either case , the effects of manipulating the H3K4 methylation pathway is more apparent in older mice , suggesting a critical role for such epigenetic pathways in aging cells and tissues . The changes in gene expression observed in response to PTIP deletion are surprising in that most of the well-characterized podocyte-specific genes appear unaffected . However , changes include both activation and suppression of previously uncharacterized genes in the podocytes . Activation of the Prm1 gene in PTIP− kidneys is unusual as this gene has only been associated with sperm maturation and is thought to encode a unique chromatin binding protein [31] , [42] . Activation of the Padi4 gene could impact gene expression by deimination of arginines in the histone H3 tail , which prevents methylation [43] . The impact of increased Padi4 is likely to be complex as arginine methylation can correlate with gene activation or repression , depending on the context and specific residues . The most compelling gene affected in PTIP− podocytes was Ntrk3 , whose expression in the glomerulus had not been previously characterized . The reduction of Ntrk3 expression in PTIP− kidneys and the phenotype of Ntrk3−/− newborn kidneys suggest that this receptor is critical for tertiary foot process pattern formation . The podocyte is a highly specialized cell with a complex network of processes that cover the glomerular basement membrane . The large primary processes are microtubule containing structures , whereas the tertiary , interdigitated foot processes contain actin microfilaments [44] . Adjacent foot processes are connected through a specialized junctional complex , called the slit diaphragm , which is essential for maintaining a functional filtration pore . Some of the essential proteins in the slit-diaphragm , such as Nephrin , Podocin , and Neph1 are well characterized and mutations are associated with severe nephrotic syndromes [45] . Yet , how foot process outgrowth is regulated and maintained is not clear . Our data suggests that Ntrk3 , and by inference its ligand NT–3 , may be important for foot process growth and patterning . NT-3 is known to promote neuronal axon guidance by stimulating actin polymerization and lamellipodia formation [46] , [47] . In cultured neuronal cells , NT-3 promotes localization of β-actin mRNA to the growth cones to stimulate motility and chemotaxis [48] , [49] . Podocytes express many proteins known to function in neurite outgrowth , such as semaphorins , neuropilins , and ephrins . A recent report even describes the release and up-take of glutamate containing synaptic-like vesicles by podocytes [50] . Furthermore , foot processes are dynamic and can retract quickly in response to polyamines like protamine sulfate [51] , [52] . This raises the possibility that sensing mechanisms are required for rapid actin dynamics; such mechanisms may be common to both podocytes and neurons . Still , reduction of Ntrk3 alone is unlikely to cause the phenotypic changes in PTIP− podocytes over time , as other genes whose functions are not well understood are also impacted . Histone methylation by Trithorax or Polycomb complexes can imprint positive and negative epigenetic marks on chromatin during development . More recently , histone methyltransferases have been associated with cancer and other disease states . However , in many cases it is not clear whether changes in the expression of epigenetic modifiers are the cause or the result of disease progression . The results presented here suggest that mutations in an epigenetic pathway , which result in alterations of H3K4 methylation patterns , can lead to a chronic disease through subtle changes in gene expression patterns . This implies a direct function for HMTs in maintaining gene expression and the differentiated state in healthy organisms .
Mice carrying the Paxip1 null ( Paxip1− ) and floxed ( Paxip1fl ) alleles were previously described and genotyped as indicated [14] , [53] . To obtain the specific deletion of the Paxip1fl allele in glomerular podocytes , these mice were crossed with the previously characterized 2 . 5P-Cre mice [24] , [25] , which express the Cre recombinase under the control of the human NPHS2 promoter ( CreNPHS2 ) . Among the next generations , mice carrying the Cre allele ( Paxip1fl/fl:CreNPHS2 and Paxip1fl/−:CreNPHS2 mice ) were considered as conditional null mutants ( PTIP− ) , whereas littermates that did not express the Cre recombinase were used as controls ( PTIP+ ) . All animal procedures were approved by the University Committee on Use and Care of Animals ( UCUCA ) of the University of Michigan and performed in compliance with ULAM recommendations . Rabbit polyclonal antibodies used to detect Nephrin ( 1∶1000 ) and Podocin ( 1∶500 ) were kindly provided by L . B . Holzman ( University of Pennsylvania , Philadelphia , PA ) . Chicken anti-PTIP was described previously [54] . Additional antibodies were commercially available: mouse clone 6F-H2 anti-WT1 ( 1∶1000 , DAKO , Carpinteria , CA ) , anti-H3K4me3 and anti-H3K27me3 ( AbCam , Cambridge , MA ) , anti-Magi2 ( Sigma-Aldrich , St . Louis , MO ) , anti-Ntrk3 ( AF1404 , R & D Systems , Minneapolis , MN ) , Alexa Fluor 488 F ( ab′ ) 2 fragment of goat anti-rabbit IgG , Alexa Fluor 568 F ( ab′ ) 2 fragment of goat anti-mouse IgG , Alexa Fluor 488 donkey anti-goat IgG ( 1∶500; Molecular Probes , Life Technologies , Carlsbad , CA ) . Mice had access to a standard breeder chow ( Purina 5008 ) and water ad libitum . Urine was collected early in the afternoon for three consecutive days from individual mice at 1 , 3 , 6 and 12 months of age and stored frozen until use . After thawing , 2 µL urine was run on a SDS-PAGE and stained with Coomassie Blue to test for the presence of proteins/albumin , using recombinant mouse albumin ( Sigma-Aldrich , St . Louis , MO ) as a control . Quantitative assessment of urine albumin and creatinine concentrations were determined by ELISA using the Albuwell M and Creatinine Companion kits ( Exocell Inc . , Philadelphia , PA ) . Mice at 1 , 3 , 6 , and 12 months of age were sacrificed and their kidneys were perfused , fixed , and processed for histology , indirect fluorescence and electron microcopy analyses . Briefly , mice were anesthetized by intraperitoneal injection of 40 mg/kg sodium pentobarbital and prepared for systemic perfusion . A saline solution was first injected through the abdominal aorta to the entire mouse body at a pressure of approximately 70 mmHg as previously described [55] . As soon as the general bloodstream had been cleared , a solution of 4% paraformaldehyde in PBS was substituted . It was left to perfuse at the same flow conditions for approximately 10 minutes . Kidneys were removed , decapsulated , cut into pieces , and incubated for 2 additional hours in the appropriate fixative solution before being processed for histology , indirect immunofluorescence , and electron microscopy . Kidneys were fixed in 4% paraformaldehyde , embedded in paraffin , sectioned at 5 microns , and stained with Periodic Acid Shiff or Masson Trichrome . For immunofluorescence analyses with Nephrin , PTIP , WT1 and Magi2 , sections were dewaxed , rehydrated , and microwaved for 10 minutes in a citric acid-based antigen unmasking solution ( Vector Laboratories , Burlingame , CA ) . Sections were permeabilized with 0 . 3% Triton X-100 in PBS and blocked with 10% goat serum in PBS . Primary antibodies were incubated overnight at 4°C in PBS , 0 . 1% Triton , 2% goat serum . Sections were washed twice and incubated with the secondary fluorescent antibodies and DAPI in PBS , 0 . 1% Triton , 2% goat serum for 1 hour in the dark at room temperature . The sections were washed again and mounted in Mowiol . Stained and fluorescent-labeled sections were analyzed under a Nikon ES800 microscope . Micrographs were taken with a digital spot camera , using equivalent exposure times among sections . For Ntrk3 staining , fresh frozen sections were dried , fixed in methanol at −20°C and washed in PBS , 0 . 1% Tween 20 before incubation with anti-Ntrk3 antibodies at 1 µg/ml . For quantitation of immunofluorescent signals , ImageJ 1 . 42 was utilized . H3K4me3 stained sections were digitally captured and light intensity measured by placing a fixed size circular area over the nuclei of cells and summing all pixels over the given area . At least 6 podocytes and 6 control cells , either mesangial or endothelial , were measured for each of 8 glomerular tufts ( at least 48 podocytes and 48 other cells for each genotype ) . The average signal intensity was then expressed as a ratio of podocyte intensity to non-podocyte cell intensity for each of the glomerular micrographs taken . For Cre activity detection , the Rosa26-lacZ reporter strain was used [56] . Mice carrying CreNPHS2 and Paxip1fl/fl were crossed to Rosa26-stop-lacZ:Paxip1fl/+ to generate Paxip1fl/fl:CreNPHS2:Rosa26-lacZ animals . Kidneys were excised at 1 month of age and stained for β-galactosidase activity as described [57] . Longitudinal slices of kidneys from PTIP+ and PTIP− mice fixed with 2 . 5% glutaraldehyde in 0 . 1M Sorensen's buffer ( pH 7 . 2 ) for 2 hours at room temperature were processed for scanning electron microscopy following standard procedures . Briefly , after several washes with the Sorensen's buffer alone , the samples were dehydrated by successive washes in graded ethanol solutions , critical point dried , mounted on a stub , sputter coated with gold-palladium , and examined under an AMRAY 1910 field emission scanning electron microscope . Pieces of the kidney cortex ( 1 mm3 ) , fixed with 2 . 5% glutaraldehyde in Sorenson's buffer for 2 hours at room temperature , were processed for transmission electron microscopy following standard procedures . They were embedded in PolyBed 812 resin ( Polysciences Inc . ) , cut into 1-micron slices and stained with toluidine blue . Sample areas were selected based on the presence of glomeruli and cut into ultra-thin sections for analysis under a Philips CM-100 transmission electron microscope . The selected SEM and TEM images are representative of at least 10 different glomeruli per kidney . Glomeruli were isolated from the kidneys of individual mice by sieving as described [58] . Briefly , 1 month-old mice were sacrificed by CO2 inhalation and kidneys were removed . After decapsulation , the kidneys were finely minced on ice and passed sequentially through nylon meshes of 90 and 41 microns ( Sefar Filtration Inc . , Depew , NY ) . The glomeruli-enriched fraction ( GEF ) was retained on top of the 41-micron mesh , while kidney tubules were flushed through . RNA was isolated directly from the mesh . Total RNA was extracted from the GEF of individual 1-month-old mice using the RNeasy Tissue Micro Kit ( Qiagen , Valencia , CA ) following the manufacturer's instructions . RNA concentration and purity were determined by nanodrop analysis on an Agilent Bioanalyzer 2100 ( Agilent Technologies , Santa Clara , CA ) . Using the Ovation RNA Amplification System V2 ( NuGEN Technologies , San Carlos , CA ) , 500 ng total RNA was reversed transcribed and linearly amplified into single-stranded cDNA , which concentration and purity were determined by nanodrop analysis on an Agilent Bioanalyzer 2100 ( Agilent Technologies ) . Microarray analyses were done by the University of Michigan Comprehensive Cancer Center ( UMCCC ) Affymetrix and Microarray Core Facility . The FL-Ovation cDNA Biotin Module V2 kit ( NuGEN Technologies , San Carlos , CA ) was used to produce biotin-labeled cRNA , which was then fragmented and hybridized to a Mouse 430 2 . 0 Affymetrix GeneChip 3′ expression array ( Affymetrix , Santa Clara , CA ) . Array hybridization , washes , staining , and scanning procedures were carried out according to standard Affymetrix protocols . Expression data were normalized by the robust multiarray average ( RMA ) method and fitted to weighted linear models in R , using the affy and limma packages of Bioconductor , respectively [59] , [60] . Only probe sets with a variance over all samples superior to 0 . 1 , a p-value inferior or equal to 0 . 05 after adjustment for multiplicity using the false discovery rate [61] , and a minimum 2-fold difference in expression were selected for the analysis . The complete data set is available from the Gene Expression Omnibus database ( accession number GSE17709 ) . Microarray data were confirmed by real-time quantitative PCR analysis . 25–50 ng single-stranded cDNA was amplified in triplicate in a 384-well plate , using the 7900HT Fast Real Time PCR system ( Applied Biosystems , Foster City , CA ) and expression levels of selected genes was determined by SYBR Green or TaqMan assays ( Applied Biosystems ) . PCR primers pairs and TaqMan probes used in this study are presented in Table S2 . Glomeruli were isolated from 6 PTIP+ and 6 PTIP− kidneys by sieving as described above . Glomeruli were resuspended in 1 ml PBS and cross linked with 1% formaldehyde for 10 minutes with rocking at room temperature . Chromatin preparation , immunoprecipitation , and PCR analysis was essentially as described previously [13] . Primers pairs for the Ntrk3 locus were as follows: P1 , 5′- CAATGTATTTTGCTTCCTTGCC , 5′- AAGAAAGGGTTAGGGGAATCCG; P2 , 5′- AACCCGTGCGTTTCGTAAGG , 5′- GGAGGAAGGAGGAGAAGGAAGATG; P3 , 5′- GCATCTTCCTTCTCCTCCTTCCTC , 5′- AAGTCACCAAGTCCCACCTCCTAG; P4 , 5′- TTTGCCTTCCCACCGTCTGTTG , 5′- TGCCTTTGAAACGCCGAAC .
|
While all cells contain essentially the same genome , adult differentiated cells have specific patterns of gene expression for unique physiological functions . Gene expression depends on specific proteins that activate some genes and repress others so that a stable pattern of expression is maintained . During embryonic development , epigenetic modifications of the genome may compartmentalize the genome into actively expressed or repressed domains through the methylation of specific histone residues on chromatin . We studied a specific pathway of histone H3 lysine 4 methylation by deleting the co-factor PTIP in a differentiated cell type . We then asked whether this epigenetic pathway is still important for maintaining the correct pattern of gene expression . Using the podocyte cells of the glomerulus as a model system , mice that carry deletions of the PTIP protein only in these podocytes show changes in gene expression patterns over time and exhibit a slowly progressing chronic disease phenotype . Chromatin immunoprecipitation showed a loss of PTIP binding and lower H3K4 methylation at the Ntrk3 locus , whose expression was significantly reduced . These data demonstrate the need for maintaining the correct epigenetic pattern in an aging , differentiated cell type and point to modifications in epigenetics as potential disease causing factors .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"developmental",
"biology/aging",
"genetics",
"and",
"genomics/disease",
"models",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"nephrology/chronic",
"kidney",
"disease",
"genetics",
"and",
"genomics/epigenetics"
] |
2010
|
Altering a Histone H3K4 Methylation Pathway in Glomerular Podocytes Promotes a Chronic Disease Phenotype
|
Upon viral infection , cells undergo apoptosis as a defense against viral replication . Viruses , in turn , have evolved elaborate mechanisms to subvert apoptotic processes . Here , we report that a novel viral mitochondrial anti-apoptotic protein ( vMAP ) of murine γ-herpesvirus 68 ( γHV-68 ) interacts with Bcl-2 and voltage-dependent anion channel 1 ( VDAC1 ) in a genetically separable manner . The N-terminal region of vMAP interacted with Bcl-2 , and this interaction markedly increased not only Bcl-2 recruitment to mitochondria but also its avidity for BH3-only pro-apoptotic proteins , thereby suppressing Bax mitochondrial translocation and activation . In addition , the central and C-terminal hydrophobic regions of vMAP interacted with VDAC1 . Consequently , these interactions resulted in the effective inhibition of cytochrome c release , leading to the comprehensive inhibition of mitochondrion-mediated apoptosis . Finally , vMAP gene was required for efficient γHV-68 lytic replication in normal cells , but not in mitochondrial apoptosis-deficient cells . These results demonstrate that γHV-68 vMAP independently targets two important regulators of mitochondrial apoptosis-mediated intracellular innate immunity , allowing efficient viral lytic replication .
Apoptosis , or programmed cell death , has a key role in a variety of eukaryotic biological processes such as development and immune responses . Whether initiated by receptor ligation at the cell surface or through signal transduction from intracellular organelles , both caspase-dependent and -independent apoptotic pathways degrade cellular components , leading to the complete dismantling of targeted cells . Particularly , mitochondria serve as gatekeepers for the life-or-death decision , conveying apoptotic signals by releasing death-promoting factors ( e . g . , cytochrome c , apoptosis-inducing factor , and DIABLO/Smac from the intermembrane space [1–4] ) . Recent studies have largely elucidated the complex mechanism that eukaryotes have evolved to regulate the permeability of the mitochondrial outer membrane during apoptosis , particularly with regard to Bcl-2 homology ( BH ) family proteins [5–7] . Bcl-2 family members are classified as either anti-apoptotic ( e . g . , Bcl-2 , Bcl-xL , Bcl-w , and A1 ) or pro-apoptotic , the latter of which can be further divided into the multi-BH domain and BH3-only subgroups . Similar to anti-apoptotic proteins , the multi-BH domain members , such as Bax and Bak , adopt a globular fold consisting of up to nine a-helices that form an extended hydrophobic cleft on the surface [8] . This cleft serves as an authentic BH3-binding site that mediates their translocation and oligomerization within the mitochondrial outer membrane and , ultimately , mitochondrial permeabilization [6 , 9] . Additional regulation of the mitochondrion-dependent apoptosis is mediated by the permeability transition pore ( PTP ) , a complex composed of the voltage-dependent anion channel ( VDAC ) and adenine nucleotide translocator ( ANT ) . VDAC is located in the mitochondrial outer membrane , and its transmembrane can form a barrel with a pore size consistent with the estimated open-channel diameter to allow the escape of cytochrome c . A direct interaction between Bcl-2 family members and VDAC may control cytochrome c permeation across membranes [7 , 10] . Alternatively , VDAC-mediated closure of the PTP supercomplex may prevent ATP/ADP exchange across the membrane , which ultimately results in outer membrane permeabilization and subsequent release of pro-apoptotic factors from the intermembrane space; this , in turn , triggers apoptosis in the cytoplasm or nucleus [10] . Although current studies favor the latter model , these hypotheses may not be mutually exclusive in defining the nature of the PTP complex . Apoptosis is part of host innate immunity that eliminates the propagation of intracellular pathogens such as viruses . As a countermeasure , viruses have evolved to encode numerous open reading frames ( ORFs ) to circumvent cellular apoptotic pathways . Herpesviruses , particularly γ-herpesviruses , including human Epstein-Barr virus , Kaposi sarcoma–associated herpesvirus , and murine γ-herpesvirus 68 ( γHV-68 ) , provide unique models for dissecting the molecular mechanisms of apoptotic regulation . A common aspect of these γ-herpesviruses is that they encode viral homologs of Bcl-2 , designated vBcl-2 , and these vBcl-2 genes effectively prevent cells from undergoing apoptosis upon various stress responses [11 , 12] . vBcl-2 has been speculated to serve a vital role in the virus lifecycle by inhibiting the premature apoptotic death of host cells during acute replication , allowing completion of viral replication cycle and favoring the spread of progeny virus . However , the previous studies have found that γHV-68 vBcl-2 has no role during acute infection but its activity is critical specifically to latent infection [13–15] . Thus , to identify viral protein that evades host apoptosis-mediated innate immunity during acute infection , we searched for mitochondrial proteins within the γHV-68 genome . We report that viral mitochondrial anti-apoptosis protein ( vMAP ) specifically interacts with cellular Bcl-2/Bcl-xL and VDAC and that these interactions effectively dampen host apoptotic processes , which ultimately contributes to efficient lytic replication in culture .
To test the role of the mitochondrion-dependent apoptosis in γHV-68 replication , wild-type ( wt ) and Bax−/−Bak−/− double knockout ( DKO ) murine embryonic fibroblasts ( MEFs ) were infected with a GFP-containing γHV-68 ( γHV-68ΔK3-GFP ) . DKO MEFs supported γHV-68ΔK3-GFP replication with accelerated kinetics as compared with wt MEFs: the virus titer in DKO MEFs was 50–100 times higher than that in wt MEFs ( Figure S1A and S1B ) . Furthermore , as previously shown [16] , DKO MEFs displayed markedly reduced cell death upon γHV-68 replication than did wt MEF ( Figure S1C and S1D ) . These results indicate that mitochondrion-mediated apoptosis plays a negative role in γHV-68 lytic replication , suggesting that γHV-68 needs to deregulate this pathway to maximize its propagation . To search for the potential mitochondrial protein ( s ) encoded by γHV-68 , we employed two computer programs , MITOProt ( http://mips . gsf . de/cgi-bin/proj/medgen/mitofilter ) and PSORT ( http://psort . nibb . ac . jp/ ) , which assess the likelihood of each candidate viral gene product to be targeted to the mitochondrion . This survey identified an M8 gene product [17] that we have named vMAP . vMAP is present in the second exon of ORF57 and shares the identical nucleotide sequence with ORF57 ( Figure 1A ) . However , vMAP has a +1 shift in reference to ORF57 frame , thus encoding a polypeptide of distinct amino acid sequence from ORF57 . vMAP contains 157 amino acids , with a predicted mitochondrial targeting sequence ( MTS ) at its N-terminus and a putative transmembrane domain at its C-terminus ( Figure 1B ) . vMAP protein was readily detected during γHV-68 lytic replication with an apparent molecular weight of 16 kDa ( Figure 2A , left panel ) . However , we failed to detect vMAP protein in γHV-68 latently infected S11 cells under the same conditions ( unpublished data ) . When vMAP expression vector was transfected into NIH3T3 cells , however , vMAP migrated as 7- , 14- , and 16-kDa species ( Figure 2A , right panel ) . Mutational analyses indicated that the 7- and 14-kDa proteins were derived from translational initiation at the third ( Met70 ) and second ( Met21 ) internal initiation codon , respectively ( unpublished data ) . Intracellular fractionation demonstrated that vMAP was present exclusively in the mitochondrion-enriched heavy membrane ( HM ) fraction ( Figure 2A , right panel ) . The position and integrity of the fraction was confirmed by the presence of the mitochondrial resident protein COX4 . Confocal immunofluorescence microscopy also showed that vMAP was present in the cytoplasm extensively colocalized with MitoTracker , a dye that specifically labels mitochondria in living cells , and with Hsp60 and cytochrome c mitochondrial resident proteins ( Figures 2B , S2A , and S2C ) . GFP fusions containing the N-terminal region of vMAP were constructed to define the MTS of vMAP . The N-terminal 40 residues of vMAP were sufficient to target GFP to mitochondria ( Figure 2C ) . Interestingly , this N-terminal sequence of vMAP is predicted to contain α-helical structure , followed by a stretch of positively charged residues that are the potential MTS motif ( Figure 1A ) . In fact , deletion mutations within this motif considerably impaired the mitochondrial localization of GFP fusions containing vMAP N-terminal sequences in confocal microscopy and intracellular fractionation ( Figures 2C and S2B ) . Of note , the vMAP ( 1–30 ) -GFP and vMAP ( 21–49 ) -GFP also showed the nuclear localization , which was likely contributed by GFP fusion ( Figure 2C ) . These results indicate that the N-terminal 40 residues are sufficient for mitochondrial targeting activity . To assess vMAP function in apoptosis , polyclonal NIH3T3/puro and NIH3T3/vMAP stable cell lines were established . The expression of vMAP was confirmed by immunoblotting with anti-vMAP serum as shown in Figure 2D . These cells were treated with various apoptotic agents and stresses ( staurosporine [ST] , TNF-α , vesicular stomatitis virus infection , and nocodazole ) to induce apoptosis , stained with propidium iodide ( PI ) , and then analyzed by flow cytometry . vMAP expression significantly reduced the accumulation of sub-G1 cells that are considered to be apoptotic ( Figure 2D and 2E ) . These results demonstrate that γHV-68 vMAP has robust anti-apoptotic activity toward various apoptotic agents . To investigate the molecular action of vMAP , we tested whether vMAP interacts with cellular apoptotic or anti-apoptotic proteins of the Bcl-2 family . Co-immunoprecipitation analyses showed that vMAP interacted with cellular Bcl-2 in transiently vMAP-expressing 293T cells and in γHV-68-infected NIH3T3 cells ( Figures 3A and S3A ) . In addition , vMAP interaction with cellular Bcl-xL was readily detected in γHV-68-infected NIH3T3 cells ( Figure 3A ) . Despite the equivalent expression of three different species of vMAP in 293T cells , the 16-kDa vMAP predominantly interacted with Bcl-2 , suggesting that the N-terminal sequence of vMAP is likely required for Bcl-2 interaction ( Figure S3A ) . To further define the interaction between vMAP and Bcl-2 family proteins , a mammalian GST fusion protein containing vMAP ( 1–50 ) was coexpressed in 293T cells along with HA-tagged Bcl-2 , Bcl-xL , Bax , Bak , Bid , or Bad . vMAP ( 1–50 ) -GST efficiently interacted with Bcl-2 and Bcl-xL but not with Bax , Bak , Bad , or Bid ( Figure 3B and unpublished data ) . However , vMAPΔ20-GST fusion lacking the N-terminal 20 amino acids completely lost its ability to bind to Bcl-2 and Bcl-xL ( Figure 3C ) . Interestingly , the N-terminal sequence containing the first 20 amino acids of vMAP is predicted to adopt an amphipathic helical structure and share limited similarity with the BH3 peptide of Bad ( unpublished data ) . However , deletion mutation analysis indicated that unlike the BH3 peptide binding that requires the BH1 , BH2 , and BH3 domains of Bcl-2 , the vMAP binding required the Bcl-2 85–186-aa region containing the BH1 and BH3 domains only in living cells ( Figure S3B ) . This was further supported by the results that Bcl-2 G145A mutation in the BH1 domain abolished vMAP binding , whereas Bcl-2 W188A mutation in the BH2 domain did not affect vMAP binding ( Figure 3D ) . By contrast , both mutations of Bcl-2 completely abrogated its Bid-binding activity under the same conditions ( Figure S3C ) . This suggests that Bcl-2 binding to vMAP is different from its binding to BH3-only proteins . Bcl-2 family members are found in the cytoplasm , the endoplasmic reticulum , and the nuclear membrane where they act as sensors of cellular damage or stress . Upon stress , members of Bcl-2 family proteins relocate to the mitochondrial surface where they exert their activity [6 , 18 , 19] . Thus , mitochondrial recruitment of Bcl-2 is considered an important step during the pro- or anti-apoptotic decision . To examine whether vMAP interaction affected Bcl-2 intracellular localization , the distribution of Bcl-2 was examined by subcellular fractionation . Whole-cell lysates were subjected to sequential centrifugation to obtain light membrane ( LM ) fraction containing microsomes derived from the endoplasmic reticulum or the trans Golgi network , HM fraction enriched with mitochondria , and cytosolic fraction . Densitometry quantification of immunoblotting revealed that approximately 70% and 30% of Bcl-2 was present in LM and HM of NIH3T3/puro cells , respectively , whereas 30% and 70% of Bcl-2 was present in the LM and HM of NIH3T3/vMAP cells , respectively ( Figure 4A ) . vMAPΔ20 mutant that failed to interact with Bcl-2 showed only a little effect on Bcl-2 localization compared with wt vMAP ( Figure 4A ) . It should be noted that despite the 20-aa deletion at the N-terminus , vMAPΔ20 mutant was still present primarily in the mitochondrion-enriched HM , as shown in Figures 4A and S2D , suggesting that vMAP may contain at least two independent motifs for its mitochondrial localization . To further test if vMAP recruited Bcl-2 into the mitochondrion in vitro , HM factions were used for mitochondrial association assay with [35S]-labeled Bcl-2 translated in rabbit reticulocyte lysates . Bcl-2 mitochondrial association activity increased approximately 2-fold in the HM fractions of NIH3T3/vMAP cells compared with those of NIH3T3/puro cells ( Figure 4B ) . These results collectively indicate that vMAP actively recruits Bcl-2 to mitochondria , and that this vMAP activity requires the specific interaction with Bcl-2 . The ability to associate with BH3-only molecules correlates with the anti-apoptotic activity of Bcl-2 and Bcl-xL [20 , 21] . Because vMAP facilitated the mitochondrial recruitment of Bcl-2 , we tested whether vMAP expression affected the ability of Bcl-2 and Bcl-xL to associate with BH3-only molecules . NIH3T3/puro and NIH3T3/vMAP cells were transfected with Flag-tagged Bcl-2 and HA-tagged Bad or Bid . At 36 h post-transfection , cell lysates were used for immunoprecipitation with anti-HA , followed by immunoblotting with anti-Flag . Surprisingly , the interaction of Bcl-2 with Bad or Bid was considerably higher in NIH3T3/vMAP cells than in NIH3T3/puro cells ( Figure 5A and 5B ) . vMAP expression also had a similar effect on the Bcl-xL-Bad interaction ( Figure 5C ) . In contrast , vMAP expression affected neither Bcl-2/Bcl-xL interaction with Bax/Bak , nor Bcl-xL dimerization under the same conditions ( Figure S3D , S3E , and unpublished data ) . vMAPΔ20 , which failed to interact with Bcl-2 and Bcl-xL but still localized at mitochondria , had no effect on their interactions with Bid or Bad ( Figure 5A and 5C , lane 4 ) . To further demonstrate the effect of vMAP on Bcl-2 interaction with BH3-only molecules in living cells , a flow cytometry–based fluorescence resonance energy transfer ( FRET ) assay was used to quantify the interaction between Bcl-2 and Bid . NIH3T3/puro , NIH3T3/vMAP , and NIH3T3/vMAPΔ20 cells were transfected with EYFP-Bcl-2 and ECFP-Bid expression vectors . At 36 h post-transfection , these cells were subjected to a FRET assay . The Bcl-2–Bid interaction increased approximately 2- to 3-fold in NIH3T3/vMAP cells as compared with NIH3T3/puro cells , whereas no significant difference was detected between NIH3T3/puro and NIH3T3/vMAPΔ20 cells ( Figure 5D ) . Furthermore , consistent with immunoprecipitation results ( Figure S3D and S3E ) , vMAP did not significantly alter the Bcl-2–Bax interaction in the FRET assay ( Figure S3F ) . These results demonstrate that vMAP specifically facilitates the interactions of Bcl-2/Bcl-xL with Bid/Bad . BH3-only pro-apoptotic proteins transduce death signals from the cell surface or intracellular apoptotic pathways by inducing a conformational change in Bax . Subsequently , Bax translocates to mitochondria and oligomerizes within the outer membrane , which ultimately leads to membrane permeabilization and release of pro-apoptotic factors from the intermembrane space [5 , 22] . To test if vMAP expression affected Bax activation , mouse monoclonal antibody 6A7 that specifically recognizes an epitope in the pro-apoptotic Bax conformer was used to assess the level of Bax activation [23] . NIH3T3/puro and NIH3T3/vMAP cells were treated with ST ( 1 μM ) for 4 h , lysed with 1% CHAPS buffer , and subjected to immunoprecipitation with the 6A7 monoclonal antibody or the P-19 rabbit polyclonal antibody that reacts with total Bax . NIH3T3/vMAP cells reproducibly showed lower levels of the pro-apoptotic Bax conformer ( Figure 6A ) . This difference was not due to a reduced level of Bax expression , as immunoprecipitation and immunoblotting with the P-19 antibody showed the equivalent amounts of Bax in both cells ( Figure 6A ) . Confocal immunofluorescence microscopy also showed that vMAP expression substantially suppressed the ST-induced Bax activation in HeLa cells ( Figure 6B ) . In contrast , vMAPΔ20 , which failed to interact with Bcl-2/Bcl-xL , did not affect Bax activation ( Figure 6B ) . Quantification of 6A7 Bax antibody–positive cells showed that over 70% of vMAPΔ20-expressing cells were positive to the 6A7 Bax conformer antibody at 4 h after ST treatment , whereas only 25% of vMAP-expressing cells were positive ( Figure 6B ) . Subcellular fractionation was further used to examine the mitochondrial translocation of Bax . NIH3T3/puro , NIH3T3/vMAP , and NIH3T3/vMAPΔ20 cells were treated with ST ( 1 μM ) for 4 h , and equivalent amounts of HM were used for immunoblotting . Upon ST treatment , endogenous Bax efficiently translocated into mitochondria in NIH3T3/puro and NIH3T3/vMAPΔ20 cells , whereas a significant reduction of Bax mitochondrial translocation was detected in NIH3T3/vMAP cells ( Figure 6C ) . Equivalent amounts of Bax expression were detected in all three cells ( Figure 6C ) . Collectively , these data indicate that vMAP expression significantly suppresses Bax activation as well as its mitochondrial translocation . While the N-terminal 50 residues of vMAP were sufficient for interacting with Bcl-2 and this interaction exhibited a pleiotropic effect on Bcl-2 family proteins , the loss of Bcl-2 interaction did not completely impede vMAP-mediated anti-apoptosis ( see below ) . This suggests that vMAP might have additional cellular targets to achieve anti-apoptotic activity . To test this idea , we used the yeast two-hybrid screen with vMAP 50–157-aa region as bait to search for vMAP-interacting cellular protein ( s ) . This study identified cellular VDAC1 as a vMAP-interacting protein . An in vitro GST pull-down experiment showed that vMAP specifically bound to cellular VDAC1 ( Figure 7A , right panel , lane 1 ) . vMAP contains two leucine-rich ( LLxL , LIxL , and LxLV ) hydrophobic regions consisting of residues 50–66 and 135–157 ( Figure 7A , dark grey box ) . To test whether these hydrophobic regions mediated the interaction with VDAC1 , bacterial GST-vMAP ( 50–157 ) , GST-vMAP ( 66–135 ) , GST-vMAP ( 50–135 ) , and GST-vMAP ( 66–157 ) fusion proteins were used for in vitro GST pull-down assays , followed by immunoblotting with antibody to VDAC1 . GST fusions containing either hydrophobic region 50–66 or 135–157 effectively interacted with endogenous VDAC1 . GST alone or GST-vMAP ( 66–135 ) did not interact with VDAC1 under the same conditions ( Figure 7A ) . Furthermore , the vMAP L/A mutant carrying the replacement of the leucine and isoleucine residues in both hydrophobic regions with alanines no longer interacted with VDAC1 ( Figure 7B ) . NIH3T3 cells stably expressing wt vMAP , vMAPΔ20 , or vMAP L/A at equivalent levels ( Figure S2E ) were tested for vMAP and VDAC1 interaction . Both vMAP wt and vMAPΔ20 efficiently interacted with VDAC1 , whereas vMAP L/A mutant did not interact with VDAC1 , indicating that the leucine-rich motifs within the hydrophobic regions of vMAP are required for its interaction with VDAC1 ( Figure 7C ) . Finally , vMAP interaction with VDAC1 was readily detected in γHV-68-infected NIH3T3 ( Figure 7D ) . Cellular VDAC1 is located at the mitochondrial outer membrane and has a role in the release of pro-apoptotic factors such as cytochrome c upon apoptotic stress [24] . To test whether the vMAP–VDAC1 interaction affected cytochrome c release , NIH3T3/puro and NIH3T3/vMAP cells were treated with ST for 4 h and subjected to intracellular fractionation , followed by immunoblotting with antibodies to cytochrome c and COX4 . This showed the significant release of mitochondrial cytochrome c to the cytosol in NIH3T3/puro cells , whereas only minimal cytochrome c leaked from mitochondria in NIH3T3/vMAP cells ( Figure 7E ) . Finally , NIH3T3/vMAP L/A cells displayed considerable release of mitochondrial cytochrome c upon ST treatment; however , the extent of release was relatively lower in NIH3T3/vMAP L/A cells than in NIH3T3/puro cells ( Figure 7E ) . Taken together , these results indicate that vMAP interacts with mitochondrial outer membrane VDAC1 and that this interaction robustly inhibits cytochrome c release . The biological significance of the vMAP interactions with Bcl-2/Bcl-xL and VDAC1 in anti-apoptosis was tested by examining the effect of wt vMAP and its mutants ( Δ20 , L/A , Δ20&L/A ) on Bcl-2–Bid interaction , Bax translocation and activation , cytochrome c release , and apoptosis ( PI staining ) . While vMAPΔ20 had no effect on the Bcl-2–Bid interaction and Bax translocation/activation , the vMAP L/A mutant potentiated the Bcl-2–Bid interaction and inhibited Bax translocation/activation as efficiently as wt vMAP ( Figure 8A–8C ) . Functionally , vMAP activity to inhibit cytochrome c release was detectably impaired by both mutations ( Δ20 and L/A ) , consistent with the finding that Bax permeabilizes the mitochondrial outer membrane to release cytochrome c ( Figure 8D ) . These data indicate that vMAP targets two mitochondrial apoptosis checkpoint proteins , Bcl-2/Bcl-xL and VDAC1 , in a genetically separable manner . To further investigate the significance of vMAP interactions with Bcl-2 and VDAC1 in the inhibition of apoptosis , NIH3T3 cells expressing wt vMAP or its mutants were treated with ST for 16 h . ST treatment induced extensive apoptosis in NIH3T3/puro cells , whereas wt vMAP efficiently blocked ST-induced apoptosis ( Figure 8E ) . In contrast , L/A and Δ20 mutations significantly impaired the anti-apoptotic activity of vMAP under the same conditions ( Figure 8E ) . Finally , the Δ20&L/A double mutations completely abrogated the ability of vMAP to inhibit ST-induced apoptosis ( Figure 8E ) . These results indicate that the Bcl-2 interaction displays a more pronounced role in the vMAP-mediated anti-apoptosis than the VDAC1 interaction , and both interactions lead to the comprehensive inhibition of the mitochondrion-mediated apoptosis . To investigate the effect of vMAP on γHV-68 lytic replication , the bacterial artificial chromosome system was used to generate recombinant γHV-68 ΔvMAP that contained the removal of first Met1 and second Met21 residues and the insertion of a stop codon without affecting ORF57 coding sequences ( for details of nucleotide changes , please see Materials and Methods ) . γHV-68 ΔvMAP KanR Bac was initially constructed and subsequently used to generate γHV-68 ΔvMAP and the revertant virus , called γHV-68 Rev , which contained wt vMAP sequence ( Figure S4A ) . vMAP protein was readily detected in wt and revertant γHV-68-infected cells but not in γHV-68 ΔvMAP-infected cells ( Figure S4B ) . To test if vMAP played a role in virus lytic replication , wt γHV-68 , γHV-68 ΔvMAP , and γHV-68 Rev were used to infect NIH3T3 , wt MEF , and Bax−/−Bak−/− DKO MEF cells and their replication kinetics were determined by plaque assay . γHV-68 ΔvMAP replicated at levels that were 10- to 30-fold lower in NIH3T3 and wt MEF cells throughout replication cycle than they were in wt γHV-68 and γHV-68 Rev ( Figure 9A and 9B ) . However , this reduced replication capacity of γHV-68 ΔvMAP was considerably diminished in Bax−/−Bak−/− DKO MEFs: γHV-68 ΔvMAP replicated at similar kinetic and slightly reduced peak titer compared to wt γHV-68 ( Figure 9B ) . To further define the role of vMAP in mitochondrial cell death during viral replication , mouse fibroblast cells were infected with wt γHV-68 or γHV-68 ΔvMAP , and Bax activation was then examined by immunoprecipitation and confocal microscopy with 6A7 Bax conformer antibody . After treatment with ST ( 1 μM , 4 h ) , wt γHV-68-infected cells showed the greatly reduced level of activated Bax compared to mock-infected or γHV-68 ΔvMAP-infected cells ( Figure 9C and 9D ) . These results demonstrate that the vMAP gene is required for efficient γHV-68 lytic replication in normal fibroblast cells , but not in mitochondrial apoptosis-deficient fibroblast cells . This indicates that vMAP may serve a vital role in γHV-68 lytic replication by inhibiting the premature mitochondrial apoptotic death of host cells during acute replication , allowing completion of viral replication cycle .
Here , we report the identification of a novel mitochondrial anti-apoptotic vMAP of γHV-68 , of which its N-terminal MTS is sufficient for the mitochondrial localization . vMAP interacts with Bcl-2 and increases Bcl-2 mitochondrial localization , leading to the neutralization of BH3-only pro-apoptotic molecules . Additionally , vMAP binds to the mitochondrial VDAC1 through its internal and C-terminal hydrophobic sequences , thereby inhibiting cytochrome c release upon apoptotic stresses . Taken together , these data indicate that vMAP engages cellular Bcl-2 and VDAC apoptosis checkpoint proteins to comprehensively inhibit the mitochondrion-mediated intracellular innate immunity , which allows completion of efficient viral lytic replication ( Figure 9E ) . Previous functional studies have classified BH3-only proteins as either death agonists such as Bid and Bim or survival antagonists like Bad [22 , 25] . Recently , this has been formally proposed as the “hierarchy model” , which postulates that the survival antagonist mainly promotes apoptosis by neutralizing anti-apoptotic Bcl-2 members , the death agonist induces a conformational change , oligomerization , and activation of Bax/Bak through a “hit-and-run” mechanism , thereby amplifying apoptotic signaling with a limited amount of cleaved Bid or dephosphorylated Bim [26] . In normal cells , the death agonist is held in check by anti-apoptotic Bcl-2 family proteins . When apoptosis is triggered , the survival antagonist binds to anti-apoptotic Bcl-2 proteins that release the death agonist , which subsequently activates Bax or Bak . This model places the survival antagonist upstream of the death agonist . Thus , BH3-only proteins display a synergistic effect in activating Bax/Bak and inducing apoptosis [22 , 25] . Our data is in support of the hierarchy model regarding the action of BH3-only molecules . Conceivably , vMAP interaction may activate Bcl-2/Bcl-xL to adopt their anti-apoptotic conformation , effectively neutralizing the pro-apoptotic BH3-only molecules through a direct interaction , which ultimately blocks Bax activation . It has been shown that cellular orphan nuclear receptor Nur77 has an activity to modulate Bcl-2 conformation and convert Bcl-2 into a pro-apoptotic molecule [27] . This suggests that vMAP may resemble Nur77 by altering cellular Bcl-2/Bcl-xL conformation , but this activity transforms Bcl-2 into an anti-apoptotic conformation rather than a pro-apoptotic form . However , it should be noted that the vMAP exhibits no specificity in Bcl-2 interactions with the death agonist Bid or the survival antagonist Bad . It is possible that vMAP increases the pool of activated anti-apoptotic Bcl-2 in general , which is reflected by the elevated interaction with Bid as well as Bad . In contradiction to the hierarchy model , BH3-only molecules have recently been proposed to induce apoptosis primarily through neutralization of Bcl-2 anti-apoptotic proteins , but not through activation of Bax/Bak [28] . The reason for this seemingly discrepancy between these two models is not clear and may be derived from various different mutant proteins and functional tests that each study relied on [26 , 28] . Thus , further studies are required to resolve this issue regarding BH3-only molecules . Nevertheless , our data indicate that the vMAP facilitates the interactions of Bcl-2/Bcl-xL with Bid/Bad BH3-only molecules , which neutralizes Bid/Bad pro-apoptotic activity , inhibits Bax activation , and thereby likely raises the threshold for cells to execute apoptosis . In addition to Bax/Bak activation , the PTP complex represents an additional apoptotic checkpoint within the mitochondrial membrane . Along with accessory components , the PTP complex is mainly composed of ANT and VDAC that connect the mitochondrial outer membrane with its inner membrane at contact sites . Although ANT and VDAC may not be required for mitochondrion permeabilization , accumulating data indicate that they are implicated in releasing pro-apoptotic factors from the intermembrane space [7 , 24 , 29 , 30] . Cellular Bcl-2 family proteins and viral polypeptides differentially modulate the PTP complex through a direct interaction with ANT , VDAC , or their accessory components , and therefore regulate apoptosis [7 , 31–33] . Our present study adds vMAP to the expanding family of proteins that influence the permeability transition by virtue of protein–protein interactions . The two vMAP LLxL repeats independently mediate an interaction with VDAC1 that is required to efficiently inhibit cytochrome c release , an indication of the mitochondrial permeability transition . We have defined two functional domains within vMAP: the N-terminal Bcl-2-binding domain and the central/C-terminal VDAC1 interaction domain . It is reasonable to speculate the potential presence of a ternary complex consisting of vMAP , Bcl-2 , and VDAC1 . Given that Bcl-2 physically and functionally interacts with VDAC1 [7] , the introduction of vMAP to this complex may influence the Bcl-2/VDAC1 interaction if a ternary complex is present . However , we observed no detectable effect of vMAP on the interaction between Bcl-2 and VDAC1 ( Figure S5 ) . On the other hand , it is also possible that vMAP independently binds to cellular Bcl-2 or VDAC1 . Consistent with this , the aforementioned two binding domains of vMAP are genetically separable in that mutations within each domain only affect its corresponding interaction , leaving the other interaction intact . Alternatively , the formation of a ternary complex may be dependent on the integrity of a lipid bilayer . Therefore , additional approaches other than traditional immunoprecipitation are required to assess vMAP interactions with Bcl-2 and VDAC1 . While both Bcl-2 and VDAC1 interactions are essential for vMAP-mediated inhibition of apoptosis , Bcl-2 binding seems to be more functionally important in vMAP-mediated inhibition of apoptosis than that of VDAC1 . This is consistent with the findings that Bax and Bak are the essential players that open the mitochondrial gate to the cell death program [6 , 9] . In addition , Bcl-2 family proteins are also important in regulating cytochrome c release during apoptosis . Indeed , vMAPΔ20 that no longer bound to Bcl-2 partially lost its activity to inhibit cytochrome c release ( Figure 8C ) . While vMAP L/A mutant that no longer bound to VDAC1 significantly failed to block cytochrome c release , a detectable amount of cytochrome c was still retained in the mitochondrion in these cells ( Figure 7E ) . These data support the idea that vMAP interaction with Bcl-2 also plays a role in the inhibition of cytochrome c release . Taken together , these results indicate that vMAP interactions with both Bcl-2 and VDAC1 synergistically contribute to its inhibition on the mitochondrion-mediated apoptosis . The N-terminal MTS of vMAP entirely overlaps with its Bcl-2 binding motif . This suggests that the N-terminal deletion mutation , vMAPΔ20 , may ablate two functions of vMAP simultaneously: Bcl-2 binding and mitochondrial targeting . However , the results of cell fractionation and confocal microscopy showed that the majority of vMAPΔ20 still localized to the mitochondrion ( Figures 4A and S2D ) . This indicates that vMAPΔ20 mutation ablates only the Bcl-2 binding activity without significantly affecting vMAP mitochondrial localization . This also suggests the presence of additional mitochondrial targeting sequence . Unfortunately , our GFP fusion strategy failed to identify additional linear sequence for its mitochondrial localization ( unpublished data ) . This implies that , perhaps , a higher order of structure of vMAP may be required to function as an MTS as seen with VDAC [34] . The cleavable N-terminal MTS generally sends protein into the matrix or interior membrane of mitochondria , while the uncleavable N-terminal MTS targets protein into the outer membrane of mitochondria [35] . In fact , vMAP N-terminal MTS appeared to be not cleaved based on its molecular weight in SDS-PAGE ( Figures 2A and 4A ) . Additionally , a mutation at the potential mitochondrial signal peptide cleavage site of vMAP did not affect its mitochondrial localization ( unpublished data ) . These results collectively suggest that an alternative MTS exists in addition to the N-terminal MTS . The expression profile and function activity of γHV-68 vMAP has unique features . First , vMAP is present within the second exon of the ORF57 transcript , which encodes an immediate early gene product that is essential for viral replication , a homolog of herpes simplex virus ICP27 [17 , 36] . This suggests that the vMAP gene has co-evolved with ORF57 and likely plays a critical role in γHV-68 replication . Second , the putative N-terminal amphipathic α-helical region of vMAP has both mitochondrial targeting and Bcl-2-binding activities . Of note , the vMAP contains an unidentified MTS , in addition to its N-terminal MTS , that potentially targets vMAP to the outer membrane of mitochondria . Finally , vMAP targets two mitochondrial proteins , Bcl-2 and VDAC1 , to affect a comprehensive inhibition of the mitochondrion-dependent apoptosis . Interestingly , human cytomegalovirus ( HCMV ) vMIA uses a similar strategy to deregulate mitochondrion-dependent apoptosis: vMIA neutralizes and inhibits permeability transition pore activity through its direct interactions with Bax [32 , 37 , 38] . Intriguingly , vMIA is also encoded within the first exon of UL37 of HCMV , an immediate early gene transcript that is required for viral replication [39 , 40] . However , inactivating the vMIA expression by mutagenesis did not dramatically reduce HCMV lytic replication , similar to what we reported here for vMAP deletion for γHV-68 lytic replication in tissue culture [41] . Thus , it is surprising that these viruses have undergone convergent evolution , evolving independently to encode mitochondrial proteins with similar molecular mechanisms but without discernable sequence similarity .
NIH3T3 , NIH3T12 , BHK21 , COS-1 , and 293T cells were grown in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% fetal bovine serum , 2 mM l-glutamine , 100 U/ml streptomycin and penicillin . BJAB and S11 cells were grown in RPMI 1640 supplemented with 10% fetal bovine serum , l-glutamine , and antibiotics . Fugene 6 ( Roche ) , lipofectamine ( Invitrogen ) , or calcium phosphate ( Clontech ) was used for transient expression of vMAP in COS-1 , NIH3T3 , and 293T cells . NIH3T3 stable cell lines were established using a standard protocol of puromycin selection at 2 μg/ml . wt MEFs or Bax−/−Bak−/− DKO MEFs transformed with SV40 were generously provided by Stanley Korsmeyer ( Dana-Farber Cancer Institute , Harvard Medical School ) and grown in complete DMEM medium . γHV-68 wt and GFP viruses were amplified in NIH3T12 cells . After three rounds of freeze-thawing , supernatants of virus-infected cells were used for de novo infection of BHK21 or NIH3T12 cells . Plaque assays were performed in BHK21 cells overlaid with 1% methylcellulose as described previously [42] . γHV-68ΔK3-GFP was kindly provided by Dr . Philip Stevenson . To generate vMAP deletion recombinant virus , the bacterial artificial chromosome [43] containing the entire γHV-68 genome was used . Due to the fact that the vMAP coding sequence overlaps with ORF57 , we generated γHV-68 ΔvMAP by changing the first and second initiation codons of vMAP to AAG and introducing two stop codons ( T > A changes at nucleotide positions of 76016 , 76049 , 76076 , and a G > A change at nucleotide position of 76079 based on Entrez accession number U97553 ) . These changes did not affect the ORF57 coding sequence . Initially , the vMAP coding sequence was replaced with a KanR/LacZ cassette ( γHV-68 Bac ΔvMAP KanR ) by allelic exchange [42] ( Figure S4A ) . Since ORF57 was required for γHV-68 replication , vMAP sequences containing various mutations were amplified by PCR and electroporated into BHK21 cells together with γHV-68 Bac ΔvMAP KanR DNA to allow homologous recombination . Viruses obtained from BHK21 cells , γHV-68 ΔvMAP or γHV-68 Rev , were plaque purified and further amplified in BHK21 cells ( Figure S4A ) . Virion DNAs were used for PCR amplification to clone ORF57 and vMAP DNA fragments , followed by DNA sequence to confirm the presence of designed changes and the absence of aberrant changes . Virion DNA was also digested with BamHI or BglII to exclude the possibility of viral genomic rearrangement ( Figure S4C ) . Viruses were then used to infect NIH3T3 or BHK21 cells to determine replication kinetics . His-tagged vMAP ( 1–135 ) was purified from E . coli strain BL21 using Ni-conjugated agarose according to manufacturer's instructions ( Qiagen ) . Urea was removed using Centricon ( Millipore ) , after which protein concentration was measured with the Bradford method ( Bio-Rad ) . Polyclonal antibodies against His-tagged vMAP ( 1–135 ) were produced in two rabbits . Unless specified , all constructs were derived from pcDNA5/FRT/TO ( Invitrogen ) or pEF-IRES-puro [44] . A DNA fragment corresponding to the γHV-68 vMAP coding sequence was amplified from S11 genomic DNA by PCR and cloned into pcDNA5/FRT/TO between BamHI and XhoI . To generate GFP fusion proteins , vMAP or vMAP ( 1–49 ) and its derivatives were PCR amplified and ligated to pEGFPN1 digested with BglII and SalI . Mutations in the vMAP gene were generated by PCR using oligonucleotide-directed mutagenesis . All constructs were sequenced with an ABI PRISM 377 automatic DNA Sequenzer . Constructs expressing the HA-tagged and GFP-tagged Bcl-2 family proteins were kindly provided by J . Marie Hardwick ( John Hopkins University ) , Beth Levine ( UT-Southwestern Medical Center ) , and Richard J . Youle ( National Institutes of Health ) , respectively . HA-Bcl-xL and Flag-Bax were amplified by two-step PCR using GFP-Bcl-xL and GFP-Bax as templates and cloned into pcDNA5/FRT/TO between the BamHI and XhoI sites . Bcl-2 was cloned into pEYFP-C1 at the BamHI site , and Bax and Bid were cloned into pECFP-C1 between the BamHI/XhoI sites , respectively . To produce His-tagged vMAP protein from E . coli , vMAP ( 1–135 ) was PCR amplified and cloned into pQE40 ( Qiagen ) . To express GST fusion proteins in mammalian cells , vMAP sequences were PCR amplified and cloned into pDEF3 digested with KpnI and EcoRV . For immunoblotting , polypeptides were resolved by SDS-PAGE and transferred to PVDF membranes ( Bio-Rad ) . Immunoblot detection was performed with anti-V5 ( 1:5000; Invitrogen ) , M2 anti-Flag ( 1:5000; Sigma ) , anti-GST ( 1:2000; Santa Cruz Biotechnology ) , anti-VDAC ( porin 31HL ) ( 1:1000; CALBIOCHEM ) , anti-Bax ( 6A7 , 1:1000 , Pharmingen ) , anti-COX4 ( 1:100; Clontech ) , anti-Bcl-2 ( 1:100 , Santa Cruz Biotechnology ) , anti-Bcl-xL ( 1:100 , Pharmingen ) , or anti–cytochrome c ( 1:200 , Clontech; 1:1000 , Pharmingen ) . Rabbit anti-vMAP serum was diluted 1:3000 for immunoblots . Proteins were visualized using a chemiluminescence detection reagent ( Pierce ) and quantified using a Fuji Phosphor Imager . For immunoprecipitation , cells were harvested and then lysed with 1% CHAPS buffer ( Cell Signaling ) supplemented with 1 mM dithiothreitol ( DTT ) and protease inhibitor cocktail ( Roche ) . After pre-clearing with protein A/G–agarose beads for 1 h at 4 °C , whole-cell lysates were used for immunoprecipitation . Generally , 1–4 μg of commercial antibody or 1 μl of vMAP antiserum was added to 1 ml cell lysate , which was incubated at 4 °C for 8 to 12 h . After addition of protein A/G–agarose beads , the incubation was further extended for 1 h . The beads were extensively washed with lysis buffer , and the immunoprotein precipitates were eluted with SDS loading buffer by boiling for 5 min . Sixteen hours after transfection or electroporation , cells were fixed with 4% paraformaldehyde for 15 min , permeabilized with 0 . 2% ( v/v ) triton X-100 for 15 min , blocked with 10% goat serum in PBS for 30 min , and reacted with diluted primary antibody in PBS for up to 2 h at room temperature . After incubation , cells were washed extensively with PBS , incubated with the appropriate secondary antibody diluted in PBS for 30 min at room temperature , and washed three times with PBS . Mitochondrial staining was performed with 250 nM MitoTracker ( Molecular Probes ) for 20 min followed by washing with PBS at room temperature for 5 min . Alternatively , mitochondria were visualized using antibodies against mitochondrial proteins including cytochrome c and HSP60 . Confocal microscopy was performed using a Leica TCS SP laser-scanning microscope ( Leica Microsystems ) fitted with a 100× Leica objective ( PL APO , 1 . 4NA ) and Leica imaging software . Images were collected at 512 × 512-pixel resolution . The stained cells were optically sectioned in the z-axis , and the images in the different channels ( photo multiplier tubes ) were collected simultaneously . The step size in the z-axis varied from 0 . 2 to 0 . 5 μm to obtain 16 slices per imaged file . The images were transferred to a Macintosh G4 computer ( Apple Computer ) and Photoshop ( Adobe ) was used to render the images . For immunofluorescence microscopy , antibodies were directed against cytochrome c ( 1:200 ) , Bax 6A7 ( 1:100; Pharmingen ) , HSP60 ( 1:100; Santa Cruz Biotechnology ) , vMAP ( 1:100 ) , V5 ( 1:100; Invitrogen ) , or VDAC 31HL ( 1:100; CALBIOCHEM ) . All conjugated secondary antibodies were obtained from Molecular Probes and diluted at 1:1000 or 1:500 . These included Alexa 488–conjugated goat anti–rabbit IgG , Alexa 568–conjugated goat anti–mouse IgG , Alexa 488–conjugated goat anti–mouse IgG , Alexa 594–conjugated donkey anti–goat IgG , Alexa 488–conjugated donkey anti–rabbit IgG , and Alexa 594–conjugated donkey anti–mouse IgG . In vitro GST pull-down was similar as previously described [45] . Briefly , E . coli ( BL21 ) cells were induced with IPTG ( 1 μM ) for 2 h and lysed with PBS containing 0 . 1% sarcosyl . GST or GST fusions were purified with glutathione-conjugated Sepharose beads and either used for immediate experiments or stored at −20 °C for future experiments . Then , whole-cell lysates were mixed with loaded glutathione-conjugated Sepharose and binding was extended at 4 °C for up to 2 h . After extensive washing , protein precipitates were resolved by SDS-PAGE and transferred to PVDF membrane , followed by immunoblotting . For mammalian GST pull-down , 293T cells expressing GST fusion proteins and Bcl-2 family proteins were harvested and lysed with 1% CHAPS buffer ( 50 mM HEPES [pH 7 . 4]; 100 mM NaCl; 10 mM Tris; 1 mM EDTA; 1% CHAPS ) supplemented with protease inhibitor cocktail ( Roche ) . Supernatants after centrifugation procedures were pre-cleared with 15 μl of protein A/G beads at 4 °C for 1 h , after which 40 μl of 50% glutathione-conjugated Sepharose beads was added and binding was extended for 2 to 3 h at 4 °C . Protein precipitates were washed extensively with lysis buffer and analyzed by immunoblotting . The yeast two-hybrid screen was performed as previously described [45 , 46] . Yeast strain AH109 bearing the Gal4-vMAP ( 50–157 ) fusion gene plasmid was used to screen a cDNA library generated from EBV-immortalized B cells . To obtain the mitochondrion-enriched HM fraction , 293T or NIH3T3 cells were harvested and washed with ice-cold PBS and resuspended with hypotonic buffer ( 10 mM Tris; 250 μM sucrose; 20 mM HEPES [pH 7 . 4]; 0 . 2 mM EDTA ) supplemented with 1 mM DTT and protease inhibitor cocktail . The suspension was incubated on ice for 15 min and lysed with Nitrogen Bomber ( Parr Instrument Company ) at 250 psi for 15 min . Nuclei and unbroken cells were removed by centrifugation at 700g for 10 min . The supernatant at this point was regarded as whole cell lysate and was subjected to centrifugation at 6 , 000g for 15 min . The pellet was then resuspended with hypotonic buffer and centrifuged at 6 , 000g for 15 min; this process was repeated twice to obtain the mitochondrion-enriched fraction . The supernatant was further centrifuged at 100 , 000g for 90 min to yield the cytosolic fraction ( supernatant ) , and the pellet was collected and regarded as the light membrane fraction . When cytochrome c release was examined , the step to obtain light membranes was skipped . Stably transfected NIH3T3 cells were grown in complete DMEM with 2 μg/ml puromycin . Cells , 1 × 106 per well , were used for apoptosis induction with 1 μM ST for up to 16 h . After treatment , cells were harvested with cell dissociation buffer ( Sigma ) and fixed with 70% ethanol overnight . After staining the cells with PI for 30 min at room temperature , DNA content was measured with flow cytometry and analyzed with CellQuest ( BD Biosciences ) . Alternatively , cells were treated with other apoptogenic agents ( TNF-α plus cycloheximide , VESICULAR STOMATITIS infection , nocodazole ) , and sub-G1 cells were measured as described above . For cytochrome c release , NIH3T3 cells were induced with 1 μM of ST for 4 h and harvested . The mitochondrion-enriched fraction was obtained as described above . Cytochrome c and mitochondrial membrane marker COX4 were detected by immunoblotting using the antibody from the ApoAlert kit ( Clontech ) . For Bax mitochondrial translocation , cells were stimulated with 1 μM ST for 4 h and processed for intracellular fractionation as described above . For Bax activation , cells were stimulated with 1 μM ST for 4 h , and whole-cell lysates in 1% CHAPS buffer were precipitated with the mouse 6A7 monoclonal antibody or the rabbit P-19 polyclonal antibody , and protein precipitates were analyzed by immunoblotting . For γHV-68ΔK3-GFP infection in wt and Bax−/−Bak−/− DKO MEF cells , the activated caspase 3 was measured by intracellular staining according to manufacturer's instruction ( BD Pharmingen ) , followed by flow cytometry analysis . Meanwhile , dead cells were stained with trypan blue and counted under light microscope . Bcl-2 was in vitro translated from TNT-coupled reticulocyte lysate systems ( Promega ) and labeled with [35S]-methionine/cysteine . The mitochondrion-enriched HM fraction was obtained as described above from NIH3T3/puro or NIH3T3/vMAP cells . Protein was measured with the Bradford method and 50 μg of mitochondrial proteins were mixed with 100 μl of reticulocyte lysates containing the radioactively labeled Bcl-2 . The mixtures were incubated at 30 °C for 2 h and mitochondria were pelleted at 13 , 000 rpm for 15 min . While the supernatants were collected as membrane-free Bcl-2 , the pellets were washed twice with hypotonic buffer and resuspended in PBS as mitochondrion-associated Bcl-2 . Proteins were then analyzed by gradient ( 4%–12% ) SDS-PAGE and Phosphor Image reader . NIH3T3/vec , NIH3T3/vMAP , and NIH3T3/vMAPΔ20 were transfected with pEYFP-Bcl-2 and pECFP-Bid . At 36 h post-transfection , cells were harvested with dissociation buffer and washed twice with room temperature PBS containing 0 . 5% bovine serum albumin . Next , cells were subjected to flow cytometry that is equipped with fluorochrome excitation capability .
The Entrez Nucleotide ( http://www . ncbi . nlm . nih . gov/sites/entrez ? db=Nucleotide ) accession numbers for the sequences discussed in this paper are vMAP ( U97553 ) ; human Bcl-2 ( NM_000633 ) ; human Bcl-xL ( NM_138578 ) ; mouse Bax ( BC053380 ) ; human Bak ( U16811 ) ; mouse Bid ( U75506 ) ; mouse Bad ( L37296 ) .
|
Apoptosis is a conserved cell death program that contributes to restriction of viral replication and elimination of infected cells . Whether triggered via internal inducers such as DNA damage or via external stimuli such as engagement of the death receptor , apoptosis takes place through a cascade of regulated internal proteolytic digestion , resulting in a collapse of cellular infrastructure , mitochondrial potential , genomic fidelity , and cell membrane integrity . Indeed , apoptosis represents a predominant form of virally infected cell demise . In response , viruses have evolved numerous ways of circumventing this host-cell apoptosis . Most of the DNA viruses including murine γ-herpesvirus 68 ( γHV-68 ) are genetically equipped with anti-apoptotic ability to ensure viral replication and propagation . The authors have identified a new viral mitochondrial protein ( vMAP ) of γHV-68 that interacts with Bcl-2 and voltage-dependent anion channel 1 ( VDAC1 ) in a genetically separable manner . These interactions markedly suppress Bax mitochondrial translocation and activation and inhibit cytochrome c release , leading to the comprehensive inhibition of mitochondrion-mediated apoptosis . The authors also demonstrate that vMAP gene is required for efficient γHV-68 lytic replication in normal cells , but not in mitochondrial apoptosis-deficient cells . These findings are entirely novel and significantly advance our understanding of how virus escapes host intracellular apoptosis-mediated innate immunity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"mus",
"(mouse)",
"viruses",
"eukaryotes",
"virology"
] |
2007
|
A Novel Inhibitory Mechanism of Mitochondrion-Dependent Apoptosis by a Herpesviral Protein
|
Schistosoma japonicum is a major public health concern in the Peoples' Republic of China ( PRC ) , with about 800 , 000 people infected and another 50 million living in areas at risk of infection . Based on ecological , environmental , population genetic and molecular factors , schistosomiasis transmission in PRC can be categorised into four discrete ecosystems or transmission modes . It is predicted that , long-term , the Three Gorges Dam ( TGD ) will impact upon the transmission of schistosomiasis in the PRC , with varying degree across the four transmission modes . We undertook longitudinal surveillance from 2002 to 2006 in sentinel villages of the three transmission modes below the TGD across four provinces ( Hunan , Jiangxi , Hubei and Anhui ) to determine whether there was any immediate impact of the TGD on schistosomiasis transmission . Eight sentinel villages were selected to represent both province and transmission mode . The primary end point measured was human incidence . Here we present the results of this five-year longitudinal cohort study . Results showed that the incidence of human S . japonicum infection declined considerably within individual villages and overall mode over the course of the study . This is also reflected in the yearly odds ratios ( adjusted ) for infection risk that showed significant ( P<0 . 01 ) downward trends in all modes over the follow-up period . The decrease in human S . japonicum incidence observed across all transmission modes in this study can probably be attributed to the annual human and bovine PZQ chemotherapy . If an increase in schistosome transmission had occurred as a result of the TGD , it would be of negligible size compared to the treatment induced decline seen here . It appears therefore that there has been virtually no immediate impact of the TGD on schistosomiasis transmission downstream of the dam .
Zoonotic schistosomiasis , caused by Schistosoma japonicum , is a chronic debilitating disease in the south of the People's Republic of China ( PRC ) , with about 800 , 000 infected and 65 million people at risk of infection . [1] , [2] The majority of transmission occurs in the lake and marshland areas of Jiangxi and Hunan , and in Jiangsu , Anhui and Hubei provinces; schistosomiasis is also endemic in the hilly and mountainous regions of Sichuan and Yunnan ( Figure 1 ) . [1] , [3] , [4] Based on ecological and environmental factors and Oncomelania snail population genetics , Davis et al . [3] , [4] categorised schistosomiasis transmission in the PRC into four discrete ecosystems or ecogenetic transmission modes . The characteristics of modes I–III are presented in Table 1 . The Three Gorges Dam ( TGD ) is one of several huge engineering projects transforming China's environment . [5] It is located in the Three Gorges region in the upper reaches of the Yangtze River ( the world's third-largest river; 5920 km long ) ( Figure 1 ) . It spans the Yangtze at Sandouping Island , just west of the city of Yichang in Hubei province . [1] , [6] , [7] The main justification for the dam is flood control; by regulating water flow it is designed to prevent disastrous floods that have occurred every decade along the lower plains regions of the Yangtze River . [1] , [6] , [7] Construction commenced in 1994 and by 2003 the TGD was closed to a height of 135 metres . In 2009 , it reached its full height of 185 metres and its hydropower station began to generate 18 , 600MW of power . [8] , [9] By 2009 the 2 , 300 m long dam had resulted in a 600 km long serpentine reservoir that inundated 115 , 000 acres of cultivated land , requiring the resettlement of some two million people [9] Long-term the dam is predicted to impact with varying degrees upon the transmission of schistosomiasis across the four transmission modes [6]–[8]; modes I–III are located downstream with mode IV upstream of the dam . In mode I , it is predicted that long term the TGD will reduce annual flooding along the Yangtze , thus stabilising the populations of Oncomelania hupensis snails – the intermediate hosts of S . japonicum – resulting in increased transmission of schistosomiasis . In mode II , there will be increased permanent marshlands in some areas with increasing snail population stability; in other areas near the Yangtze River there is predicted to be continuing but somewhat decreased snail population instability associated with snail transport . Overall , in the long-term it is predicted that snail populations will increase considerably with a resultant increase in S . japonicum prevalence for mode II . In mode III , new snail habitats will be created due to the degeneration of rice paddies into marshlands as a result of underground water levels rising , with long-term predicted increased S . japonicum transmission . [7] , [8] We recently described the results [4] of 5-year longitudinal surveillance of S . japonicum transmission we undertook in Shian , in the Anning River Valley , a schistosomiasis-endemic village located upstream of the TGD in Sichuan province , typical of mode IV . The results of this study showed no effect on transmission over the study period [4] , thereby corroborating our prediction of little or no immediate impact of the dam on schistosomiasis transmission in Sichuan . Nevertheless , we recommended continued surveillance as changes in transmission patterns may take upwards of 10 years to be realised as the water flow slows down and silt deposits settle , forming new marshland areas suitable for the propagation of Oncomelania snails . Snail dispersal and population movements will also be required to introduce schistosomes into this locality . [4] Here we report the results of a similar study we undertook over the same period ( 2002–2006 ) in eight sentinel villages , representative of transmission modes I–III , downstream of the TGD located in Hunan , Jiangxi , Hubei and Anhui provinces .
We carried out a prospective longitudinal cohort study ( 2002–2006 ) in eight villages ( Table 1 ) , representative of schistosomiasis transmission modes I–III , to determine the potential impact of the closure of the TGD on schistosome incidence over time . At baseline , two stool samples were collected and a questionnaire administered to all subjects usually resident in the eight study villages . Stool samples were examined microscopically using the Kato-Katz thick smear technique , with three slides per stool read blinded , to determine S . japonicum prevalence and intensity of infection . [10] The questionnaire consisted of questions relating to demographics , medical history and history of water contact . [11] , [12] A stool sample was also collected from all bovines ( water buffaloes and cattle ) in the study villages and examined for S . japonicum prevalence using the miracidial hatching test ( 3 individual hatches read blind; 50 grams of faeces per hatching ) and intensity of infection , using a traditional Chinese sedimentation method . [11] Following the baseline survey , a fixed cohort of all individuals aged 5–65 in each village was monitored for schistosome infection for the duration of the study . The cohort inclusion criteria were that an individual: a ) must have been a resident of the village for more than 12 months; b ) should be aged 5–65 years; c ) did not intend migrating out of the village for the next 4 years; and d ) should continuously reside in the village for the study period . A water contact questionnaire , consisting of questions relating to each participant's yearly water exposure by season , was administered to all cohort members annually . [12] Two stool samples were collected from all cohort subjects and one stool sample was obtained from all bovines to determine outcome measures; these included incidence and intensity of infection for cohort members , and infection rates and intensity of infection for bovines . At baseline , all village residents and bovines found positive for S . japonicum were treated with praziquantel ( PZQ ) ( humans: 40 mg/kg; bovines: 25 mg/kg ) , in accord with WHO recommendations , [11] , [13] until the infection was cleared so that no fecal eggs were present . Our previous schistosomiasis studies around the Poyang and Dongting Lakes showed 85–95% efficacy for a single PZQ dose in humans and bovines , with 100% efficacy following re-examination and re-treatment . [13]–[16] At follow-up , all cohort members and all bovines found schistosome egg-positive were again treated with the same WHO recommended dose of PZQ until cleared of infection . A snail survey was performed annually in April for each village to measure the prevalence of infection in snails and the density of infected snails per unit area . The survey used a Chinese random quadrat sampling method ( 0 . 11 metres2 sized frames , 20 metres between frames ) . [17] Water level readings were taken every 7–10 days from hydrological stations close to each village in modes I and II and directly measured in the canals within the villages in mode III . These were collected for the duration of the study ( 2002–2006 ) . A MS ACCESS based database was designed specifically for this project and was used for data management . [18] A positive human schistosome infection was indicated by the presence of at least one egg in any Kato-Katz smear . Egg counts were transformed to eggs per gram and geometric mean intensities were calculated using the log-transformed egg counts . Confidence intervals ( CIs ) were calculated using standard formulae based on the binomial distribution ( annual incidence of infection ) and the lognormal distribution ( intensity ) . Each cohort member was assigned a water contact score for each year preceding infection status assessment . This was determined by adding season-specific sub-scores based on the frequency of water contact obtained through the water contact questionnaires . Formal analyses of annual human incidences , both crude and adjusted ( for water contact , using the water contact score ) , used a generalized linear model ( GLM ) with a logit link and a binomial error distribution . Generalised equation estimators of parameters with an unstructured variance-covariance matrix were used to account for repeated measures on individuals over time . Analyses used the GENMOD procedure of SAS software ( version 9 . 1; SAS Institute , Inc , Cary , NC ) to calculate odds ratios ( OR ) and 95% confidence intervals ( 95% CIs ) .
The prevalence and density of infected snails fluctuated substantially over the five year study period and consistent trends could not be determined for any of the transmission modes . This was probably because of the high levels of snail sampling variability due to spatial aggregation effects that we have observed previously . [19] Water level patterns in transmission modes I and II were similar over the duration of the study with peak levels in the summer coinciding with the rainy season . The levels in mode III fluctuated by about 50 cm around the maximum height of the canal systems in Maling and Guhu over the course of the study .
We describe the results of 5-year ( 2002–2006 ) longitudinal surveillance in eight sentinel villages representative of modes I–III of the four S . japonicum transmission modes described by Davis et al . [3] These villages are located downstream of the TGD in four provinces ( Hunan , Jiangxi , Hubei and Anhui ) and our aim was to determine whether there was any impact on S . japonicum transmission following closure of the dam and the commencement in 2003 of filling of the Three Gorges reservoir . [6] The incidence of human S . japonicum infection declined considerably in all surveyed villages over the course of the study ( Tables 3 , 4 ) . This was also reflected in the yearly adjusted odds ratios for infection risk which indicated significant ( P<0 . 01 ) downward trends in all three transmission modes over the follow-up period ( Figure 3 ) . Regression analyses also showed that modal trends were significantly different from one another , thus indicating that the degree of decline in each mode was heterogeneous . The greatest decline was in mode I ( 73 . 4% ) , followed by modes II ( 57 . 3% ) and III ( 48 . 9% ) . The decrease in human incidence observed in transmission modes I–III may be attributable to the annual PZQ treatment , which on ethical grounds , had to be administered to all infected individuals . Major flooding of the lakes and marshland areas downstream of the TGD can drown adult snails resulting in decreased transmission , [1] but the water level records indicated that no major flood event had occurred during the study period . A number of reports have shown that bovines are the major transmission source for human schistosomiasis in the lakes and marshland areas of Southern China , and that interventions targeting bovines can reduce the incidence of human infection . [20]–[22] Over the follow-up period ( 2003–2006 ) , bovine infection rates decreased in transmission modes I and III , but remained relatively stable for mode II , whereas infection intensity decreased in transmission modes I and II , but remained stable for mode III . Therefore , the decrease in human S . japonicum incidence observed across the three transmission modes can probably be attributed to the annual human and bovine PZQ treatment . The differences in the downward trends evident between modes I–III may be due to the varying declines in bovine infection rates and infection intensity . Mode I had the greatest downward trend and declines in both bovine infection and intensity of infection , Mode II showed a decline only in bovine infection intensity , and whereas the infection rate declined in mode III , the infection intensity remained stable . It is noteworthy that there was no decrease in human exposure as water contact patterns did not change over the duration of the study , indicated by the similarity in crude and adjusted ( for water contact ) ORs . If an increase in schistosome transmission had occurred as a result of the TGD , it would have been negligible compared with the treatment-induced decline we observed . It appears , therefore , that there had been no or only very limited impact of the TGD on schistosomiasis transmission downstream of the dam over the 2002–2006 study period . It is well recognised that schistosomiasis emergence or re-emergence has resulted following other large-scale hydro-projects such as the Gezira-Managil Dam in Sudan , the Aswan Dam in Egypt , the Melkasadi Dam in Ethiopia , and the Danling and Huangshi Dams in China . [23]–[26] It has been predicted that the TGD will alter water and sand distributions downstream that will have a significant impact on ecological systems; these include the Dongting and Poyang lakes and the canals of Hubei , where S . japonicum transmission is generally projected to increase , although decreased transmission is projected for other locations . [1] , [6] , [23] , [24] Specifically , it is anticipated that the TGD will result in large changes to the flow , depth and sedimentation load of the Yangtze so that the distribution and numbers of schistosome-infected Oncomelania snails will be altered , increasing transmission of schistosomiasis in some areas and its re-introduction into others where the infection is currently under control . [1] , [5]–[8] , [23] , [24] Mathematical modelling has suggested a marked elevation in snail-breeding areas , increased infection rates of S . japonicum and greater associated morbidity . [7] , [16] Seto and colleagues [6] believe that the lower more stable water levels downstream created by the TGD will result in decreased overall snail densities , but that the density of infected snails and corresponding human infections may increase due to the co-location of bovine grazing areas , snail habitats , and human activity that may occur with the more stable water levels . [1] , [6] The “Return Land to Lake” program currently underway in the PRC will significantly extend the water surface area in Dongting and Poyang Lakes , with the result that large numbers of farmers and fishermen are being resettled closer to lake water and Oncomelania snail habitats . [1] , [24] This will also likely impact on schistosomiasis transmission as water contact and the prevalence and intensity of infection will increase . [1] , [25] , [27] Another important consideration is that the TGD reservoir , which has submerged wholly or in part 13 cities and 466 towns , has displaced up to 2 million people [5] , [9] from non-endemic schistosomiasis localities upstream of the dam . Many of these villagers , having been relocated to downstream schistosome-endemic areas near lakes and wetlands in the Yangtze River Basin , will have no immunity to schistosomiasis , and hence will readily acquire a schistosome infection on exposure , and likely develop severe disease as a result . Whilst no immediate effect of the TGD on schistosome transmission was evident in this study it may be that the predicted changes will take longer to eventuate . The dam , with its 1080 km2 reservoir , reached its full height in 2009 and this may be the critical time point that marks the start of environmental changes that will begin to impact on S . japonicum transmission . Continued surveillance should be undertaken to monitor the future ecological impacts of the dam . [28] , [29] Accordingly , we have commenced a new study to monitor environmental changes and undertake longitudinal surveillance ( 2010–2014 ) of infection rates and intensity of S . japonicum infection in snails , humans and bovines in a further eight villages below the TGD to determine its effect on schistosome transmission dynamics . Findings from the study will be of considerable relevance for the PRC and other settings where schistosomiasis is endemic and where large water resource development projects are planned or are underway .
|
Schistosomiasis , caused by Schistosoma japonicum , is a significant parasitic disease and public health problem in China . How the parasite is transmitted there can be categorized into four distinct modes ( modes I–IV ) and it is predicted that the Three Gorges Dam , recently completed , will affect the way schistosomiasis is spread in these modes . We monitored transmission for a 5-year period ( 2002–2006 ) in eight villages , representative of the three modes ( I–III ) below the dam across four provinces ( Hunan , Jiangxi , Hubei and Anhui ) to determine whether there was any immediate impact of the dam on schistosomiasis spread . Human schistosomiasis incidence declined considerably within individual villages and each mode , and the yearly odds ratios ( adjusted ) for infection risk showed significant downward trends in all three modes over the follow-up period . The decreased human S . japonicum incidence recorded across transmission modes I–III was probably attributable to annual human and bovine praziquantel drug treatment . If an increase in schistosome transmission had occurred as a result of the dam , it would be of negligible size compared with this treatment-induced decline . There had thus been virtually no immediate impact of the TGD on schistosomiasis transmission downstream of the dam over the 5-year surveillance period .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"public",
"health",
"and",
"epidemiology",
"infectious",
"disease",
"epidemiology",
"microbiology",
"parasitic",
"diseases",
"parasitology",
"preventive",
"medicine",
"neglected",
"tropical",
"diseases",
"infectious",
"diseases",
"environmental",
"epidemiology",
"epidemiology",
"biology",
"public",
"health",
"schistosomiasis"
] |
2012
|
Five-Year Longitudinal Assessment of the Downstream Impact on Schistosomiasis Transmission following Closure of the Three Gorges Dam
|
Animals use taste to sample and ingest essential nutrients for survival . Free fatty acids ( FAs ) are energy-rich nutrients that contribute to various cellular functions . Recent evidence suggests FAs are detected through the gustatory system to promote feeding . In Drosophila , phospholipase C ( PLC ) signaling in sweet-sensing cells is required for FA detection but other signaling molecules are unknown . Here , we show Gr64e is required for the behavioral and electrophysiological responses to FAs . GR64e and TRPA1 are interchangeable when they act downstream of PLC: TRPA1 can substitute for GR64e in FA but not glycerol sensing , and GR64e can substitute for TRPA1 in aristolochic acid but not N-methylmaleimide sensing . In contrast to its role in FA sensing , GR64e functions as a ligand-gated ion channel for glycerol detection . Our results identify a novel FA transduction molecule and reveal that Drosophila Grs can act via distinct molecular mechanisms depending on context .
Animals use gustatory systems to evaluate the quality of food . Gustation is essential not only to prevent ingestion of toxic chemicals but also to ensure ingestion of essential nutrients such as sugars , amino acids , and lipids . The detection and consumption of energy-dense foods can confer a survival advantage , especially when food is scarce . Lipids are more calorie-rich than proteins or sugars , so it is unsurprising that lipid sensing has emerged as a new candidate taste modality in addition to the five basic taste modalities in mammals: sweet , umami , bitter , sour , and salt . Dietary lipid sensing was thought to be mediated by texture and olfaction [1–3] , but the recently discovered taste receptors for fatty acids ( FAs ) in mammals indicate gustatory systems can also detect lipids [4 , 5] . Two G-protein coupled receptors ( GPCRs ) , GPR40 and GPR120 , are present in the taste receptor cells of mammals [5 , 6] and are partly requried for FA preference [5] . FA-induced responses depend on phospholipase C ( PLC ) and its downstream signaling molecules like transient receptor potential channel type M5 ( TRPM5 ) [7] , suggesting that FA taste is mediated by a phosphoinositide-based signaling pathway . Drosophila melanogaster can detect several taste modalities including sweet , bitter , salt , and amino acids [8 , 9] . Most taste modalities are detected by the direct activation of ion channels expressed in gustatory receptor neurons ( GRNs ) . The 68 members of the gustatory receptor ( Gr ) gene family in the Drosophila genome include the main taste receptors for the sweet and bitter modalities [10 , 11] . Although GRs have seven transmembrane domains , these proteins are not GPCRs . They have an opposite membrane topology [12 , 13] and function as ligand-gated ion channels [14 , 15] . Ionotropic receptors ( Irs ) , which are distantly related to ionotropic glutamate receptors [16] , are involved in the detection of low salt , pheromones , polyamines , and amino acids [17–20] . In contrast to other taste modalities , Drosophila FA taste signaling is mediated by the PLC pathway [21] . Mutation of norpA , a Drosophila orthologue of PLC , results in reduced attraction to FAs . The introduction of a norpA cDNA into sweet GRNs of norpAP24 flies rescues their deficit in FA sensing , suggesting PLC in sweet GRNs is essential for FA sensing . FA detection requires PLC signaling in sweet GRNs , but no other signaling molecules have yet been implicated . Here , we show that Gr64e , which is known as a glycerol receptor [22] , is required downstream of PLC for the detection of FAs . The precise deletion of the Gr64 cluster via CRISPR/Cas9 reduces FA palatability . By screening individual Gr64 cluster gene mutant flies , we identified a requirement for Gr64e in FA sensing . We also found the re-introduction of Gr64e into Gr64 cluster deletion mutants rescues their behavioral attraction to FAs and FA-evoked action potentials . Gr64e seems to function as a ligand-gated ion channel for glycerol sensing because the co-expression of Gr64e and Gr64b confers glycerol responses independent of PLC on sweet GRNs , the low-salt sensing GRNs , and bitter GRNs of Gr64 cluster mutant flies . In contrast , the introduction of TrpA1 , which can couple to PLC signaling [23 , 24] , in sweet GRNs of flies lacking Gr64e rescues their deficit in FA sensing but not glycerol sensing . In addition , Gr64e expression in TrpA1 mutants can only rescue their deficit in aristolochic acid ( ARI ) sensing [23] , which is PLC-dependent . Gr64e expression does not rescue the TrpA1 mutant defect in N-methylmaleimide ( NMM ) sensing , which proceeds via direct TRPA1 activation [25] . Together , our results reveal a novel component in Drosophila for signal transduction in FA detection and suggest Drosophila Grs can function via multiple molecular mechanisms depending on their cellular and molecular context .
We were prompted to test whether the Gr64 cluster is involved in FA sensing because the Gr64 cluster is required for the detection of most phagostimulatory substances [26–31] . The Gr64 cluster comprises six tandem Gr genes ( Gr64a-Gr64f ) transcribed as a polycistronic mRNA ( Fig 1A ) [26 , 29 , 31] . Because deletion of the whole Gr64 cluster ( ΔGr64 ) is lethal due to the additional deletion of neighboring genes [31] , we used CRISPR/Cas9 to generate a new Gr64 cluster deletion ( Gr64af ) covering only the Gr64 cluster coding region ( Fig 1A ) . We confirmed the deletion of the Gr64 loci by genomic PCR and DNA sequencing ( Fig 1A ) . In contrast to ΔGr64 , Gr64af is viable and fertile . As expected , we found Gr64af flies show a reduced proboscis extension reflex ( PER ) to sucrose , glucose , fructose , trehalose , and glycerol ( Fig 1B ) . PER responses to low salt are slightly increased compared to wild-type ( Fig 1C ) , suggesting Gr64af does not have a general defect in gustatory function . Furthermore , optogenetic activation of sweet GRNs expressing red activatable channelrhodopsin ( ReaChR ) [32] induces PER in wild-type and Gr64af flies ( Fig 1D ) , confirming that sweet GRNs of Gr64af are functional . We , next asked whether the Gr64 cluster is required for FA sensing . Although wild-type flies show a robust PER response to hexanoic acid ( HxA ) , octanoic acid ( OcA ) , oleic acid ( OA ) , and linoleic acid ( LA ) , Gr64af flies show severely reduced PER responses to all the FAs we tested ( Fig 1E ) . We were also able to confirm that the other sweet Grs ( Gr5a , Gr43a , and Gr61a ) are not required for FA sensing ( Fig 1F ) . To determine which of the six Grs in the Gr64 cluster are required for FA sensing , we examined PER responses to HxA in flies carrying mutations in the individual genes of the Gr64 cluster ( S1 Fig ) . norpAP24 flies , which carry a mutation in the Drosophila orthologue of PLC [33] , show reduced PER responses to HxA like Gr64af flies ( Fig 2A ) [21] . Of the various Gr64 cluster mutants , we found Gr64cLEXA and Gr64eLEXA flies show reduced PER responses to HxA like the norpAP24 and Gr64af mutants ( Fig 2A ) . To confirm the requirement of Gr64c and Gr64e for HxA sensing , we further characterized the Gr64c and Gr64e mutants . Although Gr64cLEXA flies show reduced PER responses to HxA , glycerol , and sucrose ( Fig 2B ) , the expression of a Gr64c cDNA in Gr64cLEXA flies using Gr5a-GAL4 , which labels sweet GRNs [34] , does not rescue this defect . This suggests the Gr64cLEXA phenotype cannot be attributed to the loss of Gr64c in labellar sweet GRNs . This result is also consistent with the strong FA preference of ΔGr64a2 flies , which harbor a deletion of the protein-coding sequence of Gr64a and Gr64b as well as a third of the protein-coding sequence of Gr64c at its N-terminus ( S1 Fig , Fig 2A ) . Gr64e is known as a glycerol receptor [22] . Gr64eLEXA flies show reduced PER responses to glycerol and to several FAs ( i . e . , HxA , OcA , OA , and LA ) ( Fig 2C and 2D ) . Expression of a Gr64e cDNA in the Gr64e mutant background using Gr5a-GAL4 rescues glycerol and FA responses to wild-type levels , indicating Gr64e is required for both glycerol and FA detection ( Fig 2C and 2D ) . In addition , the expression of Gr64e using Gr5a-GAL4 rescues the HxA responses of Gr64af flies , suggesting Gr64e is the only Gr in the Gr64 cluster required for FA sensing ( Fig 2E ) . Silencing the labellar Gr64e-expressing GRNs by expression of the potassium channel Kir2 . 1 [35] abolishes PER to HxA , suggesting that preference to HxA is mediated by Gr64e-expressing GRNs ( S2 Fig ) . To better understand FA sensing in the labellum , we examined electrophysiological responses to HxA . HxA elicits action potentials mainly in S-type sensilla of wild-type flies ( Fig 3A ) . In a few cases , we also observed HxA-evoked firing in I-type sensilla , but such responses were rare . Consistent with our PER results , we did not observe any responses to HxA in Gr64af or Gr64eLEXA flies ( Fig 3B and 3C ) . Gr64cLEXA flies show robust , wild-type-like HxA responses , indicating that the reduced attraction of Gr64cLEXA flies to HxA cannot be attributed to a peripheral defect in FA detection ( Fig 3B and 3C ) . In addition , Gr5a-GAL4-driven expression of Gr64e in Gr64eLEXA and Gr64af flies restores HxA-evoked action potentials , which suggests Gr64e is the only Gr in the Gr64 cluster required for FA sensing ( Fig 3D and 3E ) . Gr64e is required in GRNs for electrophysiological and behavioral responses to glycerol [22] . To determine whether the molecular function of Gr64e is the same in the detection of glycerol and FAs , we next asked whether PLC is required for glycerol sensing . We found no difference between wild-type and norpAP24 flies in glycerol-evoked action potentials or PER responses ( Fig 4A–4C ) . This indicates Gr64e plays distinct molecular roles in the detection of glycerol and FAs . It remains unclear whether Gr64e alone is sufficient for glycerol detection . Ectopic expression of Gr64e in olfactory receptor neurons confers glycerol responses [27] , but Gr64e requires Gr64b as a co-receptor to confer glycerol responses on sweet GRNs [36] . To address this ambiguity , we used Gr5a-GAL4 or Ir76b-GAL4 , which labels low-salt sensing GRNs [20] , to misexpress Gr64b alone , Gr64e alone , or Gr64b and Gr64e together in sweet GRNs or low-salt sensing GRNs of Gr64af flies , respectively . The misexpression of Gr64b and Gr64e together confers glycerol sensitivity in both sweet GRNs and low-salt sensing GRNs of Gr64af flies ( Fig 4D–4G ) . Co-expression of Gr64b and Gr64e together in sweet GRNs of Gr64af flies restores their PER responses to glycerol ( Fig 4H ) . In addition , introduction of Gr64b and Gr64e in bitter GRNs of Gr64af flies under the control of Gr66a-GAL4 , which labels bitter GRNs [34] , confers glycerol response ( S3 Fig ) . These data suggest glycerol detection occurs through the direct activation of heteromeric ion channels formed by Gr64b and Gr64e . Although both Gr64e and PLC are required for FA detection in sweet GRNs , it is unclear how they function together . It is possible that Gr64e acts as a GPCR that detects HxA and functions upstream of PLC . This is unlikely , however , because sweet GRNs of L-type sensilla expressing Gr64e do not respond to HxA . To exclude the possibility that sweet GRNs of L-type sensilla lack other factors required for PLC signaling , we used Gr5a-GAL4 to express either Gαq/norpA or Gr64e/Gαq/norpA in sweet GRNs . Neither of these combinations confers HxA responsiveness on the sweet GRNs of L-type sensilla ( S4 Fig ) . A second hypothesis relating the function of Gr64e to PLC is that Gr64e functions downstream of PLC . Drosophila trpA1 is expressed in a subset of bitter GRNs and required for avoidance to NMM [25] , a tissue damaging reactive electrophile and ARI [23] , a plant drived antifeedant . TRPA1 can be activated directly by NMM[25] and has also been associated with PLC signaling in ARI avoidance [23] . We hypothesize that if both TRPA1 and GR64e function downstream of PLC , TRPA1 and GR64e should be able to substitute for one another with regard to PLC signaling . We misexpressed either the thermosensory isoform TrpA1 ( B ) or the chemosensory isoform TrpA1 ( A ) in sweet GRNs of Gr64af flies to explore whether TRPA1 can replace the function of GR64e in FA sensing but not glycerol detection . We found TrpA1 expression in sweet GRNs of Gr64af flies rescues HxA-evoked electrophysiological responses in their S-type sensilla and their HxA-evoked PER responses ( Fig 5A–5C , S5 Fig ) . It does not , however , rescue glycerol detection . Furthermore , we also confirmed that functional replacement of GR64e with TRPA1 was dependent on PLC . Expression of TrpA1 or Gr64e in sweet GRNs of norpAP24 , Gr64af double mutant flies does not restore the response to HxA ( S6 Fig ) . We next asked whether GR64e can replace the function of TRPA1 in sensing noxious chemicals . We found that ARI elicits similar electrophysiological responses in wild-type and TrpA11 flies expressing Gr64e in their bitter GRNs ( Fig 5D and 5E ) . TrpA11 flies expressing Gr64e in bitter GRNs do not , however , respond to NMM , a direct TRPA1 activator . These data further support Gr64e acts downstream of PLC for FA detection .
Here , we show that Gr64e—a sweet clade Gr required for glycerol detection [22]—is also essential for the gustatory detection of FAs . Although Gr64e is required in sweet GRNs for the detection of both glycerol and FAs , the molecular mechanisms by which it does so are different . Glycerol evokes action potentials in sweet GRNs in L- , I- , and S-type sensilla in a PLC-independent manner ( Fig 4A and 4B ) [22] . Freeman et al . reported that single sweet GRs alone confer the responses to various sugars including glycerol when they mis-express them in olfactory neurons [27] . Only the combination of Gr64b and Gr64e , however , confers glycerol responsiveness on the sweet GRNs [36] , low-salt sensing GRNs , and bitter GRNs of Gr64af flies . This suggests Drosophila GRs form heteromeric complexes for sensing sugars . Since Gr64b/Gr64e-misexpressing low-salt sensing GRNs or bitter GRNs produce fewer glycerol-evoked action potentials than sweet GRNs , we speculate that there are unknown additional Grs in sweet GRNs that facilitate the formation of high affinity glycerol receptors . This would be similar to our findings with the L-canavanine receptor [15] . Based on the characterization of GRs for bitter sensing [15 , 37] , the detection of glycerol occurs through the direct activation of ion channels formed by Gr64b and Gr64e ( Fig 6A ) , but it remains unclear whether unknown intracellular signaling components also contribute to the function of sweet GRs . FAs selectively activate sweet GRNs in S-type sensilla in a PLC-dependent manner . Of the nine sweet clade Grs ( i . e . , Gr5a , Gr43a , Gr61a , and Gr64a-f ) , only Gr64e is required for FA detection . Gr64e seems unlikely to be a FA receptor for several reasons . First , the sweet GRNs in L- and I- type sensilla , where endogenous Gr64e is expressed [28] , respond only to glycerol , not FAs ( Fig 3 ) . Second , overexpression of G-protein signaling components ( Gαq and norpA ) alone or together with Gr64e ( Gr64e , Gαq , and norpA ) in sweet GRNs of L-type sensilla does not endow FA sensitivity ( S4 Fig ) . Finally , although there are reports that the distantly related olfactory receptors function as both GPCRs and ionotropic receptors [38 , 39] , the inverse topology of GRs relative to GPCRs is further evidence that Gr64e is unlikely a direct FA receptor . We were unable to exclude the possibility that Gr64e acts as an accessory protein for an unknown FA-responsive GPCR or the possibility that the absence of other accessory proteins ( i . e . , CD36 [40] ) in sweet GRNs of L-type sensilla explains their inability to respond to HxA . Furthermore , the functional redundancy we identified between GR64e and TRPA1 in PLC-specific functions ( e . g . , FA but not glycerol detection by GR64e and ARI but not NMM detection by TRPA1 ) suggests Gr64e functions downstream of PLC ( Fig 6B ) . Although GR64e and TRPA1 are functionally interchangeable downstream of PLC , it remains unclear whether they share the same molecular mechanism of activation . GR64e can be activated by hydrolysis of phosphoinositide by PLC , elevation of intracellular calcium , or diacylglycerol . Alternatively , Gr64e may be a voltage-gated channel that is not directly coupled to the PLC pathway . Two Drosophila species , D . psedoobscura and D . persimilis carry pseudogenized versions of Gr64e and do not respond to glycerol [22] . If these two species have also lost gustatory sensitivity to FAs , it will confirm the evolutionary conservation of this dual function for Grs . Because this is the first time a Drosophila GR has been found to function downstream of PLC , our results extend the molecular repertoire of the GR family of proteins . This is particularly intriguing because there are Grs expressed in the antenna [28 , 41] and in the enteroendocrine cells of the gut [42] . Rather than acting in the direct detection of ligands in these non-gustatory cells , these GRs may mediate novel sensory modalities via distinct molecular mechanisms . FAs act as sources of energy , but also as structural components of membranes . In addition , they have multiple biological roles in metabolism , cell division , and inflammation [43] . In flies , changes in the FA composition of membranes via FA deprivation influences cold tolerance and synaptic function [44 , 45] . Dietary FAs also modulate mitochondrial function and longevity [46] . Thus , animals must ingest dietary FAs for survival . Indeed , regular laboratory Drosophila foods also contain FAs [45] . It is unsurprising that FA taste is well-conserved between mammals and flies , which are required for PLC pathway in contrast to other taste modalities in flies . Since GPR40 and GPR120 are strong FA receptor candidates in mammals [5] , an FA-sensitive GPCR may also be selectively expressed in the sweet GRNs of S-type sensilla in flies . It will be interesting to determine whether the Drosophila orthologue of the mammalian FA receptor or any other GPCRs are involved in FA detection .
Flies were maintained on cornmeal-molasses-yeast medium at 25°C and 60% humidity with a 12h/12h light/dark cycle . The fly medium recipe is based on the Bloomington recipe ( https://bdsc . indiana . edu/information/recipes/molassesfood . html ) and composed of 3% yeast ( SAF Instant Yeast ) , 6% cornmeal ( DFC-30102 , Hansol Tech , Korea ) , 8% molasses ( extra fancy Barbados molasses , food grade , Crosby Molasses Co . , Ltd . of Canada ) , and 1% agar ( DFA-30301 , Hansol Tech ) for the nutrients and the hardener . It also includes 0 . 8% Methyl 4-hydroxybenzoate ( H5501 , Sigma-Aldrich , Saint Louis , MO ) , 0 . 24% propionic acid ( P1386 , Sigma-Aldrich ) , and 0 . 0028% phosphoric acid ( 695017 , Sigma-Aldrich ) as preservatives . For optogenetic experiments , instant fly food was purchased from Carolina ( Burlington , NC , #173200 ) . Gr64d1 was described previously [47] . Gr5a-GAL4 , Gr66a-GAL4 , Gr43aGAL4 , Gr5aLEXA , Gr64aGAL4 , Gr64bLEXA , Gr64cLEXA , Gr64eLEXA , and Gr64fLEXA were provided by H . Amrein . ΔGr64a1 , ΔGr64a2 , and ΔGr61a1 were provided by J . Carlson . UAS-Gr64b , UAS-Gr64c , and UAS-Gr64e were provided by A . Dahanukar . Gr64ab , Ir76b-GAL4 , and TrpA11 were provided by C . Montell , UAS-TrpA1 ( A ) 10a , UAS-TrpA1 ( A ) 10b , and UAS-TrpA1 ( B ) 10a were provided by P . Garrity , and LexAop-Kir2 . 1 was provided from B . Dickson , respectively . UAS-ReaChR ( BL53741 ) , norpAP24 ( BL9048 ) , UAS-norpA ( BL26273 ) , UAS-Gαq ( BL30734 ) , Gr64e-GAL4 ( BL57667 ) , and UAS-Kir2 . 1 ( BL6595 ) were obtained from the Bloomington Stock Center . nos-Cas9 ( #CAS-0001 ) was obtained from NIG-FLY . All the mutant lines and transgenic lines were backcrossed for five generations to the w1118 control genotype . For clarity , the w1118 line is referred to as wild-type throughout the manuscript . We used CRISPR/Cas9 system to generate Gr64af flies [48] . We selected two target sites for deletion of the whole Gr64 cluster using DRSC Find CRISPRs ( http://www . flyrnai . org/crispr ) and CRISPR optimal target finder ( http://tools . flycrispr . molbio . wisc . edu/targetFinder ) : one near the 5’ end of Gr64a ( GAATCCTCAACAAACTTCGGTGG , the Protospacer Adjacent Motif is underlined ) and one near the 3’ end of Gr64f ( GGTCGTTGTCCTCATGAAATTGG ) . We synthesized oligomers and cloned them into the BbsI site on pU6-BbsI-ChiRNA ( Addgene #45946 ) . After injecting two pU6-ChiRNA targeting constructs into nos-Cas9 embryos at 500 ng/μl each , we screened the resulting flies for deletions via PCR of genomic DNA isolated from the G0 generation . The primers we used for deletion confirmation were as follows: TCTCGGCAGCTAATCGAAAT and GCGACCATTCTTTGTGGAAT . We collected 3–5-day-old flies in fresh food for 24 hours . Then , we starved them for 18 hours in vials containing 1% agarose . After anaesthetizing the flies on ice , we mounted them on slide glasses with melted 1-tetradecanol ( 185388 , Sigma-Aldrich ) . We then allowed the flies to recover for 1–2 hours and ensured they were satiated with water before the assay . For each test solution , we used a 1 ml syringe with a 32-gauge needle to apply a single droplet directly to the labellum . We dissolved FAs in 4% ethanol . Each experimental group contained 24 flies , half were mated males and half were mated females , attached to a slide glass . All PER experiments were performed at the same time to eliminate any circadian effects . We report PER responses as the number of responding flies/total flies . We performed tip recordings as previously described [49 , 50] . Briefly , we immobilized 5–7-day-old flies by inserting a reference electrode—a glass capillary filled with Ringer’s solution—through the thorax and into the labellum . Then , we stimulated the indicated labellar sensilla with a recording electrode ( 10–20 μm tip diameter ) containing test chemicals in 30 mM tricholine citrate ( TCC ) as the electrolyte . After connecting the recording electrode to a 10X preamplifier ( TastePROBE; Syntech , Hilversum , The Netherlands ) , we recorded action potentials at 12 kHz with a 100–3 , 000 Hz band-pass filter using a data acquisition controller ( Syntech ) , sorted the spikes based on amplitude , and analyzed them with the Autospike 3 . 1 software package ( Syntech ) . We purchased hexanoic acids ( 153745 ) , octanoic acids ( 2875 ) , oleic acids ( 01008 ) , linoleic acids ( L1376 ) , sucrose ( S9378 ) , α-D-glucose ( 158968 ) , D- ( - ) -fructose ( F3510 ) , D- ( + ) -trehalose dihydrate ( 90210 ) , glycerol ( G9012 ) , N-methylmaleimide ( 389412 ) , aristolochic acid I ( A5512 ) , and tricholine citrate ( T0252 ) from Sigma-Aldrich . Sodium chloride ( S0520 ) was purchased from Duchefa Biochemie ( Haarlem , Netherland ) . 3–4-day-old flies were transferred to vials containing instant Drosophila medium with or without 400 μM all trans-retinal ( R2500 , Sigma-Aldrich ) , respectively . After feeding the flies retinal for a week , they were mounted into 200 μl pipette tips . Then , they were exposed to LED light ( wavelength of 627 nm ) . PER responses were monitored by video camera and counted manually . We performed all statistical analyses using SPSS Statistics 23 ( IBM Corporation , Armonk , NY ) . We tested normality and homoscedasticity using the Kolmogorov-Smirnov and Levene tests . PER responses are displayed as means ± SEM . We used unpaired Student’s t-tests or one-way ANOVAs with Tukey post-hoc tests to analyze the PER data . All electrophysiological data are presented as medians with quartiles . We used the Mann-Whitney U-test or Kruskal-Wallis test with Mann-Whitney U post-hoc tests to determine whether the medians for each genotype were significantly different .
|
Fatty acids ( FAs ) are energy-rich nutrients that are detected through the gustatory system to promote feeding . Here , we show FA detection requires a Drosophila gustatory receptor , Gr64e . Although GR64e functions as a ligand-gated ion channel for glycerol detection , in FA sensing , it acts downstream of phospholipase C signaling . We identified a novel signaling molecule for FA sensing in Drosophila . Furthermore , our findings suggest Drosophila GRs have multiple modes of action depending on their cellular and molecular context .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"action",
"potentials",
"membrane",
"potential",
"social",
"sciences",
"electrophysiology",
"neuroscience",
"animals",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"ion",
"channels",
"monomers",
"(chemistry)",
"experimental",
"organism",
"systems",
"g",
"protein",
"coupled",
"receptors",
"drosophila",
"research",
"and",
"analysis",
"methods",
"polymer",
"chemistry",
"lipids",
"taste",
"proteins",
"chemistry",
"transmembrane",
"receptors",
"biophysics",
"insects",
"arthropoda",
"physics",
"biochemistry",
"signal",
"transduction",
"psychology",
"eukaryota",
"cell",
"biology",
"ligand-gated",
"ion",
"channels",
"physiology",
"glycerol",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"sensory",
"perception",
"fatty",
"acids",
"neurophysiology",
"organisms"
] |
2018
|
Drosophila Gr64e mediates fatty acid sensing via the phospholipase C pathway
|
Leptospira interrogans are pathogenic spirochetes responsible for leptospirosis , a worldwide reemerging zoonosis . Many Leptospira serovars have been described , and prophylaxis using inactivated bacteria provides only short-term serovar-specific protection . Therefore , alternative approaches to limit severe leptospirosis in humans and morbidity in cattle would be welcome . Innate immune cells , including macrophages , play a key role in fighting infection and pathogen clearance . Recently , it has been shown that functional reprograming of innate immune cells through the activation of pattern recognition receptors leads to enhanced nonspecific antimicrobial responses upon a subsequent microbial encounter . This mechanism is known as trained immunity or innate immune memory . We have previously shown that oral treatment with Lactobacillus plantarum confers a beneficial effect against acute leptospirosis . Here , using a macrophage depletion protocol and live imaging in mice , we established the role of peritoneal macrophages in limiting the initial dissemination of leptospires . We further showed that intraperitoneal priming of mice with CL429 , a TLR2 and NOD2 agonist known to mimic the modulatory effect of Lactobacillus , alleviated acute leptospiral infection . The CL429 treatment was characterized as a training effect since i . ) it was linked to peritoneal macrophages that produced ex vivo more pro-inflammatory cytokines and chemokines against 3 different pathogenic serovars of Leptospira , independently of the presence of B and T cells , ii . ) it had systemic effects on splenic cells and bone marrow derived macrophages , and iii . ) it was sustained for 3 months . Importantly , trained macrophages produced more nitric oxide , a potent antimicrobial compound , which has not been previously linked to trained immunity . Accordingly , trained macrophages better restrict leptospiral survival . Finally , we could use CL429 to train ex vivo human monocytes that produced more cytokines upon leptospiral stimulation . In conclusion , host-directed treatment using a TLR2/NOD2 agonist could be envisioned as a novel prophylactic strategy against acute leptospirosis .
Leptospira belong to the phylum Spirochetes and are the etiological agents of leptospirosis , a reemerging worldwide zoonosis [1] . Pathogenic leptospires , including Leptospira interrogans , can infect and colonize multiple hosts , and leptospirosis constitutes a global burden affecting both human health and agricultural economics . Leptospires enter their hosts through abraded skin or mucosa and then disseminate in blood , potentially leading to renal infection [2] . Leptospirosis is underdiagnosed at its onset because of unspecific symptoms common to other viral and bacterial infections , such as fever , headaches and jaundice . Leptospirosis can cause multiorgan failure in humans and leads to a mortality rate of 6–10% . More than 300 serovars of pathogenic leptospires have been described [2] . Thus , vaccination is possible only in certain contexts , like in endemic areas when the circulating serovars are known . A few human vaccines based on inactivated whole bacteria are available , but they confer only serovar-specific and short-term protection [3] . In the absence of prophylactic drugs , preventive antibiotic therapy and wearing adapted clothes constitute the most effective prevention against this neglected disease . Recently , we have developed a mouse model of leptospirosis using bioluminescent L . interrogans [4] . Upon intraperitoneal ( IP ) administration , bacteria seem to be initially controlled , but after 2 days post-infection , they are found in blood where they progressively replicate until 3 to 4 days post-infection . The bacteria are then eliminated from the blood circulation but progressively reappear in the kidneys . L . interrogans establish their niche in the proximal renal tubules , and live spirochetes excreted during urination contribute to the spread of the disease . Chronically infected rodents , such as rats and mice , are the most prominent reservoir hosts of leptospires , although other mammals , including humans can also shed Leptospira in urine [5] . Pattern recognition receptors ( PRRs ) are a subset of membrane-associated or cytosolic proteins of the innate immune system that sense microbe-associated molecular patterns ( MAMPs ) [6] . MAMPs are molecules that are absent in the host but essential and ubiquitously present in microbes , such as lipopolysaccharide for Gram-negative bacteria , peptidoglycan and lipoproteins . MAMPs are recognized by PRRs , which trigger antimicrobial responses , such as the production of detrimental compounds for invaders , i . e . , nitric oxide ( NO ) , reactive oxygen species ( ROS ) and antimicrobial peptides . Moreover , PRR activation triggers the production of pro-inflammatory cytokines and chemokines that prime the immune system , recruit phagocytes to the site of infection and initiate an adaptive response . Altogether , these responses usually result in inflammation and clearance of the invading agent [7 , 8] . We have shown that L . interrogans escape some pathways of innate immune signaling . In humans , for instance , Toll-like receptor 4 ( TLR4 ) , the PRR-sensing bacterial lipopolysaccharide ( LPS ) , cannot detect leptospiral LPS , whereas its murine counterpart is able to sense it [9] . Consistently , in opposition to wild-type ( WT ) mice , which are resistant to acute leptospirosis , TLR4-sensing deficiency results in sensitivity to infection with L . interrogans [10] , such as in C3H/HeJ mice [11] . Additionally , we have recently shown that leptospires also escape the signaling from Nucleotide-binding Oligomerization Domain-containing protein ( NOD ) 1 and NOD2 [12] , the cytosolic receptors of muropeptides , which are the building blocks of bacterial peptidoglycan [13] . PRRs are expressed in phagocytes , and their stimulation is important for the activation of immune cells such as macrophages [7] . The role of macrophages during leptospirosis is not clear [14] , and several ex vivo experiments have suggested that leptospires can escape the macrophage response or survive within these cells in naive hosts [15–17] . It has also been reported that NO produced by macrophages and other cells is a potent antileptospiral effector , but it is also deleterious for the host as it was linked to nephritis [18] and renal fibrosis [19] . The consequences of PRR escape by Leptospira on the functions of phagocytes are not yet known . Recently , we have shown that repeated oral administration of Lactobacillus plantarum ( L . plantarum ) to C3H/HeJ mice alleviates acute leptospirosis and that this effect is associated with myeloid cells [20] . Therefore , we wondered whether this protection could be due to an innate immune memory effect . Indeed , recent studies have indicated that immune cells that have first been exposed to certain MAMPs respond differently than naïve cells upon a secondary encounter with the same or different MAMPs [21] . Not all MAMPs are able to induce such “memory” , and of note , different MAMPs can have positive or negative impacts for the host during a second encounter [22] . The positive or enhanced response has been coined innate immune memory ( also known as trained immunity ) , whereas the negative effect is named tolerance [23] . Interestingly , trained immunity is a sustained long-term effect that can lead to protection against different pathogens , and it has been proposed as a vaccination strategy [24–26] . The cells involved in this response are innate immune cells such as macrophages , monocytes and natural killer cells ( NK ) [26] . Monocytes [27] and NK cells [28] from Bacille Calmette Guerin ( BCG ) -vaccinated patients show an enhanced response to mycobacterial and non mycobacterial challenges up to 3 months post-vaccination . In addition , BCG vaccination protects against unrelated lethal infection with Candida albicans [29] in SCID mice lacking functional B and T cells , showing that this effect is independent of the adaptive immune system . Of note , this effect can be recapitulated ex vivo using human monocytes trained with several MAMPs [22] . TI has been associated with PRR engagement as well as metabolic and epigenetic rewiring of innate immune cells [30–32] . The administration of L . plantarum and other probiotic bacteria demonstrates immunomodulatory properties [29 , 33] . For instance , L . plantarum pretreatment protects B-cell deficient mice against lethal pneumonia virus ( PMV ) [34] . Moreover , using mice deficient for TLR2 and NOD2 receptors , the L . plantarum protective effect against PMV was shown to be dependent on both TLR2 and NOD2 , and it was mimicked with a bi-functional TLR2/NOD2 ligand [35] . These data suggest an innate immune memory mechanism in which PRR pre-engagement enhances host antimicrobial responses upon secondary encounters . Therefore , in the present study , we hypothesized and showed that pretreatment with CL429 , a chimeric TLR2/NOD2 agonist [36] , could mimic the protective effect observed with L plantarum and alleviate the acute phase of leptospirosis due to innate immune memory .
The Leptospira interrogans used in this work were grown in Ellinghausen-McCullough-Johnson-Harris ( EMJH ) medium at 28°C without agitation and diluted weekly to obtain early stationary phase growth cultures at the time of the in vitro and in vivo experiments . Leptospira interrogans serovars Manilae L495 ( derivative bioluminescent strain MFlum1 ) [4] , Copenhageni strain Fiocruz and Icterohaemorraghiae strain Verdun were used in this work . Bacteria were counted in a Petroff-Hauser chamber and diluted in PBS ( D-PBS Lonza ) before use . Eight-week-old female or male mice of the following genotypes were used: C57BL/6J ( Janvier Labs , Le Genest Saint Isle , France ) , Albino C57BL/6 B6 ( Cg ) -Tyrc-2J/J ( Charles River laboratory , Saint-Germain-Nuelles , France ) and Rag2-γc ( Institut Pasteur animal facility ) . Mice were sublethally infected with 1 or 5x107 bioluminescent L . interrogans strain MFlum1 diluted in 200 μL endotoxin free D-PBS ( Lonza ) , intraperitoneally ( IP ) . Mice were monitored daily for weight variation and clinical signs of leptospirosis . For depletion experiments , liposomes containing PBS or 5 mg/mL clodronate ( Encapsula Nano Science ) were injected into the peritoneal cavity in a volume of 100 μL per mouse . Liposomes were administered 3 and 1 days before sublethal IP infection with bioluminescent L . interrogans . Priming of innate immune cells was performed in mice by IP injection of CL429 ( InvivoGen ) 2 and 1 week prior to infection , as previously reported by Rice et al . , [35] . CL429 stock was solubilized in DMSO at a concentration of 5 mg/mL and further diluted in PBS . Mice were injected IP with 25 μg/mouse CL429 in 200 μL of PBS containing 2 . 5% DMSO . Control mice were injected with PBS containing 2 . 5% DMSO . Seven days after the second treatment , mice were infected with L . interrogans or euthanized by cervical dislocation to collect cells for ex vivo stimulation and flow cytometry analysis . Mice were imaged as previously described [4] . Twenty minutes after infection , mice were injected IP with 100 μL 30 mg/mL luciferin ( XenoLight D-Luciferin—K+ Salt , Perkin Elmers ) prepared in endotoxin-free PBS ( Lonza ) and anesthetized using an O2 flow rate of 1 . 5 L/minute and 2% atmosphere of isoflurane for 10 minutes . Under the same regimen of anesthesia , sleeping mice were transferred into the in vivo imaging system ( IVIS Spectrum , Perkin Elmers ) on a dark surface and imaged ( binning 8 ) in automatic mode or for 5 minutes . Images were acquired for the ventral ( day 0 to day 5 post-infection ) or dorsal ( day 8 or 9; day 15; day 30 post-infection ) views . Quantitative data were obtained in square Region of Interest ( ROIs ) of the complete animal , excluding the tails for ventral images ( acute phase ) . For dorsal images , round ROIs were applied to the animals in the area corresponding to the kidneys . Quantitative data for the average light flux normalized by the time and area of ROIs were defined as light units/second/cm2/steradian ( p/s/cm2/sr ) . Images were analyzed using Living Image software ( Perkin Elmer ) , adjusting the photon count to 1x104 to 1x106 ( p/s/cm2/sr ) for all images shown . All protocols were undertaken in compliance with EU Directive 2010/63 EU and the French regulation on the protection of laboratory animals issued on February 1 , 2013 . They are part of project number # 2014–0049 , which was approved by the Institut Pasteur ethics committee for animal experimentation ( Comité d’Ethique en Expérimentation Animale CETEA registered under #89 ) and was authorized under #8562 by the French Ministry of Research , the French Competent Authority . Human blood was ordered under convention number C CPSL UNT-N° 12/EFS/134 ( to JQ ) from the Établissement Français du Sang , which collected informed consent from healthy volunteers . Cells were obtained from peritoneal lavage of C57BL/6 mice . Briefly , the skin was peeled off and clamped , followed by injection of 3 mL complete RPMI ( RPMI GlutaMax ( Gibco ) , 10% heat-decomplemented Foetal calf serum ( FCS ) ( Lonza ) , 1 mM HEPES ( Gibco ) , 1 mM non-essential amino acids ( Gibco ) , 1 mM sodium pyruvate ( Gibco ) ) supplemented with 1X antibiotic-antimycotic ( Anti-Anti Gibco ) into the peritoneal cavity . Next , a small incision was created in the preclamped skin and peritoneal content collected , passed through 100 μm strainer and centrifuged . Only cells from peritoneal lavage pellets lacking red blood cell contamination were used for experiments . The spleen was gently removed without damaging it and kept in complete RPMI medium . Spleens were macerated using 25G curved needles , and cells were subsequently passed through 100 μm and then 70 μm strainers . Following lysis of red blood cells using Red Blood Cells Lysis Buffer ( Sigma-Aldrich ) , the splenocytes were washed once with complete medium . Peritoneal macrophages and spleen cells were seeded at a final concentration of 1x106 cells/mL and 2x106 cells/mL , respectively , and 200 μL per well was distributed into a 96-well plate ( TPP ) . Stimulation of peritoneal or spleen cells was performed 2 to 3 h after plating . Since macrophages are adherent , peritoneal cells were washed once with PBS before stimulation to remove non-adherent cells . Cells were stimulated for 24 h after plating an equal number with 100 ng/mL E . coli LPS ( EB InvivoGen ) or with live L . interrogans at a multiplicity of infection ( MOI ) of 100 . Femurs from female and male mice were obtained after euthanasia cleaned , and the heads of the bones were removed . Two mL of medium was passed through the bones using a 21G needle to flush out the bone marrow cells . Bone marrow cells were centrifuged ( 300 g , 5 minutes ) and treated with Red Blood Cell Lysis Buffer ( Sigma-Aldrich ) for 10 minutes , followed by PBS washing . The cells were enumerated , and 5x106 cells were seeded in 100-cm2 cell culture dishes in 10 mL RPMI supplemented with 10% FCS , 1 mM sodium pyruvate , 1X Anti-Anti and 10% L929 cell supernatant . Cells were kept at 37 °C in the cell culture incubator ( 5% CO2 humidified environment ) for 7 days . At day 3 , 5 mL of the same medium was added . Seven days after seeding the cells , the medium was removed , and 3 mL of cell dissociation buffer ( Gibco ) was added . The cells were collected , centrifuged , enumerated and seeded in 96-well plates at a density of 2x105 cells per well and then stimulated with different leptospires at a MOI of 1:100 . At 24 h post-stimulation , nitric oxide and the cytokine dosage in the cell supernatant were assessed by the Griess reaction and ELISA , respectively . Differentiation of bone marrow cells into macrophages was assessed by flow cytometry on the day of collection using antibodies against the CD11b , F480 and CD11c markers . Peritoneal and bone marrow cells were seeded in 96-well plates at a density of 2x105 cells per well in 200 μL and infected with bioluminescent L . interrogans serovar Manilae . At 24 h post-stimulation , 100 μL of the cell supernatants were distributed into white flat-bottom 96-well plate to monitor the presence of live bacteria by bioluminescence measurements in a Centro LB960 microplate luminometer ( BERTHOLD ) , injecting 100 μL 1 . 5 mg/mL solution of luciferin ( XenoLight D-Luciferin—K+ Salt , Perkin Elmers ) in PBS . Human monocytes from blood were purified using the Pan Monocyte Isolation Kit ( Miltenyi Biotec ) according to the supplier’s instructions . After purification , the cells were suspended at 1x106 cells/mL in RPMI without serum , and 1x105 cells were seeded into a flat-bottom 96-well plate ( TPP ) . For training , the cells were incubated for 24 h in a 5% CO2 atmosphere at 37 °C in the presence of the indicated agonists . Next , the cells were washed twice with prewarmed PBS , and 200 μL fresh RPMI with 10% pooled human serum ( Zenbio ) was added . The cells were allowed to rest for 5 days , with half of the medium replaced with fresh RPMI with 10% human serum at day 3 . After resting , the cells were washed twice with prewarmed PBS and rechallenged as indicated in RPMI without serum for 24 h in a 5% CO2 atmosphere at 37 °C . Cell supernatants were frozen for subsequent cytokine analysis . Nitric oxide ( NO ) in the cell supernatant was quantified immediately after collection using the Griess reaction . Cells supernatants were kept at -20°C until the cytokine dosage , which was performed using ELISA Duo-set kits from R&D Systems for mouse IL-1β , RANTES , IL-6 , KC and human IL-6 , according to the instructions provided by the supplier . For surface staining , a total of 2x105 cells per experimental condition were washed once in a round bottom 96-well plate with MB buffer ( 1X PBS without Ca2+ and Mg2+; 0 . 5% v/v FCS , 2 mM EDTA ) and resuspended in 50 μL staining buffer containing 2 μg/mL of FcBlock ( anti CD16/CD32 ) and 1 μg/mL of antibody ( see Table 1 ) for 20 minutes on ice . Next , 50 μL 2 μg/mL e780 fixable viability dye ( eBioscience ) was added for 5 minutes , followed by 50 μL of 4% PFA for fixation for 5 minutes . For intracellular staining , after fixation the cells were washed and resuspended in Inside Perm kit reagent containing 2 μg/mL of the corresponding antibody . For intracellular staining of IFN-γ , 1X Brefeldin A solution ( eBioscience ) was added 3 h before the collection of the cells to block the intracellular release of IFN-γ . After staining , the cells were washed 2 times with MB buffer and resuspended in 200 μL of MB buffer for acquisition . Stained cells ( 100 μL ) were acquired ( at least 50000 events for phenotyping , 25000 cells for intracellular iNOS , and 550000 cells for IFN-γ ) on a MACSQuant analyzer ( Miltenyi Biotec ) cytometer after calibration and color compensation . Data were analyzed using FlowJo V10 .
First , to delineate the potential role of phagocytic cells during Leptospira infection , we injected mice with clodronate liposomes to deplete phagocytes in the peritoneal cavity 3 and 1 days before infection with leptospires ( Fig 1A ) . Phagocyte depletion was checked on the alleged day of infection in the peritoneal cavity and in blood by flow cytometry ( Fig 1B ) according to the gating strategy shown in ( S1A Fig ) . We observed specific depletion of macrophages in the peritoneal cavity . In blood , no changes were observed in circulating monocytes and neutrophils ( Fig 1B ) . Since leptospires can grow on fatty acids , we also checked that PBS and clodronate liposomes did not favor leptospiral growth in vitro; as expected , cultures of the bioluminescent L . interrogans MFLum1 strain were not affected by the presence of liposomes ( S2A Fig ) . Next , the effect of macrophage depletion on Leptospira infection was tested by challenging mice with a sublethal dose of bioluminescent L . interrogans ( 1x107 bacteria/mouse ) , which was injected into the peritoneal cavity . Using live imaging , we observed that depletion of macrophages had a marked effect on the initial control of leptospires ( Fig 1C , left panel ) . As previously shown , at day 1 post-infection , PBS-treated ( nondepleted ) infected mice had a reduced initial number of bacteria in the peritoneum . In sharp contrast , leptospiral loads in clodronate-depleted mice remained at their initial numbers until day 2 post-infection and then progressively declined . At day 5 post-infection , both groups completely controlled leptospires to the same extent . During the chronic phase , we observed a significant increase in the amount of kidney colonizing bacteria in the depleted group in comparison to mice from the PBS-treated group ( Fig 1C , right panel ) . Of note , renal colonization in depleted mice was proportional to the extent of systemic dissemination , as previously shown [4] . To achieve higher colonization in PBS-treated mice , we performed the same experiment but with an increased sublethal dose ( 5x107 bacteria/mouse ) of bioluminescent leptospires . Although before infection clodronate treatment provoked slightly more marked weight loss than PBS-liposomes , after infection clodronate-treated mice lost more weight than their PBS counterparts ( Fig 1D ) . One month post-infection , live imaging of the kidney revealed that PBS-treated mice had the expected levels of colonization and that depleted mice presented increased kidney colonization ( Fig 1D ) . These results indicate that peritoneal macrophages play a role in the initial control of L . interrogans administered IP in the mouse model . The protective role of macrophages is partial since no lethality was observed upon depletion . These data also suggest that training of macrophages could potentially have a positive impact on leptospirosis outcomes . We have recently shown that oral treatment with Lactobacillus plantarum can alleviate acute leptospirosis in mice through a myeloid cell-mediated effect [20] . Interestingly , Rice et al . , [35] have shown that the L . plantarum effect can be mimicked with CL429 , a synthetic bifunctional TLR2/NOD2 agonist corresponding to murabutide , a NOD2 agonist that is linked to Pam2cys , a TLR2 agonist [36] . Therefore , using TLR2 or NOD2-transfected HEK293T NF-κB reporter cell lines , we first checked the specificity of our strain of L . plantarum ( strain 256 ) . As expected , we found that both CL429 and L . plantarum 256 triggered the activation of TLR2 and NOD2 ( S2B Fig ) . We also checked that CL429 , which was reported not to affect mice [36] , also did not affect in vitro cultures of leptospires ( S2A Fig ) . Hence , we hypothesized that CL429 might , such as L . plantarum , exert a protective effect upon infection with L . interrogans . We treated C57/BL6 mice IP with CL429 at 2 weeks and 1 week before IP infection with a sublethal dose of the bioluminescent L . interrogans MFLum1 strain ( Fig 2A ) . Infected CL429-treated mice showed slight but significant reduced weight loss compared with PBS-infected controls , which could be a clinical sign of enhanced bacterial control ( Fig 2B ) . We repeated this experiment and tracked the infection in CL429-treated or control albino mice by live imaging of bioluminescent Leptospira ( Fig 2C ) . We observed a similar control of leptospires at day 1 post-infection between the two groups ( Fig 2C ) . However , from day 2 to day 4 post-infection , we observed a marked reduction of the bacterial burden in treated mice , corresponding to a lack of dissemination from the peritoneal cavity , whereas control mice had systemic dissemination of leptospires ( Fig 2C , lower left panel ) . We also evaluated kidney colonization during the chronic phase and observed a reduction in the kidney load at day 15 post-infection in CL429-treated mice , although at day 30 post-infection , the kidney loads were equivalent in both groups ( Fig 2C , right panels ) . In conclusion , CL429 treatment alleviated the acute phase of infection , although it did not result in kidney clearance . These findings further substantiate what we observed previously with Lactobacillus plantarum-treated C3H/HeJ mice [20] . To understand whether CL429 treatment , which helped mice to control Leptospira , could have boosted the macrophage response as expected for innate immune memory , mice were treated with CL429 , but instead of being infected , mice were sacrificed to collect peritoneal cells for ex vivo stimulation with pathogenic leptospires ( Fig 3A ) . Two weeks after the first CL429 injection , we checked the peritoneal composition by flow cytometry and observed an increased percentage of macrophages among total peritoneal cells upon CL429 treatment ( Fig 3B ) , as well as a reduction in T cells ( S1B Fig ) , although the proportions of other cell types were not affected . Since morphological changes have been associated with a trained cellular state [22] , after plating an equal number of peritoneal cells and washing to discard non-adherent cells , we examined non stimulated cells by bright field microscopy and observed significant morphological changes in cells from CL429 pretreated mice compared with the PBS controls ( Fig 3C ) . An enhanced pro-inflammatory response has been defined as one of the key aspects of trained immunity . Therefore , we checked the production of pro-inflammatory mediators at 24 h post-stimulation with LPS or live leptospires by ELISA . We observed a significant increase in the production of IL-6 , IL-1β , and RANTES after stimulation with 3 different pathogenic serovars of L . interrogans ( Manilae , Verdun and Fiocruz ) , but the amount of KC produced by these cells was not affected by the treatment ( Fig 3D ) . Of note , the lack of IL-1ß production after LPS stimulation was expected , since IL-1ß secretion is tightly regulated in mice and requires two signals that we have previously shown to be provided by live L . interrogans through TLR2/4 and the NLRP3 inflammasome [37] . Importantly , cytokine production upon stimulation with the 3 different pathogenic serovars of leptospires was equally enhanced . Finally , since murine macrophages produce nitric oxide ( NO ) , which has been described as a potent antileptospiral mediator as well as a key pro-inflammatory signature , we tested NO production in peritoneal cells 24 h post-stimulation . Interestingly , we found that peritoneal cells from CL429-treated mice showed enhanced NO production when rechallenged with leptospires or LPS , although strikingly , cells treated with PBS produced minimal amounts of NO upon L . interrogans stimulation ( Fig 3E ) . Finally , we tested the ability of peritoneal cells to kill bioluminescent leptospires at 24 h post infection . We found fewer live leptospires in peritoneal cells of CL429-treated mice compared with the level of bacteria in cells of PBS-treated mice ( Fig 3F ) . Altogether , these ex vivo data suggest that CL429 priming enhances the ability of macrophage to combat leptospires . Trained immunity has been shown to be independent of the adaptive immune response [26] . Therefore , to understand whether the CL429-mediated effect is independent of lymphoid cells , we used Rag2-γc mice deficient in B , T and NK cells . We treated these mice with CL429 as previously performed with WT mice ( Fig 4A ) and collected the cells for ex vivo functional and microscopy analysis . After CL429 treatment , we observed a change in morphology of plated peritoneal cells from Rag2-γc mice ( Fig 4B ) , as was observed with wild-type ( WT ) mice . We observed enhanced cytokine production of IL-6 and RANTES , but not KC , after stimulation with 3 leptospiral strains ( Fig 4C ) , reproducing the effect observed in immune-competent mice . However , we did not find any NO production in peritoneal cells from PBS- or CL429-treated Rag2-γc mice ( Fig 4D ) . These results obtained using mice deficient for B , T and NK cells indicate that our observed phenotype on enhanced cytokine production and morphology is driven by innate myeloid cells and is independent of the adaptive T and B cell compartment , which is consistent with trained immunity . Since the observed peritoneal effect was localized to the site of injection , we aimed to understand whether CL429 treatment could also have systemic consequences . First , we assessed if CL429 injection in vivo ( Fig 5A ) led to changes in blood or splenic cell populations and using flow cytometry , we observed no significant changes upon CL429 administration to mice ( S1C and S1D Fig ) . Next , the production of inflammatory mediators was assessed in splenocytes from PBS- or CL429-treated mice , and we observed by ELISA significantly enhanced production of interferon-γ ( IFN-γ ) upon Leptospira infection in comparison with the PBS-treated counterparts ( Fig 5B ) . Next , using flow cytometry , we examined which cell type produced IFN-γ in splenic cells . We observed significantly increased IFN-γ secreted by splenic NK cells upon stimulation with different pathogenic serovars , although no induction of IFN-γ was observed in T cells under our conditions ( Fig 5C ) . Additionally , we tested whether progenitors from the bone marrow could also have been primed by CL429 treatment . At 2 weeks post-CL429 or -PBS treatment , we collected bone marrow cells and differentiated them into bone marrow macrophages ( BMMs ) ( Fig 5A ) . First , we assessed whether the differentiation process of bone marrow cells led to macrophages . We observed that after a week of differentiation , more than 95% of the cells were macrophages ( CD11b+/F4/80+ ) ( Fig 5D ) . In addition , we did not observe any major morphological differences in BMMs by microscopy when comparing BMMs obtained from PBS and CL429-treated mice . We then studied the secretion of pro-inflammatory compounds by BMMs . Upon restimulation with different leptospiral serovars , we observed increased production of IL-6 and KC but the same production level of the chemokine RANTES ( Fig 5E ) . Finally , we observed increased production of antimicrobial NO upon stimulation with different pathogenic serovars ( Fig 5F ) . Since NO has been associated with leptospiral killing , we assessed bacterial survival using a bioluminescent Manilae strain . We observed enhanced bactericidal activity in BMMs from CL429-treated mice compared with PBS-treated mice ( Fig 5G ) , as we observed in peritoneal cells . Altogether , these results indicate that treatment with CL429 , a TLR2 and NOD2 agonist , leads to increased production of cytokines and antibacterial mediators ex vivo , in the peritoneal cavity , the bone marrow and the spleen , which are remote from the site of infection . This local and systemic effect is consistent with trained immunity . Therefore , we wondered whether the BMMs originating from the CL429-treated mice could express more PRRs , as has been previously described for BCG inducing trained immunity and TLR4 upregulation [38] . We measured the expression of NOD1 , NOD2 and TLR4 in BMMs by qRT-PCR but did not detect an upregulation of those receptors in non stimulated BMMs derived from trained mice compared with PBS-treated mice ( S3 Fig ) . Upon restimulation with leptospires ( Manilae serovar ) , TLR2 mRNA expression was clearly upregulated compared with non-stimulated cells , although TLR4 and NOD2 were not or were only modestly upregulated , respectively . However , no difference was observed between stimulated BMMs harvested from CL429- or PBS-treated mice . To assess whether the effects of CL429 could be sustained over time , we repeated the experiments ( with two CL429 injections , Fig 6A ) and collected peritoneal , splenic and bone marrow cells at 8 weeks or 3 months post-treatment rather than 2 weeks after the first injection of CL429 . By flow cytometry , we observed an increased proportion of macrophages in the peritoneum of CL429-treated compared with PBS control mice ( Fig 6B ) , although we did not detect other changes in the peritoneal cavity ( S4A Fig ) . Interestingly , this increase at 8 weeks was similar to the one observed at 2 weeks after the first injection of CL429 . Moreover , we also observed the characteristic morphological changes in peritoneal cells from CL429-treated mice that we observed at 2 weeks post-treatment ( Fig 6C ) . Next , LPS and 3 strains of live L . interrogans were used to stimulate the cells for 24 h . Enhanced production of pro-inflammatory mediators ( NO and IL-1ß ) was observed ( Fig 6D and 6E ) , and the cellular response to leptospiral challenge was similar in different strains and comparable to the response observed at 2 weeks after the first CL429 injection . However , if the KC response was unchanged , IL-6 cytokine was not as enhanced as previously observed 2 weeks after treatment ( Fig 6E ) . To determine whether the enhanced cytokine and NO production observed after ex vivo stimulation were due to an increased number of macrophages or if CL429 treatment enhanced the reactivity of those cells , we detached the adherent cells at 24 h post-infection and analyzed their phenotype by flow cytometry as well as the expression of iNOS , the enzyme responsible for NO production . Consistent with our results using peritoneal lavages , we observed an enhanced number of macrophages ( CD11+/F4/80+ ) . Interestingly , if the number of macrophages among the plated peritoneal cells was higher in the CL429-treated mice ( S4B Fig ) , the expression of iNOS was strikingly enhanced in the CL429-treated macrophages infected with LPS or leptospires , compared to PBS-treated infected cells , whereas treated noninfected cells showed the same basal expression ( S4C Fig ) . This result suggests that CL429 treatment not only enhances the number of macrophages in the peritoneal cavity but also induces innate immune memory of macrophages . The enhanced expression of iNOS in macrophages was comparable upon stimulation with the 3 strains of leptospires ( S4D Fig ) . Consistent with the results obtained at 2 weeks post-treatment , although no change in number was observed ( S5A and S5B Fig ) , splenic NK cells from mice treated 3 months prior with CL429 exhibited more intracellular IFNγ than their PBS counterparts ( S5C and S5D Fig ) . In addition , BMMs from mice 3 months post-treatment presented increased levels of IL-6 and KC and NO , although the levels of RANTES were not different ( S6 Fig ) . Altogether , these results indicate that injection of CL429 leads to profound modifications to splenic NK and macrophages in the environment of the peritoneal cavity as well as in their progenitors in the bone marrow and that the increased response to subsequent challenges with leptospires or E . coli LPS persists for at least 3 months . Trained immunity induced by PRR ligands can be recapitulated in humans primary monocytes ex vivo [22] . To understand whether CL429 could induce innate immune memory in human monocytes and increase their responsiveness towards leptospires , human primary monocytes isolated from healthy donors were individually seeded in plates and stimulated for 24 h with PBS or treated with CL429 or MDP , the NOD2 agonist used as a positive control of trained immunity [22 , 38] . After this first stimulation or “training” , the cells were allowed to rest for 5 days and then were restimulated for 24 h with LPS or live pathogenic L . interrogans strains ( Fig 7A ) . IL-6 measured by ELISA in the supernatants was used as a read-out to determine their inflammatory status ( Fig 7B ) . Several doses of MDP and CL429 were tested . A concentration of 0 . 1 μM was chosen since it did not affect cell viability after the second stimulation ( S7 Fig ) . Even though the IL-6 response within donors was very variable ( Fig 7B ) , we observed training for both MDP and CL429 in 6/6 independent donors after LPS and Fiocruz restimulation and in 5/6 donors after Manilae stimulation . Strikingly , in all donors tested except donor 1 , CL429 treatment enhanced IL-6 production to a greater extent than the positive control MDP . These results show that CL429 can boost human monocytes and enhance their responsiveness to Leptospira infection .
In this study , we explored the role of macrophages in a mouse model of leptospirosis . On the one hand , we used a depletion approach with a treatment with clodronate liposomes that suggests a role for macrophages in the early stages of leptospiral infection . On the other hand , we showed that pretreatment with CL429 , a bifunctional TLR2/NOD2 ligand , induces protection against leptospiral infection . Pretreatment with CL429 modified the environment of the peritoneal cavity with macrophage enrichment and boosted the responses of macrophages against a subsequent leptospiral infection . These effects in mice correspond to the described trained immunity characteristics and were recapitulated in human monocytes . These results suggest that CL429 pretreatment could be used to boost human and animal cellular responses , which could potentially open up prophylactic strategies to control acute L . interrogans infection . Our clodronate macrophage depletion findings indicate that the lack of these cells influenced both the acute and chronic phases of leptospirosis , although the depletion treatment had only been done before infection . First , we consistently observed an initial decrease of bacteria at day 1 post-infection in WT mice [4] , which was not present upon IP clodronate depletion treatment . Therefore , we attribute this initial decline in bacterial burden to a protective effect of peritoneal macrophages , which killed some leptospires , as we showed in vitro , but not in sufficient numbers to avoid systemic infection . Although IP depletion of macrophages had no lethal effect , it led to enhanced bacteremia during the acute phase and to increased bacterial colonization of the kidneys . This result is consistent with the findings of Isogai et al . , using silica intravenous ( IV ) depletion , in which bacteria were more prominent in IV depleted mice during the acute phase , without a lethal outcome [39] . Since we did not observe monocyte depletion in the blood , it is unlikely that the IP clodronate treatment could have also depleted the macrophages in the renal tissue . This result suggests , as was previously observed [4] , that early control of bacteria predicts the future extent of renal colonization . It is also in line with the recent study by Ferrer et al . , [40] , based on a clodronate IP depletion regimen that was maintained after infection with L . interrogans serovar Copenhageni strain Fiocruz . They also showed that macrophages controlled the leptospiral burden in mice . Since macrophage depletion increased the renal bacterial load , we may also conclude , in accordance with Ferrer’s study , that L . interrogans serovar Manilae was not transported as cargo by macrophages to the mouse kidney , as has been previously suggested in zebrafish using the Fiocruz strain [41] . Upon injection of CL429 , the peritoneal cavity underwent a modification in its immune cells composition . Even though we did not observe any major changes in the cell populations in other compartments , there was an increased number of macrophages in the peritoneal cavity . This increase was specific since it did not affect the resident population of other cell types such as B or NK cells , although compensation was provided by a slight decrease in the percentage of T lymphocytes . This change in macrophages in the peritoneal cavity and the enhanced inflammatory response of CL429-treated peritoneal cells were sustained at 3 months post-treatment , indicating that the treatment had a long-term effect . Using CL429-treated WT or Rag-2γc mice ( lacking T , B and NK cells ) , we obtained similar enhanced cytokine responses of peritoneal cells after ex vivo leptospiral stimulation , showing a major innate immune contribution of macrophages to this phenotype . Altogether , these results indicate that the injection of CL429 led to a new homeostasis in the peritoneal cavity with long-lasting functional changes in macrophages to render them more responsive to leptospires and other stimuli . Interestingly , we demonstrated that the effect of CL429 treatment was not only local but also had systemic effects on splenic NK cells and on bone marrow myeloid progenitors , which maintained their “training” after ex vivo differentiation , as recently described for BCG and ß-glucan [42 , 43] . Future questions will address whether the trained peritoneal macrophages are resident cells and whether CL429 treatment increases their numbers or if they emanate from the bone marrow and are then recruited from the blood to be established in the peritoneum . For instance , infection with murine herpesvirus 4 leads to the replacement of lung macrophages by inflammatory monocytes , thereby conferring protection against house dust mite-induced experimental asthma [44] . Intraperitoneal administration of CL429 alleviated acute leptospiral infection in mice . The bacterial burden in mice was reduced due to the enhanced killing ability of macrophages . This treatment delayed but did not reduce kidney colonization at one month post-infection . These results resemble the observation made for C3H/HeJ mice after Lactobacillus plantarum oral treatment [20] . In our model , only two IP doses , compared with 30 oral doses of L . plantarum in [20] , were sufficient to alleviate leptospirosis in a similar fashion . Here we showed that the protective effect of CL429 was due to enhancement of the macrophage number and functions mimicking trained immunity or an innate immune memory phenotype . Interestingly , recruitment of macrophage-like cells has also been observed in the kidneys of Lactobacillus plantarum-treated mice , at 15 days post-infection with leptospires [20] . By extension , we may infer that the positive effect on acute leptospirosis previously observed with oral treatment with Lactobacillus plantarum , which is agonist of both TLR2 and NOD2 , was probably due to a trained immunity effect . Regular consumption of probiotics such as Lactobacillus plantarum , a generally recognized as safe ( GRAS ) organism , could be envisioned as a prophylactic treatment to avoid or alleviate acute leptospirosis and other infectious diseases . In our study , we did not dissect the mechanisms by which the macrophages exerted their therapeutic effects . However , we observed ex vivo that infected macrophages previously trained with CL429 had strikingly high NO production that resulted in killing of leptospires . These results strongly suggest that the enhanced NO response of peritoneal cells may participate in the reduction of leptospiral loads observed in vivo in CL429-treated mice . Whether enhanced levels of antimicrobial peptides and proteins involved in reactive oxygen species production are also triggered remains to be studied . Although upregulation of PRRs has been shown upon training with BCG [38] , we did not observe an upregulation of NOD1 , NOD2 or TLR4 mRNA in BMMs . However , we used BMM since we were limited by the number of peritoneal cells and this result can be considered with caution . Indeed , upon training with CL429 , peritoneal and bone marrow-derived macrophages did not behave exactly the same . Although secretion of IL6 was increased in peritoneal and bone marrow derived macrophages , the chemokines KC and RANTES showed opposite trends , with KC secretion increased in trained BMMs and not changed in peritoneal cells , whereas the opposite was true for RANTES . Since trained immunity has been mostly studied in human monocytes/macrophages that do not produce NO , NO has not been yet routinely associated with TI . Here , we showed that NO production , known as a potent antileptospiral compound [18 , 19] , was an excellent read-out of TI , which may also explain how mice might control the infection . Interestingly , we observed that CL429 injection led to a systemic effect , as observed for the responsiveness of splenic NKs from treated mice , which produced more IFN-γ . NO production was enhanced in presence of IFN-γ , a cytokine produced by T and NK cells . We did not observe NO production in cells from Rag2-γc mice ( lacking B , T and NK cells ) , and we showed in WT mice that NK cells , not T-cells , were the main producers of IFN-γ . Augmented cytokine production by trained macrophages could stimulate NK cells to secrete more IFN-γ , or NK cells themselves could be “trained” , as previously reported [28] . Interestingly , a protective role of IFN-γ against renal colonization with pathogenic leptospires has recently been suggested [45] . Considering that our present study shows reduced kidney load at day 15 post-infection , it would be interesting , as was proposed by Zuerner in 2011 [46] , to further prime the NK cells to limit renal colonization , which we still observed after CL429 treatment at one month post-infection . Whether NK could be trained using TLR and NOD agonists other than CL429 to limit the renal colonization after infection with pathogenic leptospires will be the focus of future studies . Consistent with a trained immunity effect that provides non-specific protection , peritoneal and bone marrow-derived macrophages from CL429-treated mice responded better than those from nontreated mice to ex vivo stimulation with LPS from E . coli and with different serovars of pathogenic leptospires isolated worldwide ( Icterohaemorragiae strain Verdun , a French isolate , Copenhageni strain Fiocruz , a Brazilian strain and Manilae strain L495 , isolated in Philippines ) . Of note , the 3 different strains stimulated trained macrophages and human monocytes to the same extent . Efficient vaccines against leptospirosis are notoriously difficult to establish because of the numerous leptospiral serovars identified on the basis of the highly immunogenic but variable O-antigen portions of the lipopolysaccharide . Therefore , only serovar-specific vaccines are available . Interestingly , our study suggests that CL429 may be used as a prophylactic treatment to alleviate the severity of leptospirosis , independently of the Leptospira serovar . In vitro studies have shown that NOD2 is associated with innate immune memory effects [22 , 38] . Conversely , TLR2 has been reported to induce tolerance [22] , the opposite functional effect of innate immune memory , in which the immune response to second encounters is significantly reduced . Our study showed that CL429 , a bi-functional TLR2/NOD2 agonist , triggered innate immune memory in a mouse model but also , in a certain low concentration range , boosted human monocyte responses . Their functional response led to the production of more IL-6 upon stimulation with several leptospiral pathogenic strains . Interestingly , at comparable molar concentrations , the CL429 effect was slightly higher than MDP , which was used as a positive control for training [38] . This synergistic effect of CL429 has been previously highlighted by Pavot et al . , [36] who first tested the effect of CL429 on the activation of dendritic cells , and by Rice et al . , [35] , who showed in mice that CL429 was better than NOD2 or TLR2 agonists alone to alleviate a secondary pulmonary viral infection . Here we further demonstrated that the immunomodulatory role of CL429 indeed corresponded to trained immunity , as previously suggested [35] . Of note , Borrelia burgdorferi , another spirochete , has been used as a non specific secondary stimulant of ß-glucan-trained human monocytes [47] , which suggests that trained immunity could also be used to limit Lyme disease . We have recently shown that pathogenic Leptospira escape the NOD1 and NOD2 response [12] . NOD2 has been involved in adaptive immunity and shown to instruct the onset of antigen-specific T and B cell immunity in vivo in mice . Moreover , NOD2 stimulation shifts the immune response towards a Th2-type profile characterized by the induction of IL-4 and IL-5 by T cells and by IgG1 antibody responses [48] . Th2 confers immunity to extracellular pathogens . Even though L . interrogans have been described intracellularly in macrophages [15 , 17 , 41] , L . interrogans are extracellular pathogens that replicate in blood or in the lumen of renal proximal tubules . Whether leptospiral escape from NOD1/2 signaling could consequently limit efficiency of macrophages to phagocytose leptospires and , thereby , alter the onset of adaptive immunity is under investigation . However , in this study , we showed that CL429 , a NOD2 agonist , activated macrophage function . Vaccine grade CL429 is available and could be used as an adjuvant [36] combined with conserved antigens of pathogenic leptospires , potentially establishing novel vaccines against leptospirosis . Likewise for BCG , which is at the origin of the discovery of trained immunity and has been used as a platform for many vaccines , we may hypothesize that CL429 could increase the efficiency of leptospiral vaccines . Thus , a recombinant BCG-based leptospiral vaccine expressing LipL32 ( the most abundant leptospiral lipoprotein ) rescued half of the treated hamsters from lethal challenge ( 2x LD50 ) with L . interrogans 70 days post-immunization compared to the rescue of one-fifth with the control BCG [49] . A limitation of the present study is the route of CL429 administration and leptospire infection . We only studied the training effect of CL429 after intraperitoneal injection and leptospire challenge using the same route , neither of which are physiologic , nor are they adapted to vaccine trials . Other mouse models of mucosal [50] or subcutaneous infection of leptospires should be used to test the effect of CL429 . In conclusion , we showed that CL429 treatment in mice and in human cells triggered innate immune memory that helped the host to fight acute leptospiral infection , independently of the serovar of L . interrogans . Our study demonstrates for the first time that manipulation of macrophages and other innate immune cells through the induction of innate immune memory could have a positive effect in vivo in mice to alleviate acute leptospirosis . Furthermore , it enhanced ex vivo the inflammatory response of human monocytes towards leptospires . This result led us to hypothesize that CL429 might be used as a prophylactic treatment to alleviate human and animal leptospirosis . Thus , this work paves the way to the design of new prophylactic strategies against leptospirosis . Although a universal vaccine against leptospires is still a long-term goal [3] , we envision the use TLR and/or NLR combined agonists , which are known to train immunity , as adjuvants for future leptospiral vaccines .
|
Leptospirosis is a worldwide zoonotic neglected disease caused by pathogenic Leptospira . In addition to cattle and pets , it affects one million people per year and causes 60000 deaths . Antibiotics are efficient when given early after infection , but because the symptoms are not specific , leptospirosis is difficult to diagnose in a timely fashion . There are ~300 serovars of pathogenic Leptospira , which complicates the development of a universal vaccine against leptospirosis . Therefore , prophylactic treatments would be welcome . Recently , a new kind of immune memory known as “innate immune memory” or “trained immunity” has been described . It is not linked to adaptive immunity but only to innate immune cells , such as macrophages . Initial stimulation by a bacterial or fungal component can “train” macrophages to respond better but nonspecifically to a second encounter with a microbe . In this work , we show that trained immunity can be induced in mice to alleviate the symptoms of experimental acute leptospirosis . Cells from treated mice stimulated ex vivo by different serovars of pathogenic leptospires secreted more antimicrobial compounds and showed enhanced leptospire killing . Moreover , this treatment was able to boost human monocytes . This work presents a host-directed therapeutic approach that could be used as a novel prophylactic strategy against leptospirosis .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"flow",
"cytometry",
"medicine",
"and",
"health",
"sciences",
"neurochemistry",
"immune",
"cells",
"leptospira",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"tropical",
"diseases",
"microbiology",
"neuroscience",
"animal",
"models",
"bacterial",
"diseases",
"model",
"organisms",
"experimental",
"organism",
"systems",
"neglected",
"tropical",
"diseases",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"spectrum",
"analysis",
"techniques",
"white",
"blood",
"cells",
"zoonoses",
"animal",
"cells",
"animal",
"studies",
"medical",
"microbiology",
"neurochemicals",
"nitric",
"oxide",
"microbial",
"pathogens",
"mouse",
"models",
"leptospirosis",
"spectrophotometry",
"biochemistry",
"cytophotometry",
"cell",
"biology",
"monocytes",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"leptospira",
"interrogans",
"macrophages",
"organisms"
] |
2019
|
Innate immune memory through TLR2 and NOD2 contributes to the control of Leptospira interrogans infection
|
In most meiotic systems , recombination is essential to form connections between homologs that ensure their accurate segregation from one another . Meiotic recombination is initiated by DNA double-strand breaks that are repaired using the homologous chromosome as a template . Studies of recombination in budding yeast have led to a model in which most early repair intermediates are disassembled to produce noncrossovers . Selected repair events are stabilized so they can proceed to form double-Holliday junction ( dHJ ) intermediates , which are subsequently resolved into crossovers . This model is supported in yeast by physical isolation of recombination intermediates , but the extent to which it pertains to animals is unknown . We sought to test this model in Drosophila melanogaster by analyzing patterns of heteroduplex DNA ( hDNA ) in recombination products . Previous attempts to do this have relied on knocking out the canonical mismatch repair ( MMR ) pathway , but in both yeast and Drosophila the resulting recombination products are complex and difficult to interpret . We show that , in Drosophila , this complexity results from a secondary , short-patch MMR pathway that requires nucleotide excision repair . Knocking out both canonical and short-patch MMR reveals hDNA patterns that reveal that many noncrossovers arise after both ends of the break have engaged with the homolog . Patterns of hDNA in crossovers could be explained by biased resolution of a dHJ; however , considering the noncrossover and crossover results together suggests a model in which a two-end engagement intermediate with unligated HJs can be disassembled by a helicase to a produce noncrossover or nicked by a nuclease to produce a crossover . While some aspects of this model are similar to the model from budding yeast , production of both noncrossovers and crossovers from a single , late intermediate is a fundamental difference that has important implications for crossover control .
Meiotic recombination is initiated by a DSB on one chromatid followed by repair using the homologous chromosome as a template , resulting in crossover ( CO ) or noncrossover ( NCO ) products [1] . In the predominant model of repair , NCOs are produced when an early intermediate – a D-loop extended by synthesis using a homologous template – is disassembled by a helicase ( Figure 1C ) , whereas COs are produced when a late intermediate – the double-Holliday junction ( dHJ ) – is cleaved by a resolvase ( Figure 1F ) . Crossover control , the ill-defined mechanisms that determine the number and distribution of crossovers , is thought to act prior to the bifurcation of CO and NCO pathways [2] . This model has been derived largely from studies in Saccharomyces cerevisiae , with strong support coming from the physical isolation of molecules with the properties expected of the key intermediates [3] , [4] . Because many key meiotic recombination proteins are conserved , it is thought that this model is also applicable to plants and animals; however , it has not been possible to isolate recombination intermediates in model metazoans to test this assumption . Here , we take a molecular genetic approach to analyzing recombination intermediates to determine what structures give rise to COs and NCOs in a model metazoan , Drosophila melanogaster . Recombination involves formation of heteroduplex DNA ( hDNA ) , regions in which the two strands of a duplex come from different parental DNA molecules ( Figure 1 ) . Sequence differences between the parental chromosomes result in base-base mismatches and insertion/deletion ( indel ) loops in hDNA and can be used as markers to map hDNA tracts . Different recombination models predict different arrangements of hDNA ( e . g . , Figure 1C vs 1H ) . In the budding yeast model , NCOs arise from synthesis-dependent strand annealing ( SDSA ) , with limited , if any , contribution from dHJ resolution or dissolution . SDSA predicts a cis configuration of hDNA , with all of the markers from the donor on one strand of the product ( Figure 1C ) . In contrast , dHJ dissolution predicts trans hDNA , with markers on different strands on opposite sides of the DSB ( Figure 1H ) . Crossovers are thought to come from resolution of dHJs by cleavage , as in the original double-strand break repair ( DSBR ) model of Szostak et al . [5] . In this model , dHJs can be resolved in either of two equally likely orientations ( Figure 1F ) . One orientation gives products with a single hDNA tract ( upper products ) and the other gives products with a tract of hDNA adjacent to a tract of gene conversion ( lower products ) . Thus , analysis of hDNA patterns in final recombination products can be used to make inferences about the structures of intermediates that give rise to COs and NCOs . The information in hDNA is usually lost because of mismatch repair ( MMR ) , resulting in either gene conversion or restoration of the original sequence ( Figure 2A ) . Attempts to recover meiotic hDNA by knocking out the canonical MMR have been made in budding yeast , animals , and plants [6]–[10] . In every case , the hDNA tracts that are recovered are complex mixtures of hDNA , gene conversion , and apparent restoration ( Figure 2B; we note that the term “half conversion” has been used in genetic studies to refer to retention of hDNA in the final recombination products , but we use “hDNA” to refer to regions of heteroduplex both in intermediates and in products of recombination ) . This complexity makes interpretations difficult because it is not possible to determine whether tracts of conversion come from synthesis-dependent processes that do not involve hDNA , such as gap repair or synthesis and dHJ resolution , or from hDNA that was repaired by a process other than the canonical MMR pathway . Similarly , apparent restoration could come from either hDNA repair or from synthesis using the sister chromatid as a template , with transitions from hDNA to restoration to conversion possibly resulting from template switching during repair . In the canonical eukaryotic MMR pathway , recognition of mismatches and indels is dependent on heterodimers of MutS homolog ( Msh ) proteins , Msh2–Msh3 and Msh2–Msh6 [reviewed in 11] . Drosophila does not have an ortholog of Msh3 [12]; it is thought that all canonical MMR uses a heterodimer between the Msh2 ortholog ( SPEL1 ) and MSH6 . In support of this hypothesis , meiotic recombination in Msh6 mutants resulted in hDNA tracts that were patchy , as described above ( Figure 2B ) , suggesting that canonical MMR was eliminated [8] . It was proposed that the patchiness resulted from a short-patch MMR system that was able to repair some mismatches and small indels within the same meiotic hDNA tract independently of each other ( Figure 2B ) . Short-patch MMR has been reported in fungi , animals , and plants , but in most cases the proteins that execute this pathway are unknown [6] , [13]–[16] . The exception is S . pombe , where a short-patch MMR system that depends on nucleotide excision repair ( NER ) operates during meiosis . This short-patch system is detected when canonical MMR is absent , and seems to repair primarily C∶C mismatches , which frequently escape canonical MMR [13] . In budding yeast , NER has recently been shown to repair mismatches containing methylated bases [17] , but this pathway is not thought to be involved in repair of non-methylated mismatches [14] . In Drosophila mei-9 mutants , a subset of meiotic hDNA tracts are able to escape both canonical and short-patch MMR [18] . MEI-9 is the Drosophila ortholog of S . cerevisiae Rad1 and mammalian XPF , the catalytic subunit of a nuclease essential for NER [19]–[22] . This suggests that NER might be involved in short-patch MMR in Drosophila; however , these studies were complicated by the fact that MEI-9 is also required to generate meiotic crossovers [18] , [22] , [23] . We now show that hDNA repair in MMR mutants in the model metazoan Drosophila melanogaster requires the NER protein XPC . XPC , the ortholog of S . cerevisiae Rad4 , is involved in the DNA damage recognition step of NER [24] and has no known or suspected role in meiotic recombination . The ability to knock out both canonical and short-patch MMR allowed us to analyze uncorrected hDNA patterns , leading to novel insights into the structures of pre-CO and pre-NCO intermediates . Our findings challenge the applicability of a central paradigm of the current recombination model from budding yeast by suggesting that NCOs and COs may arise from the same intermediate .
To recover hDNA tracts , we used a genetic assay to select for wild-type recombinants in the rosy ( ry ) gene [8] , [18] , [25] , [26] . Briefly , when flies mutant in ry are exposed to dietary purine , they die as larvae . We generated females that were heteroallelic for two ry mutations about 4 kb apart ( Figure 3A ) . Each ry allele was flanked by unique recessive markers that allowed us to determine if a recombinant was a CO or a NCO and had additional markers ( single nucleotide polymorphisms ( SNPs ) and indels ) that allowed us to map the hDNA tracts . These females were mated to males with a deletion in ry and allowed to lay embryos for three days . Purine was then added to the food; only wild-type recombinant larvae survived to adulthood . The presence of hDNA in the maternal ry allele results in mosaic larvae that have both strands of this chromosome represented in different cells or tissues . If the hDNA spans a mutant site , this results in mosaicism for ry activity , but ry is non-cell autonomous so these larvae also survive purine treatment [8] . To detect mosaicism and analyze the composition and structure of hDNA tracts , we extracted genomic DNA from the surviving recombinants and sequenced both bulk PCR product and cloned , individual molecules . To test the hypothesis that short-patch MMR in Drosophila is dependent on NER , we first asked whether tract lengths are consistent with NER tracts , which extend 22–24 nucleotides 5′ and 5–6 nucleotides 3′ of the lesion being excised [27] . We analyzed previously described recombination tracts from Msh6 mutants , which lack canonical MMR but exhibit short-patch MMR [8] . We classified each pair of adjacent markers as co-repaired ( both converted or both restored ) or not co-repaired ( one converted and one restored or one repaired and one not repaired ) ; pairs in which both were unrepaired were not counted . 40 of 42 ( 95% ) pairs of markers less than 21 bp apart , and therefore within the range of NER tracts , were classified as co-repaired ( Figure 3B ) . In contrast , when adjacent polymorphisms were further apart than the size of NER tracts , only 40 of 111 ( 36% ) were considered co-repaired ( P<0 . 0001; Figure 3B ) . This result supports the hypothesis that short-patch MMR in Drosophila is mediated by NER . We directly tested the involvement of NER in short-patch MMR by removing XPC , a key damage recognition factor in NER [24] . Previous studies of Xpc ( also known as mus210 ) did not report any apparent meiotic defects [28] . We screened 1 . 7 million larvae and recovered 66 products of meiotic recombination ( 50 crossovers and 16 noncrossovers ) between two highly polymorphic alleles of ry in Xpc; Msh6 double mutants . Among these recombinants we detected 32 hDNA tracts spanning a total of 136 markers ( Figures 4 and 5 ) . This does not include two noncrossovers that had tracts of full gene conversion with no hDNA ( Figure S1 ) ; these likely came from residual canonical MMR due to maternal MSH6 or from an alternative recombination pathway such as double-strand gap repair , so they were excluded from further analysis . Only two of the 136 markers ( 1 . 5% ) in these tracts were repaired , both as restorations within the same noncrossover . This was the only tract that was patchy , as it also had sites with unrepaired hDNA ( Figure S1 ) . In stark contrast , Msh6 single mutants repaired 274 of 334 hDNA markers ( 82%; P<0 . 0001 ) and 35 of 39 of hDNA tracts were patchy ( 90%; P<0 . 0001 ) [8] . Based on these data and previous work suggesting that the NER protein MEI-9 is involved in short-patch MMR [18] , we conclude that short-patch MMR in Drosophila is indeed dependent on NER . This is the first identification of a pathway responsible for short-patch MMR in a metazoan . It is notable that , unlike in S . pombe , where NER-dependent short-patch MMR repairs primarily C-C mismatches [13] , short-patch MMR in Drosophila appears to repair all types of mismatches and short insertion/deletion polymorphisms with similar efficiency ( Figure 3C ) . Eliminating both canonical and short-patch mismatch repair makes it possible , for the first time in a metazoan , to analyze the structures of meiotic hDNA tracts generated in the complete absence of mismatch repair , thereby providing unique insights into recombination pathways . We recovered thirteen NCOs that spanned more than one marker ( Figure 4 ) . Twelve of the thirteen NCOs occurred at the ry531 locus . This is potentially due to a difference in the ability to detect NCOs at each mutation: the nearest SNP on either side of ry531 is between 150–200 bp and the nearest SNP downstream of ry606 is 400 bp . Additionally , the markers upstream of ry606 consist of some small insertion deletion polymorphisms , while the markers around ry531 are single nucleotide polymorphisms . Previous analyses at rosy in the Msh6 mutant did not show the same bias in NCO location [8] , [18] , [26] , suggesting that mutations in XPC may influence our ability to recover NCOs that span indels ( Figure 3 ) . Surprisingly , of the thirteen NCOs with tracts that include more than one marker , only six have the cis hDNA arrangement predicted by SDSA; the other seven have two adjacent tracts of hDNA in the trans orientation ( Figure 4 , asterisks ) , an arrangement not predicted by the standard SDSA model . NCOs with trans hDNA were previously seen in Msh6 mutants [8] , [18] . It is possible that mutations in mismatch repair genes directly cause an increase in the frequency of the intermediate that gives rise to trans hDNA , possibly through mechanisms such as decreasing the frequency of heteroduplex rejection . However , the level of heterology we used in these experiments does not affect the frequency of meiotic recombination in wild-type females [29] , suggesting that heteroduplex rejection is not frequent in this context . Therefore , we focus the discussion below on other sources of trans hDNA . A small number of NCOs with trans hDNA were also reported in mei-9 mutants [8] , [18] . Radford et al . [18] hypothesized that these NCOs arose from dHJ dissolution because the MEI-9 meiotic resolvase was not available to cleave the dHJs . Since canonical MMR appears to be normal in mei-9 mutants [18] , it was suggested that hDNA persisted in these NCOs because dissolution does not leave nicks that are known to stimulate MMR [11] and because short-patch MMR is defective due to the loss of the NER function of MEI-9 . According to this model , if NCOs are normally produced by dHJ dissolution , then unrepaired hDNA should be frequent in NCOs from wild-type females; however , unrepaired hDNA is never detected in recombinants from wild-type females [8] , [26] , [30] . This argument implies that the trans hDNA in the NCOs we describe here arises through a process that generates products with nicks or gaps rather than through dHJ dissolution . Based on these considerations , we propose that the trans hDNA in Xpc; Msh6 mutants comes either from either two-ended SDSA or a process we term “two-end engagement” , wherein both ends of a break engage with the same homologous chromatid and are extended by synthesis but are not ligated to produce a dHJ ( see Figure 6 and discussion below ) . Studies of gap repair in mitotically growing yeast cells have led to the suggestion that some trans hDNA in NCOs comes from an intermediate with unligated HJs [31] . We also analyzed crossovers generated in the absence of both canonical and short-patch MMR . In the DSBR model [5] , crossovers are generated by resolution of a dHJ in either of two equally likely orientations , one of which gives products with a tract of hDNA adjacent to a tract of full conversion ( Figure 1F , upper products versus lower products ) . Because we recover only the chromatid that goes into the oocyte , this tract of gene conversion can only be detected if there is an adjacent tract of hDNA . As drawn in Figure 1F , the model predicts that all COs have hDNA tracts , but we detected hDNA in only 16 of the 50 COs ( 32% ) . This may be a consequence of low marker density in some regions ( Figure 5 ) , since tracts that do not span a marker will not be detectable . If our ability to detect gene conversion tracts was similar to our ability to detect hDNA , then among the 16 COs with hDNA it should have been possible to detect gene conversion in five COs ( 32% of 16 ) . The binomial distribution probability of recovering zero out of five is 0 . 04 . This suggests that crossovers in Drosophila are not usually associated with MMR-independent gene conversion tracts . One possible explanation for these results is that dHJ resolution is biased toward a single orientation in which nicks are made at or near the point where the 3′ end of the nascent DNA is ligated to the original resected strand ( Figure 1F , open arrowheads ) . In yeast , a similar bias has been noted by Gilbertson and Stahl [32] and later by Jessop et al . [33] . It has been proposed for both S . cerevisiae meiotic recombination and DSB repair in mammalian cell lines that newly synthesized DNA provides structural asymmetry that directs cleavage to achieve this bias [34] , [35] . An alternative explanation is that dHJs are un-ligated; nicking across from un-ligated HJs would also produce crossovers with an hDNA tract but no gene conversion ( Figure 6 ) . Models in which the dHJs are not ligated have been proposed to better fit the in vitro biochemical properties of the known structure-selective endonucleases than ligated dHJs [36] . The high frequency of trans hDNA we found among NCOs , along with previous analyses of recombination in wild-type and mutant Drosophila [8] , [18] , [26] , suggests that many or most NCOs may arise from an intermediate in which both resected DSB ends are engaged with the same chromatid from the homologous chromosome and are extended by synthesis . This intermediate is identical to a nicked-dHJ that we hypothesize to be a precursor to COs . Together , these results suggest the simple model illustrated in Figure 6 . A central feature of this model is that both NCOs and COs come from the same two-end engagement intermediate . NCOs are produced when this intermediate is disassembled by a helicase , whereas COs are produced when it is cleaved by a structure-selective endonuclease . A two-end engagement intermediate also occurs in current models of recombination based on data from yeast ( Figure 1D ) , but it is thought to be only a precursor to a final joint molecule with ligated HJs . This model seems to be at odds with the argument that unrepaired trans hDNA in the mei-9 mutants comes from dissolution of ligated dHJs ( see discussion above and ref 8 ) . We hypothesize that crossover formation involves protection of intermediates from helicase-catalyzed disassembly , perhaps by the mei-MCM complex [37] , prior to resolution by the MEI-9 complex . In the absence of MEI-9 , protection of the crossover-designated intermediate may persist until breakdown of the synaptonemal complex and recombination nodules , after which repair follows a pathway more like that in mitotic cells ( similar to return-to-growth experiments in yeast ) . This may involve immediate disassembly or cleavage of the unligated dHJ , or ligation into a dHJ and then resolution or dissolution . MMR may occur before or after these processes . The extremely low frequency of unrepaired trans hDNA in the mei-9 mutant ( only 3 of 32 NCOs ) suggests that we may have detected only a fraction of the events – those that were ligated and then dissolved prior to MMR; other intermediates may have been subject to MMR , either prior to or without ligation , or after resolution . If NCOs and COs come from the same intermediate , gene conversion tract lengths would be expected to be similar between NCOs and COs . We used a modification of TractSeq [38] to estimate lengths of hDNA tracts in NCOs and COs recovered in the absence of mismatch repair . For NCOs with trans hDNA , we considered each of the two halves to be an independent tract , since each is predicted to have the same origin as the single tract in NCOs without trans hDNA and the single tract of hDNA in COs ( Figures 1 and 5 ) . The mean length of NCO tracts was 710 bp ( n = 22; SEM = 111 bp ) , in good agreement with a previous estimate of 706 bp based on analysis of purine-selected NCO gene conversions in ry [39] . The mean length of hDNA tracts associated with COs was 773 bp ( n = 16; SEM = 243 bp ) ; this is not significantly different from the NCO tract length ( P = 0 . 7985 ) ( Figure 7A ) . Genetic studies in S . cerevisiae have found that tracts that are bi-directional , and therefore would give trans hDNA if unrepaired , are highly asymmetric in length with respect to the DSB [32] , [33] , [40] , [41] . Among the seven NCOs with trans hDNA that we recovered , the average length of the shorter sides was 361 bp and the average of the longer sides was 939 bp ( P = 0 . 0261 ) ( Figure 7B ) . This suggests that asymmetry may also be a feature of recombination in Drosophila; however , visual inspection suggests that there may be two classes of NCO , one symmetric and one asymmetric ( Figure 7B , black dotted lines versus blue dashed lines ) . It difficult to make definitive conclusions about tract length differences from our data . Although whole-chromosome [42] and whole-genome [43] analyses indicate that ry is a typical locus with regard to recombination frequency , this frequency is nevertheless quite low . Our screening of more than a million larvae still yielded a somewhat small sample size . Also , the ability to detect hDNA or gene conversion tracts and the resolution with which they can be mapped is highly dependent on marker spacing , and the particular spacing of markers in the ry alleles we used may have impacted measurements for NCOs and COs differently . It should also be noted that selection for ry+ recombinants should enrich for longer NCO tracts [39] . DSBs are thought to be made throughout the ry gene , rather than just near the 5′ end as in yeast [8] , [44] . The longer a tract is , the greater the probability it will span one of the two mutant sites , which is required to generate a ry+ allele that will survive purine selection . This selection does not impact COs the same way because any CO between the two ry mutations should be recoverable if it generates a ry+ chromatid . There may be some selection against extremely long tracts , since these may cross a mutant site . In the absence of MMR , this will only matter if the wild-type allele is fully converted to the mutant allele , but we did not detect this pattern among COs ( Figure 5 ) . Further studies either at additional loci or genome-wide analyses that do not rely on selection should provide more accurate measurements of tract lengths . In budding yeast , genetic data from several loci show that most NCO gene conversion tracts are uni-directional ( cis ) , extending to only one side of the DSB [32] , [33] , [40] , [41] . The small number of tracts in these studies that appear to be bi-directional ( trans ) have been explained as the result of multiple , closely spaced DSBs [33] or dHJ dissolution [32] . A single-end invasion intermediate has been detected in physical studies , but this is thought to be a precursor to dHJs and COs , not NCOs [4]; pre-NCO intermediates have not been detected in these assays [45] , [46] . These molecular/genetic data , combined with physical analyses of recombination intermediates , have led to a model in which most NCOs arise through SDSA and there is a split into distinct NCO and CO pathways very early in repair , prior to strand invasion [2] . We found that trans hDNA is a common feature of NCOs in Drosophila: seven of the 13 NCO tracts that spanned more than a single marker had the trans orientation , and it is likely that at least some of the other six have trans hDNA that could not be discerned because one tract did not cross a marker ( Figure 4 ) . Recombination does not occur in hotspots in Drosophila [43] , [47] so it is unlikely that any of the trans tracts are the result of multiple , nearby events . Rather , it seems most likely that trans hDNA arises when both sides of the DSB interact with a homologous template and are extended by synthesis . This can occur through any of three distinct processes . First , NCOs with trans hDNA may come from dHJ dissolution . Although the genetic studies discussed above found trans hDNA to be a rare event in budding yeast , a genome-wide analysis of meiotic recombination in mutants lacking canonical MMR found trans hDNA in at least 35% of NCOs [7] . The authors of this study proposed that these came from dHJ dissolution , although they could not rule out the possibility of two-ended SDSA . This implies that dissolution is a major contributor to NCOs and that a large fraction of dHJs are dissolved into NCOs , in stark disagreement with a wealth of molecular data supporting the conclusion that dHJs are resolved exclusively or primarily into COs [45] . While the contribution of dHJ dissolution to meiotic NCO production in Saccharomyces remains debatable , we believe , based on the arguments of Radford et al . ( 2007; see above discussion also ) , that dissolution is not the most attractive model to explain the trans hDNA we found in our studies . A second possibility is two-ended SDSA , in which both ends of the DSB participate in strand exchange and synthesis . If the choice of partners is not coordinated , the two ends may engage with different homologous chromatids or one might invade the sister chromatid . Multi-chromatid intermediates have been detected in S . cerevisiae sgs1 mutants; it is thought that Sgs1 helps disassemble such intermediates [48] . An end that has been dissociated from its original partner might then engage with a different partner , potentially giving discontinuous gene conversion tracts , as have been noted in yeast [7] , [49] . Gene conversion tracts in wild-type Drosophila are never discontinuous [8] , [18] , [26] , [50] , [51] , indicating that either multiple rounds of strand exchange , synthesis , and dissociation are not a feature of meiotic recombination or that the sister is never used as a template . Furthermore , Drosophila does not have homologs of any of the canonical partner choice proteins such as Red1 or Hop1 , suggesting that homolog bias during strand invasion may be ensured by other mechanisms . Two-ended SDSA might also occur such that both ends of the DSB invade the same homologous chromatid . It seems likely that steric hindrance would prevent two ends from invading the same template simultaneously , so two-ended SDSA with the same chromatid might require that one end invade and be extended by synthesis , then dissociate before the second end invades the same template . This might explain why some NCOs we analyzed did not have detectable trans hDNA . If the nascent sequence anneals to the second end before that second end participates in strand exchange , recombination could be completed through simple , one-ended SDSA . Conversely , if the second end does undergo strand exchange and extension then dissociation and annealing , trans hDNA might be produced . Two-ended SDSA occurring this way , or with one end invading each of the two chromatids on the homologous chromosome , could explain the frequent occurrence of trans hDNA we see . A third mechanism that can produce trans hDNA involves a two-end engagement and synthesis intermediate ( Figure 6 ) . The process generating this intermediate is mechanistically distinct from two-ended SDSA because it involves 2nd-end capture ( i . e . , annealing of the resected 2nd end of the DSB to the D-loop strand displaced by synthesis ) rather than 2nd-end strand exchange , followed by repair synthesis . Since we cannot physically detect recombination intermediates in Drosophila , we cannot distinguish between two-ended SDSA and two-end engagement; however , we favor the two-end engagement model because it also explains the absence of tracts of full gene conversion in crossover products ( Figure 5 ) . In many organisms , including S . cerevisiae , meiotic DSBs are made prior to assembly of the synaptonemal complex ( SC ) [reviewed in 52] . Recombination is then used to promote chromosome pairing and synapsis and thus the ability to carry out multiple rounds of strand invasion into different partners might be favored via unstable short D-loops . In Drosophila , chromosome pairing and synapsis are achieved without recombination , and DSB formation does not occur until after chromosomes are fully synapsed [53] , [54] . This likely has important consequences for how recombination proceeds . Since homologs are already intimately paired when recombination begins , the risk of strand invasion with an inappropriate template is greatly reduced , and the structure of the SC may enforce bias toward the homolog as a recombination partner . This may allow stable engagement with the homolog to be achieved early , making multiple cycles of strand exchange and dissociation unnecessary , and allowing both ends of the DSB to engage with a homologous template , as in the two-end engagement model . The two-end engagement model is conceptually very similar to the original DSBR model of Szostak et al . ( 1983 ) in having NCOs and COs come from a single intermediate . However , in the DSBR model , the NCO/CO outcome relies on random orientation of cleavage by resolvases , such that each dHJ resolution has an equal probability of producing NCO or CO products ( Figure 1 ) . In contrast , we propose that NCOs and COs are produced through different enzymatic activities – disassembly by a helicase and cleavage by a nuclease , respectively ( Figure 6 ) . Although current models from S . cerevisiae also have NCOs arising from helicase activity and COs from nuclease activity , our model differs critically in returning to a single intermediate . Consequently , the NCO/CO decision might be made and/or enforced much later than proposed in yeast – after this late intermediate is formed . In yeast , a key step in executing the CO decision involves loading of certain proteins , including the Msh4–Msh5 heterodimer , which is thought to protect recombination intermediates from disassembly by helicases [55] , [56] . In contrast , Msh4–Msh5 focus dynamics suggest an earlier role , perhaps prior to the NCO/CO decision [57] , [58] , and Arabidopsis msh4 mutants have defects in both COs and NCOs [59] . These observations suggest that a later NCO/CO decision , as in our model , may be widespread . This does not preclude the existence of an early decision that proceeds down an NCO pathway such as one-ended SDSA , but rather adds the possibility of introducing a second control point . In fact , studies of crossover homeostasis point to two phases of crossover designation in mice [60] . Our analysis of Drosophila meiotic recombination after eliminating both canonical and short-patch MMR reveals that trans hDNA is frequent in NCOs and that MMR-independent gene conversion tracts are infrequent in COs . Although it is possible to fit these results to current models of meiotic recombination from yeast , doing so requires the addition of two-ended SDSA as a major contributor to NCO formation and biased crossover resolution . We favor the two-end engagement model because of its simplicity , its ability to succinctly account for all of our results , and how this model correlates with other features of Drosophila meiosis ( e . g . , DSB induction after SC formation and the absence of any orthologs of the homolog bias-promoting proteins Red1 , Hop1 , and Dmc1 ) . Some of these features may be specific to Drosophila meiotic recombination . However , reports of trans hDNA in the S . cerevisiae literature suggest that what may be a major pathway of NCO formation in Drosophila might also be a minor pathway of NCO formation in yeast , and the discussion above raises the possibility that a late intermediate that can be processed into CO or NCOs may also occur in mammals and in plants . Drosophila might provide a unique opportunity to study this pathway in more detail . Important tests of this model will include more precise determination of the frequency of trans hDNA in noncrossovers , measurements of hDNA tract length distributions , and assessment of whether the MEI-9 nuclease complex has a preference for unligated HJs over ligated HJs .
Experiments were done in flies heteroallelic for two nonsense mutations in Xpc ( also known as mus210 ) and two deletion mutations in Msh6 [8] , [28] . Thirty females of the genotype XpcG1/XpcC2; P{GawB}h1J3 Msh668 ry531 cv-c/Msh610 kar ry606 red Sb were crossed to 10 males of the genotype y/Y , Dp ( 1:Y ) y+; kar ry506 cv-c . Purine selection was carried out on the progeny as in [8] . Briefly , adults were allowed to mate and lay eggs for three days before being removed , and then an aqueous purine solution was added to the medium . This treatment kills ry mutant larvae , but rare ry+ recombinants survive . Previous experiments demonstrated that larvae that are mosaic due to loss of mismatch repair survive as well as fully wild-type larvae [8] . One bottle in every tray of 25 was left untreated so adult progeny could be counted to estimate the total number of larvae screened . Previous studies of recombination at the ry locus demonstrated that essentially all recombinants arise during female meiosis [61] . This is evident in the observation that each treated bottle has zero or one surviving ry+ adult fly . In experiments reported here , however , there were six cases of clusters of ry+ progeny in a single bottle . Most or all of these appear to result from recombination between the ry531 and the TM3 balancer chromosome in the stock , prior to generating heteroallelic females . In numerous previous experiments of this type in our laboratory [8] , [18] , [26] , [51] , we have observed only a single similar case ( KP Kohl and JS , unpublished ) . The rate may be higher in the experiments here because of simultaneous reduction in both XPC and MSH6 . However , since all such events happened in one of the two stocks , it may be the presence of two balancer chromosomes ( CyO for chromosome 2 and TM3 for chromosome 3 ) that led to an increase in recombination in the ry region . These events were excluded from our analysis , since they occurred in a previous meiotic or mitotic cell cycle . Recombinant flies were homogenized to isolate DNA . Sequences from ry were amplified by PCR , using primers anchored in the ry506 deletion so as to amplify only the maternal , recombinant chromosome . To avoid PCR-mediated recombination , an extension time of one minute per kilobase was used and amplification was limited to 25 cycles . Bulk PCR product was sequenced to confirm whether the recombination event was a crossover or noncrossover and to map locations of gene conversion tracts and hDNA . To determine the orientation of hDNA markers on the two strands , PCR amplicons were isolated through Topo-TA cloning ( Invitrogen Life Technologies ) and individual colonies were sequenced . Table S1 shows the polymorphisms used in this study . Tract lengths were estimated using a modification of TractSeq [38] . Each tract has a minimum length determined by the outmost included markers and a maximum length determined by the nearest non-included markers . TractSeq uses a truncated exponential to find the most likely length of each tract . For the variable p , which is the probability of extending one additional base , we used 0 . 99717 , a value derived previously to estimate the lengths of gene conversion tracts in ry [39]; however , the same conclusions were reached when we varied p from 0 . 990 to 0 . 999 , the value used by Rockmill et al . [38] for experiments in S . cerevisiae . Our modification uses the same method for tracts that include a single marker as for tracts that include multiple markers .
|
During meiosis , breaks are introduced into the DNA , then repaired to give either crossovers between homologous chromosomes ( these help to ensure correct segregation of these chromosomes from one another ) , or non-crossover products . Meiotic break repair mechanisms have been best studied in budding yeast , leading to detailed molecular models . Technical limitations have prevented directly testing these models in multi-cellular organisms . One approach that has been tried is to map segments of DNA that are mismatched , since different models predict different arrangements . Mismatches are usually repaired quickly , so analyzing these patterns requires eliminating mismatch repair processes . Although others have knocked out the primary mismatch repair system , we have now , for the first time in an animal , identified the secondary repair pathway and eliminated it and the primary pathway simultaneously . We then analyzed mismatches produced during meiosis . Though the results can be fit to the most popular current model from yeast , if some modifications are made , we also consider a simpler model that incorporates elements of the current model and of earlier models .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"nucleotide",
"excision",
"repair",
"meiosis",
"cell",
"biology",
"chromosome",
"biology",
"molecular",
"biology",
"cell",
"cycle",
"and",
"cell",
"division",
"mismatch",
"repair",
"biology",
"and",
"life",
"sciences",
"dna",
"gene",
"conversion",
"dna",
"recombination",
"dna",
"repair",
"cell",
"processes",
"genetics",
"molecular",
"genetics",
"homologous",
"recombination"
] |
2014
|
Eliminating Both Canonical and Short-Patch Mismatch Repair in Drosophila melanogaster Suggests a New Meiotic Recombination Model
|
Porcine cysticercosis is caused by a zoonotic tapeworm , Taenia solium , which causes serious disease syndromes in human . Effective control of the parasite requires knowledge on the burden and pattern of the infections in order to properly direct limited resources . The objective of this study was to establish the spatial distribution of porcine cysticercosis in Mbulu district , northern Tanzania , to guide control strategies . This study is a secondary analysis of data collected during the baseline and follow-up periods of a randomized community trial aiming at reducing the incidence rate of porcine cysticercosis through an educational program . At baseline , 784 randomly selected pig-keeping households located in 42 villages in 14 wards were included . Lingual examination of indigenous pigs aged 2–12 ( median 8 ) months , one randomly selected from each household , were conducted . Data from the control group of the randomized trial that included 21 of the 42 villages were used for the incidence study . A total of 295 pig-keeping households were provided with sentinel pigs ( one each ) and reassessed for cysticercosis incidence once or twice for 2–9 ( median 4 ) months using lingual examination and antigen ELISA . Prevalence of porcine cysticercosis was computed in Epi Info 3 . 5 . The prevalence and incidence of porcine cysticercosis were mapped at household level using ArcView 3 . 2 . K functions were computed in R software to assess general clustering of porcine cysticercosis . Spatial scan statistics were computed in SatScan to identify local clusters of the infection . The overall prevalence of porcine cysticercosis was 7 . 3% ( 95% CI: 5 . 6 , 9 . 4; n = 784 ) . The K functions revealed a significant overall clustering of porcine cysticercosis incidence for all distances between 600 m and 5 km from a randomly chosen case household based on Ag-ELISA . Lingual examination revealed clustering from 650 m to 6 km and between 7 . 5 and 10 km . The prevalence study did not reveal any significant clustering by this method . Spatial scan statistics found one significant cluster of porcine cysticercosis prevalence ( P = 0 . 0036; n = 370 ) . In addition , the analysis found one large cluster of porcine cysticercosis incidence based on Ag-ELISA ( P = 0 . 0010; n = 236 ) and two relatively small clusters of incidence based on lingual examination ( P = 0 . 0012 and P = 0 . 0026; n = 241 ) . These clusters had similar spatial location and included six wards , four of which were identified as high risk areas of porcine cysticercosis . This study has identified local clusters of porcine cysticercosis in Mbulu district , northern Tanzania , where limited resources for control of T . solium could be directed . Further studies are needed to establish causes of clustering to institute appropriate interventions .
Porcine cysticercosis is caused by the larval stage of the tapeworm Taenia solium , which also infects human and may cause serious neurological disorders such as epilepsy [1] . The lifecycle of T . solium involves a human as the sole natural definitive host carrying the adult parasite in the small intestine resulting from consumption of insufficiently cooked meat infested with parasite larvae . A person infested with the adult parasite can transmit the infection to intermediate hosts including pigs and humans through faecal-oral transmission such as faecal contamination of foodstuffs or by other means leading to cysticercosis . Human neurocysticercosis , the infection of the central nervous system by the larval stages of T . solium , has been found to account for up to 50% of epilepsy in areas where T . solium is endemic [1] . Eradication of T . solium is considered epidemiologically possible because of several factors , including the fact that the human being is the only source of infection to the intermediate hosts , commonly pigs , and that the latter can be used to monitor the dynamics of the infection because of their short life span [2] . In addition , efficacious diagnostic and therapeutic agents for the parasite are currently available , which facilitate surveillance and control activities . Because of this perceived simplicity , the International Taskforce for Disease Eradication recommended that elimination of T . solium be implemented in a sizeable geographical area [3] . Nevertheless , T . solium has not been eliminated in developing countries because of poverty combined with a lack of sanitation , the need to involve both the agricultural and health sectors in control efforts , and the lack of political will , which may be attributed to poor knowledge on the presence , magnitude and impact of the parasite . Because of limited resources in many developing countries , control efforts for T . solium should be targeted to specific ( preferably small ) areas of high endemicity to reduce the burden , which could possibly eliminate the parasite in the long run . However , the spatial patterns of T . solium infections in many endemic settings are not well known . Such information could be used to guide efficient and effective utilization of limited financial and personnel resources through targeting interventions to areas of high endemicity . Analysis of spatial point pattern of porcine cysticercosis in an endemic situation would be very valuable in this aspect . Spatial pattern analysis has been found to be useful in the control of other human helminth infections elsewhere . It may be a key factor in understanding disease transmission when there is strong correlation between the spatial distribution of the disease and hosts or transmission risk factors . For example , a study in coastal Kenya found a significant clustering of urinary schistosomiasis around a water-contact site with high numbers of snails shedding Schistosoma haematobium cercariae [4] . Similar findings have been reported in Uganda [5] . A study in Belgium and The Netherlands established a spatial correlation of the fox tapeworm Echinococcus multilocularis in Red foxes [6] . Nevertheless , this is the first time that spatial pattern analysis is being applied in T . solium research in Africa . A study in Peru observed significant clustering of human cysticercosis seropositivity by the household whether or not the household had a tapeworm carrier [7] . Widdowson and others reported household clustering of porcine cysticercosis in Mexico [8] . However , both of these studies did not disentangle the clustering of the infections that could occur just because of natural environmental variation in pig or human population density , and hence , the studies were unable to ascertain the association between the risk factors and the observed infection clustering . Recently , a prevalence study by Morales and others found no clustering of porcine cysticercosis by household over what would be expected under natural environmental variation [9] . On the other hand , Lescano and others in Peru found an increasing clustering of porcine cysticercosis prevalence and incidence ( based on antibody detection ) towards households with tapeworm carriers [10] . The observed gaps in research for porcine cysticercosis as well as the conflicting findings from previous studies call for more research and emphasise the need for a context-specific situational analysis regarding the pattern of porcine cysticercosis to guide control efforts . In Tanzania , a previous prevalence study in Mbulu district based on lingual examination of village pigs found clustering of porcine cysticercosis at village level [11] . Nevertheless , because the study was based on a less sensitive diagnostic test and the fact that it described clustering of the infection by a risk factor , it could not guide on specific geographical areas that needed more attention given limited resources to control the parasite . A Bayesian analysis of incidence data from a randomized field intervention trial based on health education found no clustering of the incidence rate ratio between the intervention and control villages [12] . Because of small sample sizes in many villages and exclusion of clustering by household , this study could not rule out clustering at village level . We found a need to identify priority geographical areas that needed more attention towards control of porcine cysticercosis in Mbulu district given limited resources . The present paper is based on secondary analysis of prevalence and incidence data collected in Mbulu district , northern Tanzania [12] . The study examined spatial clustering of porcine cysticercosis in Mbulu district , which was confirmed using Ripley's K functions and spatial scan statistics .
This randomised field intervention trial is registered with the Australian New Zealand Clinical Trials Registry ( ANZCTR ) , one of the WHO International Clinical Trials Registries . The registration number is ACTRN12609000190202 . In Tanzania , the trial was approved by the Tanzania National Institute for Medical Research ( NIMR ) and Ministry of Health ethics review board , with reference number NIMR/HQ/R . 8a/Vol . IX/88 . The trial was also approved by Sokoine University of Agriculture ( the principal researcher's institution ) with reference number SUA/ADM/R . 1/8 , and the WHO/FAO Collaborating Center for Research and Training on Emerging and Other Parasitic Zoonoses , based at the Faculty of Life Sciences , University of Copenhagen , Denmark . Approval by these institutions was based on review of the research proposal . We obtained verbal consents from all study participants and their village authorities after the principal researcher had explained the purpose of and possible benefits from the study , as well as the freedom of the farmers to refuse participation . The verbal consents were recorded in a spreadsheet , which was finally approved by the institutional review board ( the NIMR ) before commencement of the field trial . We could not obtain written consents because of the high level of illiteracy ( approximately 30% ) in the study area and unwillingness to write by those who had some formal school education . Information from individual study participants were not disclosed to others . The use of the pigs in this research was based on the approval by the principal researcher's institution and the Mbulu District Livestock Department , which are responsible for animal care and use . Mbulu District is located in north-eastern Tanzania , between latitude 3 . 80° and 4 . 50° S , and between longitude 35 . 00° and 36 . 00° E . The altitude ranges from 1000–2400 m above mean sea level . The district contains areas having semi-arid and sub-humid climates that receive annual rainfall of <400 mm and >1200 mm , respectively . The long rainy season extends from March to mid-May and the short rainy period extends from November to December . Relative humidity ranges from 55 to 75% and mean annual temperature ranges from 15 to 24°C [13] . Currently , the District has 72 villages distributed in 16 wards , with an average of four villages per ward . The 2002 National census counted 237 882 people living in 38 729 households ( average six persons per household ) . In 1997 the Mbulu District pig population was estimated at about 35 000 , and crop and livestock production were by far the most important economic activities , employing >90% of the total labour force [13] . This study is a secondary analysis of data collected during the baseline and follow-up periods of a randomized field trial that aimed at reducing the incidence rate of porcine cysticercosis through an educational programme . The baseline consisted of a cross-sectional study conducted between July 2002 and July 2003 to enable randomisation and evaluation of a health education intervention trial [12] . The baseline study was followed by a randomized controlled community trial conducted between July 2003 and April 2004 . Sentinel piglets were provided to a random sample of pig-keeping households in the 42 villages and the effect of health education on the incidence rate of porcine cysticercosis and related transmission factors was evaluated after half of the villages had received a health education while the other half was left as control ( received no intervention ) . The randomisation process was stratified by the baseline median prevalence of infection . More details of the methodology and some findings from this randomized trial have been provided elsewhere [12] , [14] , [15] , [16] . The flow of participants during the randomised trial was described following the primary analysis of the data [12] . Figure 1 specifically shows the flow of participants selected for the present secondary analysis of the prevalence data and the incidence data based on lingual examination and antigen enzyme-linked immunosorbent assay ( Ag-ELISA ) , respectively . The village was the primary sampling unit . The choice of the number of villages to be included in the study was based on eligibility , whereby all eligible villages were included . The eligibility criteria for a village to participate in the study were that a village was keeping pigs , it had not been excessively studied for porcine cysticercosis , had ≥20 pig-keeping households , the village leaders agreed to participate , and the village was virtually independent from other villages . Based on the above criteria 42 out of 72 villages of the district qualified and they were enrolled in the study . Most of the villages excluded were administrative divisions of the studied villages [12] . In addition , two entire wards were excluded from the study because one ( Yaeda Chini ) was inhabited by a Bushmen society , which was not keeping animals , and the other ( Mbulu Mjini ) was occupied mostly by urban area of the district , where pig keeping was not popular . Selection of pig-keeping households and pigs for the prevalence study was done randomly as described previously [12] . A total of 784 pig keeping households were included in the study , with the same number of pigs ( one per household ) examined for porcine cysticercosis prevalence . The eligibility criteria for a household to participate were presence of at least one pig 2–12 months old and willingness of the owner to participate the baseline and follow-up studies . Pigs aged 2 to 12 months were examined in the baseline study to match with a planned 12 months follow-up study [12] . Unfortunately , about 53% of participants dropped out after the baseline study . The reasons for the dropout was mainly poverty , though a few farmers were worried about an outbreak of African Swine Fever in a neighbouring district . Potential for selection bias was examined before proceeding to the follow-up study . The randomized community-trial , which involved a follow-up of sentinel pigs to estimate incidence rates post-randomization , commenced immediately after the prevalence study . One out of the 14 studied wards was excluded in the present analysis because the randomization of the intervention by village assigned its village to the intervention group , hence not eligible for this analysis . Therefore , the total number of villages included in the control group for the incidence study was 21 located in 13 wards . Among the control group , a total of 295 households were provided with sentinel pigs ( one per household ) purchased in the district following screening with lingual examination . Blood samples were also collected and subsequently tested with Ag-ELISA . Briefly , approximately 3 ml of blood were collected from the jugular vein of each pig into a vacutainer tube . The samples were centrifuged in the same day and frozen at −21°C until their analysis , which was done after the end of field data collection . The Ag-ELISA was performed as described by Dorny and others [17] . The assay uses monoclonal antibodies to detect circulating T . solium antigens , an indication of presence of viable infection . All pigs included in the incidence study were free from infection based on both the lingual and Ag-ELISA . None of the pigs used in the prevalence study was used in the incidence study . Upon arrival to a household , the selected pig was restrained in a standing position using a pig snare . A wooden rod was twisted gently between the lower and upper jaw , and by using a piece of cotton cloth , the tongue was gently pulled out and the under-surface visually examined for presence of cysticerci of T . solium . The examination was performed by the principal investigator ( HAN ) who had been trained in Veterinary Medicine and Veterinary Public Health . Sentinel pigs aged 1–6 months ( median 2 months ) were reassessed once or twice following randomization to determine occurrence of cysticercosis using lingual examination and Ag-ELISA . For the pigs that completed the follow-up study , 32% were reassessed twice while 68% were reassessed once by Ag-ELISA . On the other hand , 44% and 56% were reassessed twice or once , respectively , by lingual examination . In each study household , we recorded the location using a handheld geographical positioning system ( GPS ) receiver ( Garmin , made in Taiwan ) at an accuracy of 10 metres on average . All the geographical data were transferred to a field data sheet right in the field . Data were entered , coded , and cleaned in Microsoft Office Excel . Epi Info 3 . 5 was used to analyse the prevalence of porcine cysticercosis . We examined the potential for selection bias because of the participants dropout by performing three preliminary analyses to compare the distribution of 45 baseline variables between the dropouts and participants . In the first analysis , important baseline characteristics of the dropout participants were analysed to assess if there was any important difference between those who dropped out of the intervention and those who dropped out of the control groups . A total of 15 different variables were examined in this analysis . In the second analysis , the 15 variables were assessed to see whether there was any difference in the baseline proportions between the dropout and full participant households . In the final analysis , households that participated fully in the study were analysed to examine the distribution of the 15 baseline variables between the intervention and control group . ArcView 3 . 2 was used to map the prevalence and incidence of porcine cysticercosis . We used ECESPA and SPATSTAT packages ( available at http://cran . r-project . org/web/packages/ ) to compute Ripley's K functions in R statistical software to assess for general clustering of porcine cysticercosis . Bernoulli probability model was used in SatScan to identify local clusters of porcine cysticercosis prevalence at household level . Discrete Poisson model was used for the incidence data . In both cases , we used isotonic spatial scan statistic to take into account the heterogeneity of the study population . Ward specific relative risks of infections were estimated from the prevalence and incidence data using SatScan .
The overall prevalence of porcine cysticercosis at the pig level was 7 . 3% ( 95% CI: 5 . 6 , 9 . 4; n = 784 ) based on lingual examination . The incidence rate at the end of follow up of sentinel pigs was 69 ( 95% CI: 65 , 72 ) per 100 pig-years and 25 ( 95% CI: 23 , 28 ) per 100 pig-years , using Ag-ELISA and lingual examination , respectively [12] . The spatial point pattern of porcine cysticercosis prevalence in household pigs is presented in Figure 2 . However , because of the dropout of some prevalence-study participants before their locations were recorded , the spatial map includes only 370 households . The spatial distribution of porcine cysticercosis incidence in the sentinel pigs based on Ag-ELISA and lingual examination are shown in Figure 3 . Locations of 6 of the 295 households included in the incidence study were not recorded . In addition , 53 and 48 households were lost to Ag-ELISA and lingual examination follow-up because their pigs died or got lost before reassessment . Therefore , the incidence study examined 236 , 241 , and 235 pigs by Ag-ELISA , lingual examination , and both tests , respectively . The median followed-up time was 4 months ( range: 2–9 months ) .
This study has for the first time described the spatial distribution of porcine cysticercosis in an endemic area of Africa . The study has also found that despite the low sensitivity of the lingual examination method for detecting porcine cysticercosis , it can be useful in identifying geographical areas where interventions should be directed . Nevertheless , the study suggests the need to use a combination of analytical procedures for effective identification of potential ‘hotspots’ of infections . One limitation of this study is lack of Ag-ELISA in the prevalence study , which limited assessment of the method in identifying clustering in cross-sectional studies and its comparison with the lingual examination prevalence results . In addition , the observed dropout of a large number of participants after the baseline study highlights on the need for careful recruitment of study participants intended for follow-up studies and analysis of potential for selection bias when losses to follow up are detected . This will enable the researcher to see whether or not results obtained from the remaining sample can be generalised to the target population . The present study has established an overall significant clustering of porcine cysticercosis in Mbulu district of northern Tanzania , and identified significant local clusters of the infection . Lack of clustering of porcine cysticercosis for distances smaller than 600 m observed with Ripley's K functions could be due to omission of household clustering by the study design and the scattered nature of households in certain areas of the study district , such that only a few households would be found within the 600 m radius . The general significant clustering of porcine cysticercosis incidence revealed by the K functions was well supported by the presence of significant local clusters of the infections as determined by spatial scan statistics . The pattern of clustering of porcine cysticercosis incidence described by the K functions matched well with that found by spatial scan statistics , that is , the presence of one continous cluster by Ag-ELISA and two discrete clusters based on lingual examination method . One controversial finding in this study is the identification of a significant cluster of porcine cysticercosis prevalence by spatial scan statistics , but , failure of the K functions to recognise such clustering . Nevertheless , the fact that there is approximately 50% overlap of this cluster with that of Ag-ELISA incidence , there is a strong reason to believe that it is most likely cluster . Although possible low sensitivity of K functions to detect clustering of disease with low prevalence could be speculated , further studies are needed to ascertain this . The use of various methods to identify clusters of porcine cysticercosis in this study has highlighted the potential strength of each method . The overlapping of clusters identified by the different methods strengthens the impression that the areas are ‘hotspots’ . For example , the identification of Bargish and Tlawi wards as low risk wards by all three studies suggests that the wards are of low risk despite their inclusion in the clusters . On the other hand , the identification of Sanu , Kainam , Murray , and Gehandu as high risk wards by at least two of the methods , suggests that these wards are important areas of parasite transmission , calling for urgent attention . Although clusters from both the prevalence and incidence studies included the same wards , though at varying proportions , the cluster identified by the prevalence study was slightly differently located as compared to the clusters identified by the incidence studies . This could be due to the fact that the prevalence study included household pigs which we could not ascertain whether the observed infections were acquired locally given the dynamic nature of pigs in the area . Some of these pigs could have been introduced in the households while infected . For the incidence study , all pigs were serologically confirmed free from cysticercosis at the beginning of the study , and they were restricted from movement in terms of exchange until the end of the study . Therefore , infections that developed later on were most likely acquired at the household or at most around a small geographical area , where a pig could roam . The use of pig-months at risk as the background population at risk , which took into account the dynamic nature of the pig population , makes the incidence study results probably more relevant . The pig dynamicity may also explain the observed significant difference between the point prevalence and incidence rate determined by lingual examination despite the different sampling periods . The incidence study was able to record all cysticercosis incidents in the study population during the year , while the prevalence study captured the existing cases in the population at one time point . Some of cases could have been sold , moved , or died . While the Ag-ELISA incidence identified one large cluster , the lingual examination incidence found two relatively small clusters , mostly contained within the Ag-ELISA cluster . This can be explained by the low sensitivity of the lingual examination method in detecting porcine cysticercosis , which could lead to the lingual examination being able to detect the cases later than the Ag-ELISA . Our results suggest that the lingual examination method is likely to be more sensitive where there is high infection pressure , for example , within the observed very high risk wards . It could also be possible that the actual cluster is irregular rather than circular as the SatScan could identify such a cluster as a series of small discrete clusters [22] , in which case the lingual examination incidence cluster pattern would be more real . Unfortunately , there is no gold standard to ascertain the actual pattern of clustering in the study area . Clustering of porcine cysticercosis in Mbulu district seemed to include certain wards , but not necessarily covering each ward entirely . However , for disease control purposes , we would recommend that any high risk ward included in the clusters be considered in totality because of shared administrative responsibilities and other common issues . In addition , despite the observed cluster boundaries ( circle perimeters ) , it is not easy to identify them in the real world . It should also be noted that the observed Mbulu ward boundaries are apparently arbitrary , which could lead to misclassification of points if only visual impression is relied upon . This was the reason we also assessed the ward specific relative risks of infection using the actual administrative data and ward centroids . Nevertheless , spatial scan statistics could also identify high risk areas as low risk and vice versa . Thus , the need for multiple methods to identify clustering , the approach we used in this study . The observed significant clustering of the incident cases is a possible indication of the clustering of parasite transmission risk factors . In May 2004 following the completion of the study , all study households received health education to reduce porcine cysticercosis . The health education consisting of training by a trained local livestock extension officer , a video show , and distribution of one booklet and leaflet to each participant was administered at village level . The present study recommends current situational analysis and use of combined interventions for ultimate elimination of the parasite . Few studies have examined for possible clustering of T . solium infections in endemic areas . For example , in Peru , Garcia and others observed significant clustering of human cysticercosis seropositivity by the household whether or not the household had a tapeworm carrier [7] . Widdowson and others reported household clustering of porcine cysticercosis in Mexico [8] . However , both of these studies did not disentangle the clustering of the infections that could occur because of natural environmental heterogeneity . Recently , a prevalence study by Morales and others found no clustering of porcine cysticercosis by household over what would be expected under natural environmental variation [9] . A Bayesian analysis found no village-level clustering of the incidence rate ratio of porcine cysticercosis in the randomised trial where data for the present paper were derived . It is presumed that this could be due to the small number of pigs per village [12] . Lescano and others in Peru found an increasing clustering of porcine cysticercosis prevalence and incidence ( based on antibody detection ) towards households with tapeworm carriers [10] . Establishing the spatial patterns of porcine cysticercosis in an endemic situation is an important basis for implementing focused intervention given limited locations . Understanding the causal association as assessed by most of the previous studies is an important complementary to the findings of spatial pattern of the disease in order to efficiently implement appropriate interventions . Nevertheless , we emphasise the need to establish region-specific causes for clustering given conflicting findings elsewhere . The overal prevalence of 7 . 3% observed in this study is lower than that of 17 . 4% reported previously in 21 villages of Mbulu district [11] . Several factors could explain the observed great difference in the prevalence between the two studies . The first and probably the most likely factor is the fact that the present study examined pigs from 2-12 months of age while the previous study included pigs from 2 months and above , about 84% of which were between 6 months and 5 years old [11] . The old pig population is likely to be mostly infected with cysticercosis because of the tendency of most smallholder pig farmers to keep infected pigs as breeding stocks due to lack of market . Secondly , the sampling of pig-keeping households in the previous study was haphazard , whereby any other household keeping pigs was visited after the previous household as opposed to the present study whereby the actual random sampling of the households was done . In addition , in the previous study all eligible pigs in a household were examined as opposed to the present study whereby only one pig per household was examined . Some of these factors could also account for the observed clustering of the infection by village in the previous study [11] but not in the consecutive incidence study in the area [12] . Several studies have shown that Ag-ELISA can detect two or more times as many cases of porcine cysticercosis as the lingual examination method [12] . Nevertheless , the lingual method has been reported to be highly specific ( approximately 100% ) [17] . Being easily available in the field and mostly accepted by the rural pig owners , the lingual examination method provides an avenue for monitoring porcine cysticercosis in an endemic situation . This was evident in the present study as indicated by the incidence study employing both the lingual examination method and Ag-ELISA . Both tests showed similar patterns of disease clustering , although with Ripley's K functions the lingual examination method suggested two peaks of clustering as opposed to one suggested by the Ag-ELISA . The good overlap of the clusters identified by spatial scan statistics between the two diagnostic methods further confirms the utility and potential limitation of the lingual examination . Future studies should examine for possible improvement of the lingual examination method . For example , in South Africa , Krecek and others [23] found that the use of additional light source to iluminate the oral cavity of the pig improved the accuracy of the test . This should be evaluated experimentally . Note that the use of a combination of several methods to highlight on the spatial pattern of porcine cysticercosis in this study , though has increased our confidence on the results , might be too expensive to implement in endemic areas because of limited time and financial resources . Particularly , the use of incidence studies with sentinel pigs might have some practical and time limitations . This study recommends further studies to examine the ability of seroprevalence studies to identify hotspots of infection as a rapid way to guide implemention of control measures for T . solium . Once data on geographical locations and infection statuses have been obtained , subjecting the data to a variety of spatial and other analyses can be done to establish the spatial pattern of the infection or risk factors .
|
Taenia solium is a tapeworm that causes two different disease conditions . In its adult stage , it inhabits the small intestine of human , a condition known as taeniosis , which is characterised by mild symptoms including abdominal disconfort . In the larval stage , T . solium can infect humans and various animal species , mainly pigs , causing cysticercosis . Taeniosis is acquired through consumption of inadequately cooked infected meat , while cysticercosis is acquired through ingestion of tapeworm eggs in foodstuffs contaminated with faeces from a human tapeworm carrier . Cysticercosis of human central nervous tissues ( neurocysticercosis ) causes serious syndromes such as epilepsy . Transmission of T . solium is facilitated by several factors such as presence of tapeworm carriers , poor sanitation and poor pig husbandry , which allow pigs to access human faeces . Nevertheless , the role of these factors in parasite transmission may vary with different cultural settings . Following an incidence and a prevalence studies in a rural area of northern Tanzania , there was a significant spatial clustering of porcine cysticerocis , suggesting focal distribution of transmission risk factors , which could be targeted for interventions . The study also revealed that despite the low sensitivity of the lingual examination method to detect porcine cysticercosis , it could highlight the potential ‘hotspots’ of the infection .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/helminth",
"infections",
"public",
"health",
"and",
"epidemiology/preventive",
"medicine",
"infectious",
"diseases/epidemiology",
"and",
"control",
"of",
"infectious",
"diseases",
"infectious",
"diseases/neglected",
"tropical",
"diseases"
] |
2010
|
Spatial Clustering of Porcine Cysticercosis in Mbulu District, Northern Tanzania
|
Survival within macrophages is a central feature of Mycobacterium tuberculosis pathogenesis . Despite significant advances in identifying new immunological parameters associated with mycobacterial disease , some basic questions on the intracellular fate of the causative agent of human tuberculosis in antigen-presenting cells are still under debate . To get novel insights into this matter , we used a single-cell fluorescence resonance energy transfer ( FRET ) -based method to investigate the potential cytosolic access of M . tuberculosis and the resulting cellular consequences in an unbiased , quantitative way . Analysis of thousands of THP-1 macrophages infected with selected wild-type or mutant strains of the M . tuberculosis complex unambiguously showed that M . tuberculosis induced a change in the FRET signal after 3 to 4 days of infection , indicating phagolysosomal rupture and cytosolic access . These effects were not seen for the strains M . tuberculosisΔRD1 or BCG , both lacking the ESX-1 secreted protein ESAT-6 , which reportedly shows membrane-lysing properties . Complementation of these strains with the ESX-1 secretion system of M . tuberculosis restored the ability to cause phagolysosomal rupture . In addition , control experiments with the fish pathogen Mycobacterium marinum showed phagolysosomal translocation only for ESX-1 intact strains , further validating our experimental approach . Most importantly , for M . tuberculosis as well as for M . marinum we observed that phagolysosomal rupture was followed by necrotic cell death of the infected macrophages , whereas ESX-1 deletion- or truncation-mutants that remained enclosed within phagolysosomal compartments did not induce such cytotoxicity . Hence , we provide a novel mechanism how ESX-1 competent , virulent M . tuberculosis and M . marinum strains induce host cell death and thereby escape innate host defenses and favor their spread to new cells . In this respect , our results also open new research directions in relation with the extracellular localization of M . tuberculosis inside necrotic lesions that can now be tackled from a completely new perspective .
The intracellular localization of bacterial pathogens has important consequences for the sensing by the host and the induced host immune responses . Hence , numerous studies have investigated the intracellular niche of the microbes . Particularly , phagolysosome fusion is a central cellular mechanism used by host cells to cope with infection . Intracellular pathogens like Mycobacterium tuberculosis are known to avoid lysosomal fusion through the manipulation of host signal transduction pathways [1] . Following phagocytosis by a host macrophage and/or dendritic cell , M . tuberculosis typically resides in a phagosomal compartment that maintains many characteristics of an early endosome . The maturation towards an acidified phagolysosome is blocked or retarded by M . tuberculosis [2] , [3] . This particularity is thought to be linked to the capacity of the bacterium to persist and replicate within macrophages [4] , [5] . However , as has been recently reported by cryo-immunogold electron microscopy , other cellular mechanisms that involve translocation of M . tuberculosis from the phagosome into the cytosol of the infected host cell might play crucial roles at later time points of the infection process [6] . In the same study it was shown that the attenuated Mycobacterium bovis Bacille de Calmette et Guérin vaccine ( BCG ) , lacking the 6kD Early Secretory Antigenic Target ESAT-6 ( EsxA ) and its 10kD Culture Filtrate Protein partner CFP-10 ( EsxB ) due to the deletion of the region of difference RD1 [7] , [8] remained within the phagolysosomal compartment , similar to an M . tuberculosis CFP-10 transposon mutant , which was not detected within the host cytoplasm [6] . The translocation-model of van der Wel and colleagues has the potential to explain a series of observations in conflict with the often-repeated concept of M . tuberculosis remaining enclosed in endosomal compartments that resist maturation and acidification . Especially , it reconciled the MHC class I presentation of M . tuberculosis antigens , and the increased CD8 responses during M . tuberculosis infection [9] . Nevertheless the proposed model of phagosomal escape of M . tuberculosis has remained controversial due to the complexity to interpret the results entirely based on ultrastructural observations . Hence , it still awaits broad acceptance by the scientific community through independent studies and alternative experimental techniques [6] , [10] . This situation prompted us to investigate the potential phagosomal escape of M . tuberculosis and closely related mycobacteria by the means of a recently developed fluorescence microscopy approach that proved to be a powerful tool to investigate the rupture of host endocytic vacuolar membranes during the cell invasion by Gram-negative pathogens [11] , [12] . This assay requires the loading of host cells with a chemical probe that is trapped within the host cytoplasm and sensitive to fluorescence resonance energy transfer ( FRET ) measurements [13] , [14] , and a beta-lactamase activity present on the cell surface of bacteria . FRET image analysis of single infected cells can be performed in live or in fixed cells , and the intracellular localization of the bacterial pathogen can be quantified with high precision via automated image analysis tools . We took advantage of the BlaC mediated intrinsic resistance of M . tuberculosis , BCG , Mycobacterium marinum and BlaS of Mycobacterium smegmatis to beta-lactam antibiotics [15]–[17] that worked in conjunction with our FRET assay . Proteome analyses have identified BlaC as a membrane-associated protein [18] containing a lipobox motif in its signal sequence that predicts membrane-anchored cell envelope localization [19] , [20] . This feature allowed us to follow the cytosolic contact of selected mycobacteria over a given time course , to monitor its consequences and to link phagosomal rupture with the pathogenic potential of the tested bacterial species .
The aim of our study was to analyze the capacity of selected mycobacteria to reach the cytosol during macrophage infection via a robust , sensitive and quantitative approach at the single cell level . For this , we had to adapt our FRET based reporter that was previously used to monitor the intracellular localization of Salmonella and Shigella species [11] ( Figure S1 ) to the slowly growing mycobacteria . In the host cell , membrane permeable CCF-4-AM molecules diffuse freely across the cellular plasma membrane , are subsequently trapped in the cytosol and excluded from endosomes and other organelles by anion conversion into CCF-4 upon cytosolic esterase action . The use of this cytosolic probe relies on the bacterial expression and exposure of beta-lactamase . Cleavage of CCF-4 by beta-lactamase-expressing bacteria leads to a switch of the FRET signal from 535 nm ( green ) to 450 nm ( blue ) upon 405 nm excitation . As mycobacteria are naturally resistant to beta-lactam antibiotics , such as ampicillin , due to the constitutive expression of endogenous beta-lactamases [21] , we assumed that this activity could be exploited by our assay . To ascertain that the selected mycobacterial strains showed the expected beta-lactamase activity , we first demonstrated their capacity to cleave the colorimetric beta-lactamase substrate nitrocefin ( data not shown ) , and then evaluated their cleavage of the CCF-4 probe in vitro by fluorimetry . As shown in Table 1 and Figures S2A–C ( in Supporting Information ) , all tested strains displayed the necessary enzymatic activity allowing us to follow the intracellular behavior of each strain in real-time during the process of infection of human macrophages . We decided to use the THP-1 human cell line as a host cell model because upon addition of phorbol-myristate-acetate ( PMA ) it differentiates from a cell with the characteristics of monocytes to one with characteristics of macrophages . Therefore , this cell line depicts a high state of differentiation . Our experimental protocol started with the infection of THP-1 cells by mycobacteria . Then , upon loading of cells with CCF-4 it was possible to discriminate exclusive phagosomal location or host cytosol access by the change of the FRET signal . Finally , using automated microscopy and dedicated image analysis algorithms , quantification was achieved on hundreds to thousands of cells for dozens of conditions in parallel . At first , we compared the translocation capacities of M . marinum with M . smegmatis to evaluate the adaptability of our assay to mycobacteria . THP-1 cells were infected by M . smegmatis ( at 37°C ) or M . marinum expressing DsRed ( at 30°C ) at a multiplicity of infection ( MOI ) of 1 ( Figure 1 ) . As shown in Figure 1A , the M . marinum DsRed induced the switch of the FRET signal from 535 nm to 450 nm , indicating rupture of the phagosome and contact with the cytosol after 24 h to 48 h of infection . Strikingly , after 48 h of infection , virtually all the infected cells turned blue , and only the uninfected cells remained green . On the contrary , infection of the THP-1 cells with M . smegmatis did not result in a switch of the fluorescent FRET signal showing that the bacteria remained trapped within intact phagolysosomes . These results were corroborated by quantitative analysis ( Figure 1B ) , via a dedicated algorithm segmenting individual host cells and measuring both fluorescent channels [11] . While low fluorescent ratios ( around the value of 1 ) indicated the presence of membrane-enclosed bacteria , high ratios ( above 1 ) showed the presence of cytosolic bacteria . Indeed , after 24 h of infection with M . marinum DsRed , ratios went up to 3 and even reached 5 after 48h . It is noteworthy that for M . marinum similar effects were already observed at rather early time points ( 2 h , 5 h and 20 h ) ( Figures S3A–F ) . In the case of M . smegmatis infection , the ratios remained low highlighting its phagolysosomal localization . We decided to take a closer look at the molecular requirements for phagolysosomal rupture of M . marinum by using a M . marinum M strain variant that was reported to be impaired for the secretion of ESAT-6 and CFP-10 and has been named M . marinum MVU [22] ( Figure S4A ) . As shown in Figures 1A and S3A-F , this strain did not induce a FRET shift within the infected macrophages , suggesting that it was unable to reach the cytosol of the macrophage . This effect was very strong as determined by automated image quantification of 216 randomly imaged cells after 48 h of infection ( Figure 1C ) ; not even a single infected macrophage of the inspected cells showed an increase in the cytoplasmic 450/535 nm emission ratio during the course of infection for early ( Figure S2A ) or late ( Figure 1A ) time points , although the MOIs were the same as for the ESAT-6 secretion-intact M . marinum M strain . In agreement with previously published data [23] , [24] our results indicate that the cytosolic translocation phenotype of M . marinum is dependent on intact ESAT-6 secretion . This necessity is corroborated by quantitative analyses depicted in Figures 1C , S2D–F ( ratio distribution was stalled at 1 throughout infection ) . We then studied the cytosolic access of the virulent M . tuberculosis H37Rv reference strain and the BCG vaccine strain . Unlike M . marinum , these tubercle bacilli display a very long infection cycle in macrophages that can last up to 1 or 2 weeks depending on the MOI . THP-1 cells were infected with BCG DsRed or M . tuberculosis DsRed at 37°C at an MOI of 1 for a 10 day time course . As shown in Figure 2A , BCG was unable to induce a FRET switch indicating that it remained in a membrane-enclosed compartment for the whole course of the experiment , and the same was found for the BCG::pYUB412 vector control strain ( data not shown ) . In contrast , M . tuberculosis was progressively found within macrophages displaying a FRET signal switch from 535 nm to 450 nm , indicating phagosomal rupture and contact with the cytosol starting from day 3 onwards ( Figure 2D ) . It is of note that in some cells with no apparently visible red bacteria a FRET switch was measured due to the automated focusing on the equatorial plane of the inspected cells during image acquisition . Quantification of the imaging data revealed a gradual increase of the 450/535 nm ratios over the observed time course for M . tuberculosis to reach 2 . 5 ( Figure 2D ) , whereas ratios remained at a ratio of 1 for BCG ( Figure 2C ) until the end of infection at day 10 . Our data obtained with BCG are very similar to the data obtained with uninfected or paraformaldhyde ( PFA ) -killed bacteria ( Figures S5A–D ) . As in the experiments with M . marinum and M . smegmatis all acquired images were randomly chosen by the computer driving the microscope , and subsequently were automatically analyzed by our analytical pipeline . Taken together , our results suggest that M . tuberculosis is indeed able to enter into the cytosol during macrophage infection of multiple days , while BCG does not manifest this faculty , even after prolongated infection . M . tuberculosis thus shows a delayed but otherwise similar intracellular behavior as M . marinum , whose faculty to translocate into the cytosol of host cells is well established [23]–[25] . Based on results from previous reports and our results with the M . marinum ESAT-6/CFP-10 secretion null mutant ( MVU ) , we decided to also test an M . tuberculosis ESX-1deletion mutant ( M . tuberculosisΔRD1 ) [26] , which lacks a functional ESX-1 secretion system ( Figure S4B ) . In the observed time course using an MOI of 1 , we obtained similar results as for BCG , depicted by the absence of a FRET ratio switch to higher values ( Figures 3A , D ) . This means that the ESX-1 deletion mutant bacteria reside in the phagosome until the end of our experimental time course at day 10 , emphasizing that ESX-1 effectors represent secreted key factors that allow M . tuberculosis to gain access to the host cytosol . We also complemented M . tuberculosisΔRD1 with the genomic region encompassing the ESX-1 cluster of M . tuberculosis ( M . tuberculosisΔRD1::RD1; Figure S4B ) and obtained similar results as for the M . tuberculosis strain in the FRET analysis ( data not shown ) . In order to further investigate whether the involvement of the ESAT-6 and the ESX-1 secretion system is crucial to trigger the vacuolar rupture of mycobacteria , we tested recombinant BCG::RD1 , a BCG strain that expresses and secretes ESAT-6 and CFP-10 due to the integration of a 32 kb genomic region encompassing the ESX-1 cluster of M . tuberculosis ( Figure S4B ) [27] , [28] . In a time course experiment at a MOI of 1 , this recombinant strain showed a similar phenotype as the one observed during M . tuberculosis infection . Quantitative analyses showed that a gradual increase of the number of infected cells with higher FRET ratios gradually increased starting from day 3 after infection ( Figures 3B , 3E ) . This demonstrates that the ESX-1-complemented BCG::RD1 strain progressively gained access to the host-cell cytosol . In contrast , the use of the BCG::RD1-ESAT-6Δ84–95 strain that expresses and secretes truncated ESAT-6 ( Figure S4B ) , showed no switch of the FRET ratio to higher values throughout the time course , indicating the absence of bacteria in the cytosol during the measured infection period ( Figures 3C , 3F ) . These findings indicate that the cytosolic contact observed upon complementation of BCG with ESX-1 did not occur when 12 amino-acids were deleted from the secreted ESAT-6 molecule , emphasizing the importance of ESAT-6 and its C-terminal region as an effector involved in phagosomal rupture . As depicted in Figures 2D and 3E , we observed that upon infection with M . tuberculosis strains that cause phagosomal rupture , the total number of THP-1 cells diminished extensively over the time course of the experiment , suggesting that mycobacteria-induced phagosomal rupture and cytosolic contact leads to host cell death . Strikingly , we observed this phenomenon also for M . marinum , which induced cell death within 48 h ( Data not shown ) . Furthermore , by staining THP-1 cells by the plasma membrane fluorescent marker WGA , we observed that upon infection with M . tuberculosis or BCG::RD1 the number of THP-1 cells decreased progressively from day 4 to day 10 , which was much less pronounced for macrophages infected with mycobacterial strains that did not cause phagosomal rupture ( Figure 4A ) . Live/Dead Pacific Blue staining in combination with FLICA poly-caspase staining and flow cytometry of THP-1 cells at day 7 , when strain-specific differences were most pronounced ( Figure 4A ) , confirmed that rupture-proficient strains caused more extensive host cell death . Only 2–4% of the cells infected with M . tuberculosis , M . tuberculosisΔRD1::RD1 and BCG::RD1 were found viable , i . e . without any apoptosis or necrosis marker compared to 10–20% viable cells in non-infected cultures or those infected with M . tuberculosisΔRD1 , BCG or BCG::RD1-ESAT-6Δ84–95 ( Figure 4B ) . Similar results were obtained when BCG and M . tuberculosis were tested with necrosis marker 7-AAD at day 5 by flow cytometry ( Figure S6 ) . Hence , our data support the hypothesis that cytotoxicity is linked with access to the host cytosol . Inversely , cytotoxicity was profoundly attenuated when the strains carried genetic lesions that prevented disruption of the macrophage phagolysosome . To exclude that cell death alone leads to a FRET signal switch , we induced cell death in the THP-1 cells in multiple ways and found that no FRET change occurred under these conditions ( Figures S7A , B , ) . We also investigated the capacity of mycobacterial strains to gain access to the cytosol upon chemical induction of host cell necrosis . For this purpose THP-1 macrophages were infected with BCG at a MOI of 1 prior to induction of necrosis by high concanavalin A concentration ( 100 µg/ml ) for 24 h or 48 h ( Figure 5 ) . As shown in Figure 5A , concanavalin A did elicit necrosis as seen by propidium iodide nuclei staining . Nevertheless , induction of necrosis did not lead to contact of BCG with the host cytosol , as suggested by the absence of a FRET signal change during the course of the experiment ( Figures 5A–C ) . Similarly , the concanavalin A induced necrosis conditions did not change the faculty of M . tuberculosis and BCG::RD1 to cause FRET switches only at later timepoints of infection . Samples taken at an early time point ( 24 h ) exclusively showed low 450/535 nm ratios ( Figures S8A–B ) and the same was true for comparable infection conditions when red fluorescent M . tuberculosis and BCG strains were imaged ( Figure S9 ) . Together , our data suggest that phagosomal rupture induced by mycobacterial pathogens and induction of cytotoxicity are two cellular events that are intimately linked . From the observed timescale and the performed functional assays , the sequence of events can be deduced suggesting that cytosolic contact of M . tuberculosis precedes and triggers host cellular death . We propose that these observations are highly relevant for a better understanding of the underlying mechanisms responsible for the pathogenicity of M . tuberculosis as a human pathogen .
Taking advantage of recently developed tools that allow the quantitative analysis of vacuolar rupture and cytosolic entry of beta-lactamase exposing bacteria , we provide unbiased evidence that M . marinum and M . tuberculosis gain access to the cytosol of infected THP-1 macrophages . These findings help to clarify the long-standing controversy of phagosomal escape of mycobacteria . While studies in the eighties and early nineties [29]–[31] showed some evidence of M . tuberculosis in the cytosol using traditional electron microscopy , these results remained to be confirmed by other research groups due to the lack of reliable techniques . This situation resulted in the paradigm of M . tuberculosis residing generally within endomembrane compartments along the endocytic route avoiding maturation and acidification of its niche . Although the model of exclusive phagosomal containment is in agreement with the fact that M . tuberculosis induces strong CD4+ T-cell responses , which is a typical feature of MHC II class presentation of peptides from phagocytosed pathogens [32] , M . tuberculosis also elicits pronounced CD8+ T-cell responses [9] . The latter require proteasome-degraded cytosolic proteins to be presented via MHC I class molecules . Furthermore , M . tuberculosis infection also triggers type I interferon responses [33] , and NLRP3 inflammasome activation , where ESAT-6 has been proposed to facilitate the diffusion of immunomodulatory PAMPS ( Pathogens-Associated Molecular Patterns ) to the cytosol resulting in enhanced caspase 1 activation [34] . Thus , immunological data clearly suggest that mycobacterial constituents appear and are processed within the cell host cytoplasm during infection . A plausible way to interpret these findings takes into account that intracellular mycobacterial pathogens rupture the phagosomal membrane and translocate to the cytosol at some stages of the infection and thereby initiate the cytosolic recognition pathways . Evidence for such mycobacterial translocation was reported for M . marinum , which behaved similar to Shigella flexneri forming actin comet tails inside the host cytoplasm [25] . More recently , potential phagosomal escape of M . marinum was also linked to the formation of the ejectosome [35] , and/or the trapping of the pathogens within septin cages [36] . For M . tuberculosis novel ways of sample preparation under cryo-conditions and immuno-gold staining were employed to investigate the intracellular pathogen localization challenging the common view of the exclusive phagosomal localization of M . tuberculosis [6] . However , the subject has remained controversial , and the study of phagosomal rupture using independent techniques is of major scientific interest [10] . In this respect , our fluorescence microscopy-based approach for live and fixed samples bring new quantitative insight into this phenomenon using an alternative simple , specific and sensitive experimental approach . Our data provide evidence that phagosomal rupture and contact with the host cytoplasm occurs at a species time course during the infection process . This observation is even more significant as such rupture was never observed for the different mutant strains impaired for ESAT-6 and CFP-10 secretion ( Figures 2A , 3A ) , thereby demonstrating that the observed effects did not occur at random , but in a biologically relevant fashion . Strikingly , BCG devoid of the ESX-1 secretion system due to deletion of the RD1 region [7] was unable to induce a change in the 450/535 nm ratio , whereas the ESX-1 complemented BCG::RD1 strain under the same experimental conditions clearly induced a change in the ratio from green to blue , indicating cytosolic contact ( Figures 3B , 3E ) . These results extend previous reports on the potential involvement of the ESX-1 system showing that the ESX-1 system is a central component required for phagosomal rupture . Belonging to the recently described type VII secretion systems [37] , [38] , ESX-1 is one of the five ESX systems in M . tuberculosis , designated ESX-1 to ESX-5 , that is well known for its involvement in virulence and specific immune responses [39] . Indeed , M . tuberculosis , M . bovis and/or M . marinum mutants lacking an intact ESX-1 secretion system or its substrates have been shown to be attenuated in cultured macrophages and animal models of infection , thereby exhibiting defects in cell to cell spread , altered cytokine profiles [23] , [26] , [40]–[42] or phagosome maturation arrest [43] . Furthermore , it has been suggested that ESAT-6 secreted by M . marinum could play a direct role in producing pores in membranes of the vacuoles containing mycobacteria , facilitating the escape of M . marinum from the vacuole to the cytosol and cell to cell spread [44] . Similar observations were made with purified ESAT-6 from M . tuberculosis that destabilized and lysed artificial lipid bilayers and liposomes [26] , [45] . A more distantly related ESX-1 system is also present in M . smegmatis , where it is thought to serve as a putative conjugation system [46] . As M . smegmatis is an avirulent , saprophytic and fast-growing mycobacterium that did not show any phagosomal translocation in our assay , it is possible that ESAT-6 and CFP-10 from M . smegmatis , which share only 60–70% amino acid identities with their homologues from M . tuberculosis might not fulfill the same function ( s ) and/or might miss potential interaction partners . Inspection of available genome data showed that M . smegmatis is not the only non-pathogenic fast-growing mycobacterium carrying an orthologous ESX-1 system . Other species , such as Mycobacterium vanbaalenii ( GenBank: CP000511 ) , Mycobacterium gilvum ( CP000656 ) , Mycobacterium sp . JLS ( GenBank: CP000580 ) , Mycobacterium sp . KMS ( GenBank: CP000518 ) and Mycobacterium sp . MCS ( GenBank: CP000384 ) also encode ESX-1 systems that resemble ESX-1 in M . tuberculosis in gene content and gene order . In contrast , these fast-growing mycobacterial species lack orthologs of the EspACD locus that in M . tuberculosis is involved in the regulation of ESX-1 mediated virulence [47]–[50] . As the EspACD locus is present in M . marinum , M . leprae , and M . tuberculosis , which have all been described as translocation-proficient mycobacteria [6] , [24] , [25] , it might well be that phagosomal rupture requires interaction of ESX-1 and EspACD proteins . The CCF-4 based experimental approach presented here will be helpful to investigate this question in the near future . We show that mycobacterial phagosomal rupture precedes host cell death ( Figures 4 , 5 , S8 and S9 ) . Using necrosis markers , like propidium iodide we found that infected host cells harboring cytoplasmic bacteria tested positive for necrosis within one to two days upon FRET signal switch , however cells appeared as “empty bags” at the very late time points complicating the interpretation of the results . Flow cytometry confirmed that most of M . tuberculosis and BCG::RD1 infected cells stained positively for necrosis and/or apoptosis markers , whereas this effect was substantially reduced for infection with M . tuberculosisΔRD1 and BCG ( Figures 4B , S6 ) . Triggering cell death in infected macrophages is in agreement with studies suggesting the role of the ESX-1 system during this event [24] , [26] , [51] . Independent studies have investigated the different mechanisms of host cell death that may have an impact on triggering the immune response [52] . Depending on the individual cell cycle of each of the involved bacteria , it might be sufficient if only a few bacteria become cytosolic to induce the initiation of cell death supporting the hypothesis of Fortune and Rubin on phagosomal escape of M . tuberculosis , who suggested that the cytosol might only be a brief stop on the way to escape from the intracellular environment altogether [10] . Importantly , in many hosts , large numbers of bacteria have been found to be extracellular , such as those found in necrotic caseous lesions . Thus , the paradigm of intracellular growth might represent only a part of the life cycle of the infectious organisms [10] . Such a scenario would also explain why only a certain percentage of bacteria was previously found in the cytosol by EM and other techniques due to the experimental procedures . In conclusion , our presented data on the ESX-1-dependent potential of M . marinum , BCG::RD1 and M . tuberculosis to change the FRET signal in an ESX-1-dependent manner is in excellent agreement with previous studies that have suggested an involvement of ESAT-6 in membrane and/or cell lysis . The similarities with the FRET changes caused by IpaB-secreting Shigella flexneri [11] , [53] suggest that these mycobacterial strains also gain cytosolic access via rupture and disassembly of phagosomal membranes , but at later stages of infection . However , taking into consideration that the reported ESX-1-dependent type I interferon response to infection with M . tuberculosis occurs already 24 h post infection [33] , it is plausible that rupture of the phagosomal membrane is preceded and/or initiated by pore-forming activity of ESAT-6 , which might allow small signalling molecules such as cyclic diadenosine monophosphate ( c-di-AMP ) [54] or other PAMPs to translocate prior to the escape of entire bacteria and thereby trigger an early cytosolic host response . In any case , the ESX-1 mediated access to the cytosol seems to represent a major switch for many cellular parameters and the resulting immune responses [39] , [55] , which are thus substantially different between BCG and M . tuberculosis . Hence , infection experiments using BCG as a model organism might have the disadvantage of potentially missing out on this part of cellular responses that are linked to the cytosolic access of the bacteria . The BCG::RD1 strain seems to represent a good compromise , as it is much less virulent than M . tuberculosis , but due to the access to the cytosol still elicits ESAT-6 specific immune responses similar in quality and quantity to M . tuberculosis [9] . This finding seems also to be an important feature for the enhanced protective efficacy of BCG::RD1 or other ESX-1 complemented vaccine strains relative to BCG alone [27] , [56] . In a more practical sense , our findings have direct relevance for the design of new vaccines and the potential development of new therapeutic intervention strategies that might target the ESX-1 secretion system . It is clear from our study that a more effective vaccine strain should have the capacity to access to the cytosol of target cells in order to induce the series of immunologic responses ordinary BCG is unable to induce . A potential combination of the properties of BCG and M . marinum might represent one novel route to explore . Finally , the presented CCF-4 assay might also represent a powerful new system that can be automated for cell-based screening of larger compound libraries [57] in order to identify molecules that can block phagosomal rupture and thereby a whole series of events that are linked to the pathogenicity of one of the most deadly pathogens of mankind .
Wildtype and mutant strains of M . tuberculosis ( H37Rv ) , M . bovis BCG Pasteur 1173P2 , M . marinum MVU , M . marinum M Dsred and M . smegmatis ( mc2 155 ) were grown in 7H9 liquid medium ( Difco ) supplemented with albumin-dextrose-catalase ( ADC , Difco ) or on solid Middlebrook 7H11 medium ( Difco ) supplemented with oleic acid-albumin-dextrose-catalase ( OADC ) . With the exception of M . marinum , which was grown at 30°C , all other strains were cultivated at 37°C . M . tuberculosis or M . marinum heat killed strains were obtained by heating the sample at 95°C for 30 min . PFA-fixed M . tuberculosis or M . marinum bacteria were obtained after incubation with PFA 4% for 30 min . If required , the media were supplemented with 20 µg/ml of kanamycin or 50 µg/ml of hygromycin . Shigella flexneri ( M90T-AfaI ) was grown in BTCS and Salmonella typhimurium was grown in LB broth . The human pro-monocytic cell line THP-1 was maintained in RPMI 1640 glutamax and 10% heat-inactivated fetal bovine serum ( FBS ) at 37°C under an atmosphere containing 5% CO2 . For differentiation into macrophages , cells were plated into 96-well Greiner plates that work for fluorescence microscopy ( 3x104 cells/well ) and treated with 20 ng/ml of PMA ( Sigma ) for 72 h or were seeded into 24-well plates for flow cytometry ( 5x105 cells/well ) . Before use , cells were washed twice with fresh medium . M . tuberculosis and BCG strains were grown to mid-log phase in 7H9 containing albumin dextrose catalase ( ADC , Difco ) medium at 37°C . Cultures were harvested , washed , resuspended in PBS and gently sonicated to avoid clumping . The concentration of each strain was determined by OD600 measurement . Then PMA-differentiated THP-1 cells were infected with M . tuberculosis and BCG suspensions at a MOI of 1∶1 in RPMI medium during for 2 h at 37°C with 5% CO2 . After 2 h , the medium was removed and cells were washed 3 times to remove extracellular bacteria before the addition of fresh medium . Time course measurements evaluating phagosomal rupture were done at days 3 , 4 , 5 , 6 , 7 and 10 . M . marinum strains were grown to mid-log phase in 7H9 containing ADC at 30°C . Cultures were harvested , washed , resuspended in PBS and then filtered though a syringe . The concentration of each strain was determined by OD600 measurement . Then PMA-differentiated THP-1 cells were infected with M . marinum suspensions at an MOI of 1∶1 in EM medium ( 120 mM NaCl , 7 mM KCl , 1 . 8 mM CaCl2 , 0 . 8 mM MgCl2 , 5 mM glucose and 25 mM Hepes at pH 7 . 3 ) for 2 h at 30°C . After 2 h , the medium was removed and cells were washed 3 times to remove extracellular bacteria before the addition of fresh EM medium containing Hepes and 10% FBS . Time course measurements to monitor the M . mariunum phagosomal rupture were done at day 0 , 1 and 2 to monitor the M . mariunum phagosomal rupture . Ampicilline-resistant bacteria were incubated overnight at 37°C with shaking , diluted 1/100 in BTCS ( for Shigella M90T-AfaI ) or LB broth ( for Salmonella typhimurium ) , and incubated in the same conditions for 2 , 25 h . The cultures were washed with PBS . Salmonella typhimurium and Shigella flexneri were incubated with 0 . 1 mg/ml beta-lactamase ( Sigma ) or 0 . 1 mg/ml Atto594 beta-lactamase , respectively , for 10 min . After PBS washing , cultures were diluted in 10% FBS-containing DMEM/F12 ( Gibco ) and added to the cells at a MOI of 100 at 37°C . Labelling of beta-lactamase ( Sigma ) was obtained using Atto594 protein labelling kit ( Fluka ) . At the successive stages of the time course measurements , a mix containing 50 µM CCF-4 substrate ( Invitrogen ) in EM containing 2 . 5 µM probenicid was added for 2 h at RT in the dark . Cells were washed with PBS containing 2 . 5 µM probenicid before fixing with PFA 4% for 30 min at RT in the dark . Cells were washed before performing directly fluorescence imaging or additional cellular staining . The necrotic marker propidium iodide ( Fluka ) was used just before fixation for 5 minutes at 2 µg/ml in PBS in the dark . The cell membrane marker WGA 647 ( Invitrogen ) was used after fixation at 2 mg/ml in PBS in the dark for 30 min . Samples were imaged using an automated inverted fluorescent widefield microscope Nikon Ti with 20X or 40X objective driven by Metamorph . Picture acquisition was achieved randomly and automatically for all measurements of this study on 36 to 49 fields per condition . Samples were excited at 405 nm , and emission was followed to determine the 450/535 nm intensity ratio . Automated image analysis including cell segmentation and quantification was achieved using dedicated journals written for Metamorph [11] . The used Metamorph journals are available upon request . Two different flow cytometry assays were performed to evaluate cell death caused by different M . tuberculosis and BCG strains at later timepoints of infection . THP-1 cells were labeled with FLICA Poly-Caspases ( SR-VAD-FMK ) marker ( Immunochemistry Technologies ) , which detects apoptotic cells , and LIVE/DEAD Pacific Blue Fixable Dead Cell Stain Kit ( Invitrogen ) , which detects necrotic cells , according to the manufacturer recommendations . The stained cells were fixed overnight with 4% PFA and cytometric data were acquired on a Cyan system , using the Summit software ( Beckman Coulter , Villepinte , France ) . Data were analyzed with FlowJo software ( Tree Star , OR , USA ) . A parallel experiment was performed for BCG and M . tuberculosis strains using 7-AAD staining ( BD-Pharmingen ) according to the manufacturer's recommendations . The stained cells were fixed overnight with 4% PFA and analyzed using a FACS ARIA III flow cytometer and FlowJo software ( Tree Star , OR , USA ) . M . tuberculosis strains or M . bovis BCG strains were grown in liquid medium , at pH 7 for 7 days and 4 days for M . marinum strains . Cultures were harvested by centrifugation . The supernatant was recovered after filtration through 0 . 22 µm pore size filters ( Millipore ) and protease inhibitors were added ( Complete EDTA Free; Roche Diagnostics GmbH , Mannheim , Germany ) . The cell pellet was washed twice and resuspended in PBS . Cells were ruptured by shaking with 106-µm acid washed-glass beads ( Sigma ) for 8 min at speed 30 in a Mill Mixer ( MM300; Retsch GmbH , Haan , Germany ) . The whole cell lysate consisting of the supernatant fraction recovered after filtration and debris were removed by centrifugation at 17000 g for 30 min . Total protein concentrations were determined by using Nanodrop . Samples were subjected to NuPAGE Novex Bis-Tris pre-cast gel ( 12% ) ( Invitrogen ) before blotting on iBlot Gel Transfer Stacks Mini Nitrocellulose ( Invitrogen ) with iBlot Dry Blotting System . Immunoblotting was performed with mouse monoclonal anti-ESAT6 antibody ( Hyb 76-8 , Antibodyshop ) and mouse monoclonal anti-GroEL2 antibody as lysis control as described [58] . Antibodies against GroEL2 were received as part of National Institutes of Health , National Institute of Allergy and Infectious Diseases contract entitled “Tuberculosis Vaccine Testing and Research Materials , ” awarded to Colorado State University . Beta-lactamase activity was assayed spectrophotometrically with 100 mM nitrocefin ( Calbiochem ) in 0 . 1 M sodium phosphate buffer at 37°C . Hydrolysis was monitored at 486 nm using a UV spectrophotometer . The molecular extinction coefficient of hydrolyzed nitrocefin at 486 nm is 20500 M−1cm−1 . The rate of nitrocefin hydrolysis by each protein was expressed as micrograms of nitrocefin hydrolysed per minute per microgram of protein . Fluorimetric assays were performed in 1 ml PBS containing 50 µg/ml porcine esterase liver extracts ( Sigma ) and 100 nM CCF-4-AM liveblazer ( Invitrogen ) . The esterase was added to cleave off the –AM ester moieties yielding fluorescent CCF-4 . Bacteria were washed with PBS and added to the mixture for 12 h at 37°C in the dark . Soluble lactamase 1 mg/ml was used as a positive control . Fluorescence intensity measurements were then performed using PTI Quantamaster fluorimeter in 1ml quartz cuvettes . Staurosporine ( Sigma ) , cycloheximide ( Sigma ) and TNF-α ( RD systems ) were used for inducing apoptosis whereas concanavalin A ( Sigma ) and H2O2 ( Sigma ) were used for inducing necrosis .
|
Mycobacterium tuberculosis is one of the most life-threatening pathogens of all time . Despite the development of vaccines and antibiotics , this pathogen is still a major public health problem . Also the HIV epidemic has an important impact on the rise of M . tuberculosis infections since immunodeficient people are highly susceptible . Commonly , M . tuberculosis has been thought to reside in a membrane-bound compartment within its host cells during the entire infection cycle from invasion to cell death . Using a fluorescence-based method , we provide evidence that M . tuberculosis is able to rupture its membrane-bound compartment and gain access to the host cytosol , where it can elicit cell death . Furthermore , we show that this effect is dependent on a functional type VII secretion system named ESX-1 . Most importantly , we were able to track the dynamics of infection to understand the consequences of M . tuberculosis phagosomal rupture . This revealed that phagosomal rupture results in cell toxicity and host cell death involving necrosis . Together , our data provide a new angle in the worldwide fight against M . tuberculosis and could lead to new approaches in the development of innovative treatments .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology",
"microbiology"
] |
2012
|
Phagosomal Rupture by Mycobacterium tuberculosis Results in Toxicity and Host Cell Death
|
It has long been known that the brain is limited in the amount of sensory information that it can process at any given time . A well-known form of capacity limitation in vision is the set-size effect , whereby the time needed to find a target increases in the presence of distractors . The set-size effect implies that inputs from multiple objects interfere with each other , but the loci and mechanisms of this interference are unknown . Here we show that the set-size effect has a neural correlate in competitive visuo-visual interactions in the lateral intraparietal area , an area related to spatial attention and eye movements . Monkeys performed a covert visual search task in which they discriminated the orientation of a visual target surrounded by distractors . Neurons encoded target location , but responses associated with both target and distractors declined as a function of distractor number ( set size ) . Firing rates associated with the target in the receptive field correlated with reaction time both within and across set sizes . The findings suggest that competitive visuo-visual interactions in areas related to spatial attention contribute to capacity limitations in visual searches .
It has long been known that the visual system is limited in its ability to process multiple simultaneous inputs . Psychophysical evidence for visual capacity limitations comes from the observation that irrelevant distractors impair the ability to detect or discriminate a task-relevant target . Distractor interference is of several types . Distractors positioned close to the target can impair target visibility ( i . e . , the ability to detect the target and distinguish its features , generating the phenomena of lateral masking and crowding ) [1 , 2] . The critical separation for crowding—approximately half the retinal eccentricity—well exceeds visual acuity limits , suggesting that this form of interference arises in higher-order visual processing stages [1] . The “set-size effect” is another form of interference that operates over larger distances in search tasks in which the target does not pop out from the display [3–6] . The set-size effect does not affect target discriminability but increases the time needed to find a target in the presence of a large number of distractors . Although the set-size effect has been widely documented , its neural substrates have remained unknown . In extrastriate cortical areas , blood-oxygen-level-dependent functional MRI activation decreases as more stimuli are added to a display , suggesting that neuronal populations representing individual inputs engage in mutually suppressive ( competitive ) interactions [7] . Single-neuron recordings have confirmed the presence of competitive interactions in cortical visual areas V2 and V4 , and have shown that attention biases the competition in favor of the attended stimulus while suppressing the effect of distractors [8] . However , these studies have not related neuronal competitive interactions to specific forms of visuo-visual interactions . Because the set-size effect does not impair discriminability per se , it is thought to reflect a form of attentional , rather than visual , interference . Here we tested this idea by examining how set-size affects search-related neural activity in the lateral intraparietal area ( LIP ) , an area important for spatial attention and eye movements [9] . Monkeys performed a covert visual search task in which they discriminated the orientation of a target surrounded by variable numbers of distractors without shifting gaze to the target . As expected , LIP neurons responded more if the target appeared than if a distractor appeared in their receptive field ( RF ) , thus reliably encoding target location . However , firing rates associated with both the target and the distractors decreased with an increasing number of distractors ( set size ) , reflecting the operation of competitive visual interactions . The set-size-related decline in target responses correlated with performance accuracy and reaction time . The findings suggest that the set-size effect is explained , at least in part , by long-range competitive interactions that limit the strength of signals related to spatial attention .
Two monkeys performed a covert visual search task in which they discriminated the orientation of a visual target surrounded by a variable number of distractors ( Figure 1 ) . A trial began when monkeys shifted gaze to a fixation point located at the center of a stable circular array of 2 , 4 , or 6 figure-8 placeholders ( left panels ) . The array was positioned so that , when monkeys achieved central fixation , one placeholder ( the “RF stimulus” ) entered upon a constant location in the center of the RF of the recorded neuron . After a 500 ms delay two randomly selected line segments were removed from each placeholder , revealing a search display with 2 , 4 , or 6 unique shapes . One of the shapes , a right- or left-facing letter “E” appearing at an unpredictable location , was the search target while the others were distractors . Without breaking central fixation monkeys reported the orientation of the target by releasing grasp of a bar held in the right or in the left hand . We refer to the fixation epoch prior to presentation of the search display as the presearch epoch and to the interval starting with removal of line segments and ending with the bar release as the reaction time or search epoch . To examine the effect of set size , we used interleaved trial blocks in which the stable array contained 2 , 4 , or 6 elements ( Figure 1 , top , middle , and bottom rows ) . Increasing set size was associated with higher reaction times and lower accuracy ( Figure 2 ) . The set-size effect on reaction times , estimated using linear regression ( Materials and Methods section , Equation 1 ) , was significant in 70% of sessions with an average slope of 10 . 2 ± 1 . 1 ms/item ( Figure 2A; 13 . 2 ms/item for the significant subset; both p < 0 . 01 relative to 0 ) . Fitting the population data ( Figure 2B ) yielded a very similar slope of 10 . 6 ms/item ( confidence interval ( CI ) [5 . 8 , 15 . 5]; intercept , 425 ms; regression , p < 10−5; R2 = 0 . 11 ) . Compared with correct trials , error trials had higher reaction times but a comparable set-size effect ( intercept , 459 ms; p < 0 . 05 relative to correct trials; slope 14 . 7 ms/item; CI [10 . 2 , 17 . 8]; regression , p < 0 . 05; R2 = 0 . 09 ) . Fitting the accuracy values ( Figure 2C ) yielded a slope of −2 . 2%/item ( CI [–2 . 9 , −1 . 5]; intercept , 100 . 5; regression , p < 10−5; R2 = 0 . 33 ) . Thus , each additional distractor in the display caused an increase in reaction time of ∼10 ms and a decrease in accuracy of ∼2 . 2% . A distinguishing feature of our task is that it required covert attention and a nontargeting motor report ( a grasp release ) but precluded oriented movement of either eye or limb toward the search target . However , it was possible that even while they maintained central fixation monkeys attempted to shift gaze toward the target . To examine this possibility we measured average eye position in consecutive 100 ms time bins during the search period ( 0–400 ms after search onset ) as well as the end points of the first saccade made within 300 ms after the bar release ( when the search array remained on the screen but the fixation point was removed ) . All eye position measures were uniformly distributed relative to target location ( Rayleigh test for directedness of circular distributions , n = 1 , 710 , 3 , 312 , and 4 , 698 trials for set-sizes 2 , 4 , and 6; p > 0 . 6 for all measures ) . Thus we found no direct evidence that monkeys tended to shift gaze toward the search target during or after a trial . LIP neurons are known to encode target location during visual search , whether search is accompanied by saccades [10–12] or is performed covertly , as in the present study [13] . Accordingly , the neurons that we describe here had robust target location selectivity during the active phase of search ( Figure 3 ) . In addition , their firing rates declined as a function of set size . Figure 3A shows the responses of a representative neuron , and Figure 3B the average responses of the 50 neurons tested at all set sizes . Responses are segregated according to set size ( red , green , and blue traces ) and according to whether the target or a distractor was in the RF ( solid versus dashed traces ) . During the presearch epoch ( left panel , −200 to 0 ms ) the visual array was uniform , and the RF visual stimulation was constant across set sizes . Nevertheless , firing rates declined as set size increased from 2 to 4 to 6 . Once the placeholders changed shape ( time 0 in left panel ) , neurons showed a small transient response to the visual offset at ∼50 ms latency ( see also Text S1 , note 1 ) , followed by a robust signal of target location , whereby responses became much stronger if the target was in the RF than if a distractor was in the RF . Both target and distractor responses were lower at higher set sizes , but this neural set-size effect diminished by the time of the bar release ( right panels ) . To estimate the magnitude and time course of the set-size effect we fitted firing rates as a linear function of set size using linear regression ( Materials and Methods section , Equation 2 ) . We conducted this analysis separately for the presearch epoch ( 100 ms prior to search onset; Figure 4A and 4B ) and in consecutive time bins spanning the search epoch ( Figure 4C ) . Figure 4A shows the results of the presearch analysis for the example neuron in Figure 3A . Because target location was unpredictable , all trials regardless of target location were pooled for this analysis . The neuron showed a significant set-size effect with a slope of −6 . 2 spikes/s/item ( CI [–7 . 9 , −4 . 6]; regression , p < 10−11; R2 = 0 . 31 ) . Across the sample ( Figure 4B ) , 56% of neurons had slopes significantly smaller than 0 with an overall mean of −2 . 0 ± 0 . 46 spikes/s/item ( −4 . 0 spikes/s/item for the significant subset; p < 0 . 0001 relative to 0; n = 50 ) . Thus , neurons showed a decrease in firing rates of ∼2 spikes/s , on average , for each item added to the display . This effect was present from the beginning of fixation ( i . e . , from the time when the stable placeholder entered into the RF by virtue of the monkeys' eye movements ) ( Figure S1 ) . To follow the evolution of the set-size effect during the search epoch we repeated the regression analysis in consecutive 50 ms time bins , this time segregating trials according to whether a target or a distractor appeared in the RF . Figure 4C shows the average slope as a function of time , in data aligned on search onset ( left ) and bar release ( right ) , for target and distractor trials ( circles versus triangles ) . In the first 200 ms of search the set-size effect remained comparable to that in the presearch epoch ( p > 0 . 2 relative to presearch bins , for each time bin and trial type between 0 and 200 ms after search onset ) . However , the set-size effect declined markedly thereafter ( i . e . , slopes increased toward 0 ) , and the average slope became statistically indistinguishable from 0 ( open symbols ) by 250 ms after search onset for both target and distractor trials ( each p > 0 . 05 relative to 0 ) . When the data were aligned on bar release ( right ) a small residual set-size effect was seen for distractor but not for target trials ( all distractor slopes , p < 0 . 03; target slopes , p > 0 . 73 relative to 0 ) . However , no significant differences were found between target and distractor slopes in any time bin ( paired t-tests , p > 0 . 1 ) . A two-way analysis of variance ( ANOVA ) with bin and trial type ( target or distractor in RF ) as factors confirmed that there was a highly significant effect of time ( p < 10−10 ) but no effect of trial type or interaction between time bin and trial type ( p > 0 . 1 ) . The fraction of neurons showing significant slopes reached a peak of ∼55% ( 60% for the target , 50% for the distractors ) between 100 and 150 ms after search onset and dropped to 15% ( 12% for target , 18% for distractor ) in the last 50 ms before bar release . Thus , set-size effects were comparable whether a target or a distractor was in the RF and diminished gradually from the presearch epoch to the time of the bar release . A similar pattern was found in the larger subsets of neurons tested at only two of the three set sizes ( Figure S2 ) . Because LIP neurons strongly distinguish between a target and a distractor in the RF , it is important to determine how set size affected neuronal selectivity for target location . We measured target location selectivity using receiver operating characteristic ( ROC ) analysis , which estimates the probability that an ideal observer can determine whether a target or a distractor is in the RF based on the distribution of firing rates associated with each ( see Materials and Methods section ) . A ROC index of 0 . 5 indicates no selectivity , while indices above 0 . 5 indicate preference for the target over distractors in the RF . The finding that firing rate versus set size slopes were similar for target- and distractor-related responses suggests that increasing set size reduced firing rates uniformly and thus did not change the difference between target and distractor responses . Indeed , as shown in Figure 5 , both the time course and the asymptotic ( peak ) levels of the ROC values were unaffected by set size . The population ROC value ( center panel ) became significantly greater than 0 . 5 at a similar time across set sizes ( p < 0 . 05; n = 50; 110–120 , 90–100 , and 130–140 ms for set sizes 2 , 4 , and 6 ) . Likewise , the distributions of target discrimination times in individual neurons ( see Materials and Methods section ) showed no effect of set size nor significant differences between set sizes ( one-way ANOVA followed by multiple comparisons; median times of 160 , 150 , and 150 ms for set-sizes 2 , 4 , and 6 ) . The asymptotic ROC values ( measured between 200 and 300 ms after search onset ) were also not affected by set size ( one-way ANOVA , p > 0 . 1 ) . Thus , increasing set size reduced task-related firing rates but did not reflect the magnitude or time course of target–distractor selectivity . To examine the relationship between the LIP response and performance accuracy we analyzed responses on error trials in which monkeys released the wrong bar ( Figure 6 ) . Because of the relatively low error rates few neurons had a sufficient number of trials in all trial categories at each set size . Therefore we turned to a pairwise analysis in which we separately compared correct and error trials in subsets of neurons that contributed a sufficient number of error trials at set-size 6 and set-size 2 ( Figure 6A , n = 55 neurons ) and at set-size 6 and set-size 4 ( Figure 6B , n = 53 neurons ) . Although a significant effect of set size was present in both correct and error trials ( black asterisks indicate p < 0 . 05 in 100 ms time bins ) , selectivity for target location ( difference between solid and dashed traces ) was entirely absent on error trials ( colored asterisks indicate p < 0 . 05 in 100 ms time bins for the corresponding set size ) . As discussed in relation to Figure 2 , manual latencies in error trials were longer than those in correct trials ( for the present subsets of data , 2 versus 6 , intercept , 440 ms , slope , 13 . 8 ms/item; 4 versus 6 , intercept , 460 ms , slope , 18 ms/item; all p < 0 . 05 relative to correct trials ) . However , neuronal target location selectivity was absent up to the time of the bar release ( right panels ) , ruling out the possibility that discrimination may have occurred later on error trials , commensurate with the longer reaction times . These findings suggest that at least some errors reflected failures in target selection , which were associated with a lack of target location selectivity in the LIP . Two prior studies have reported that saccade reaction times correlated with the onset of significant neuronal discrimination between target and distractor in the RF [12 , 14] . However , as shown in Figure 5 , in our data neither the time of neuronal discrimination between target and distractors nor the asymptotic level of discrimination varied across set size despite clear effects on reaction time . Indeed , we found no significant correlation between the change in ROC onset times and the corresponding change in reaction time across set sizes ( 2 versus 4 , 2 versus 6 , and 4 versus 6; all r < 0 . 02; p > 0 . 3 ) . In contrast with the constancy in the ROC signal , however , we found that the firing rates associated with the target itself did reliably correlate with reaction time ( see also [15 , 16] ) . To examine this correlation within a set size , we separated trials into subgroups in which reaction time was shorter ( thick traces ) or longer ( thin traces ) than the median for each cell ( Figure 7A ) . Target-related responses had a stronger and faster rise in trials with short reaction times than those with long reaction times , while distractor-related activity showed a much smaller dependence on reaction time . We confirmed these observations by computing trial-by-trial correlations between firing rates and reaction time as a function of time during the trial . To compute the correlation across the population ( Figure 7B ) we first normalized firing rates and reaction times by subtracting the average in each neuron's dataset . When the target was in the RF population correlation coefficients became significantly negative ( indicating that higher firing rates were associated with shorter reaction times ) starting 100–200 ms after search onset . In contrast , distractor-related responses were largely uncorrelated with reaction time except for a trend toward a positive correlation , which reached significance only for set-size 2 late in the trial ( 200–300 ms ) . The middle and right panels show the distribution of coefficients for target and distractor responses in individual neurons ( computed without normalization ) between 200 and 300 ms after search onset . Coefficients for the target response were shifted toward negative values with medians of −0 . 16 , −0 . 24 , and −0 . 21 for set sizes 2 , 4 , and 6 ( all p < 10−5 relative to 0 ) . Coefficients for distractor responses tended to be positive but had smaller absolute values ( 0 . 13 , 0 . 09 , 0; p < 0 . 05 for set-sizes 2 and 4 ) . While these analyses included only trials in which the target was in or opposite the RF , we obtained similar results when we included all distractor trials . Thus , variability in the neural response to the target , but much less in that to the distractor , was correlated with variability in reaction time . To see to what extent variability in the target response could account for the set-size effect in reaction time , we fitted reaction times as linear functions of target-related firing rates including set size as covariant ( Figure 8; see Materials and Methods section ) . The analysis yielded slopes of −0 . 51 ms/spikes/s 100–200 ms after search onset , and −0 . 49 ms/spikes/s 200–300 ms after search onset , showing that reaction times increased by about 1 ms for each 2 spikes/s drop in neural response . Although modest , these slopes were highly significant ( each p < 10−20 relative to 0 ) . Correlation coefficients were comparable to those in Figure 7 ( 100–200 ms , −0 . 21 , −0 . 17 , and −0 . 15 for set-sizes 2 , 4 , and 6; 200–300 ms , −0 . 21 , −0 . 17 , and −0 . 15; all p < 0 . 05 ) . The analysis of covariance also showed that intercepts differed significantly across set sizes . Intercepts ( measured at 0 spikes/s ) at set sizes 2 , 4 , and 6 were −12 , 13 , and 37 ms for 100–200 ms , and −14 , 14 , and 35 ms for 200–300 ms ( each p < 0 . 05 for effect of set size ) . Thus , the analysis revealed two components of the behavioral set-size effect: one , captured by the intercept , was independent of LIP responses while a second , captured by the slope , was significantly correlated with LIP target-evoked responses . In our task visual stimuli were placed at relatively large distances ( medians of 15 . 0° and 10 . 7° at set sizes 4 and 6; see Materials and Methods section ) that exceeded the distances associated with masking or crowding and may also exceed the span of some of the neurons' RF . We wondered if set-size effects represented interactions only within a RF ( i . e . , if they arose only in neurons that had more than one stimulus in the RF ) or whether they represent interactions beyond the border of the classical RF . To address this question we examined whether set-size effects were related to the RF profile as estimated from the memory-saccade task ( see Materials and Methods section ) . Figure 9 shows the average normalized visual response on the memory-saccade task at the locations used in set size 4 ( Figure 9A ) and set size 6 ( Figure 9B ) , aligned to the center of each neuron's RF ( 0° ) . Because LIP RFs can be asymmetric we sorted the data so that the two locations flanking the RF center were grouped according to their relative response strength ( i . e . , flankers associated with the stronger and weaker of the two values were averaged separately ) . Normalized responses were calculated for each neuron by subtracting the baseline firing rate and dividing by the peak response ( always in the center of the RF , 0° ) . Finally , we segregated neurons according to whether firing rates in the 200 ms before search onset were significantly larger , smaller , or equivalent to those at set-size 2 ( set-size 4 , n = 11 , 34 , and 29 , respectively; set-size 6 , n = 9 , 31 , and 14 ) . If competitive interactions were limited to the span of the RF , then neurons with significant set-size effects should have significantly stronger excitatory responses at the stronger flanking location than neurons without a set-size effect . However , this was not the case . Average responses at 90° and 60° flanking locations ( set sizes 4 and 6 ) were not statistically different from baseline ( both p > 0 . 1 , one-way ANOVA ) , showing that for most neurons the nearest flanking stimuli fell outside the visual RF [17] . Moreover , response magnitude at flanking locations did not differ between neurons that did or did not have a set-size effect . We also found no consistent tendency for neurons to show inhibitory surrounds near the borders of the RF , as there was no significant dip below baseline at the weaker of the two flanking locations . Activity at the weaker locations was also not related to the set-size effect ( p > 0 . 1 for set-size effect , one-way ANOVA ) . These findings show that competitive effects in the LIP are not straightforwardly predicted by inhibitory surrounds near the excitatory RF and can extend beyond the confines of the classical RF . Previously we have shown that during a similar covert search task LIP responses were modified by the active limb , with some neurons having stronger responses to the target if the monkey released the right bar and others preferring left bar release [13] . In the present dataset , approximately one-third of neurons showed limb effects . However , we found that the magnitude and time course of the set-size effect as well as the correlation with reaction time were equivalent in neurons with and without limb sensitivity and , for the former group , were equivalent for responses with the preferred and nonpreferred limbs ( Figures S3 and S4 ) . Thus the set-size effect did not depend on limb selectivity . A second question is whether the set-size-related decline in activity may have been related to reward probability , which declined together with the monkeys' accuracy [18 , 19] . To examine this possibility we computed correlation coefficients between session-by-session firing rates and success rate ( Figure 10A ) . We found no correlations either within a set size ( Figure 10A; coefficients of −0 . 09 , 0 . 06 , and −0 . 08 for set sizes 2 , 4 , and 6; all p > 0 . 58 ) or in computing the differences across set sizes ( Figure 10B; set size 2 versus 4 , r = −0 . 12 , p = 0 . 06; set size 2 versus 6 , r = 0 . 08 , p = 0 . 55 ) . We also considered the possibility that monkeys estimated reward probability from local sequences of 10–20 trials rather than globally across long trial blocks [19] . However , correlation coefficients between local measures of firing rate and reward probability ( measured in sliding windows of 20 trials ) were statistically significant in only 2% of neurons , less than the 5% expected by chance . This precludes the possibility that the set-size effects were due to reward expectation . Because increasing set size also increases the uncertainty of target location , an important question is whether set-size effects reflected location uncertainty ( the diminished probability that the target would appear at any one location ) rather than the number of stimuli per se [19 , 20] . While previous studies reported probability effects in saccade-based tasks [19 , 20] , it is not clear whether monkeys use probability information to adjust the distribution of covert attention . To examine this question we trained monkeys in blocks of trials in which set size was constant ( n = 6 ) but target location probability varied between 100% ( the target appeared at a single , constant location ) and 16 . 7% ( the target appeared with equal probability at each location as in the standard condition ) . We ran these conditions for 20 sessions for each monkey using long blocks of , on average , 148 trials for each condition ( range , 96–358 trials ) . Despite extensive testing we found no significant differences in either reaction time or response accuracy between the constant and the variable conditions , either in individual sessions or in the pooled data ( for reaction time , p > 0 . 1 , n = 6 , 003 and 5 , 994 trials in the constant and variable conditions; for accuracy , p > 0 . 2 , n = 40 sessions ) . We also found no differences between reaction time and accuracy in the early and late portions of the blocks ( first 25% of trials , constant location , 464 ± 32 ms , 85 ± 6% correct; variable location , 459 ± 71 ms , 83 ± 6% correct; last 25% of trials , constant location , 455 ± 74 ms , 83 ± 5% correct; variable location , 467 ± 56 ms , 83 ± 6% correct ) . These results suggest that monkeys did not take note of changes in location probability even when these changes were very large ( 100% to 16 . 7% ) . This makes it unlikely that they detected the more subtle changes in probability among set sizes 2 , 4 , and 6 ( 50% versus 25% versus 16 . 7% ) . Figure 10C shows responses in a subset of 10 neurons tested at set-size 2 ( red traces , 50% probability ) and at set-size 6 in 100% and 16 . 7% probability conditions ( blue dashed and blue solid traces ) . Responses at set-size 2 were much higher than those at set-size 6 but did not differ between probability conditions within set-size 6 . A one-way ANOVA in each 100 ms time bin from 200 ms before to 300 ms after search onset revealed highly significant differences between set-size 2 and set-size 6 and either the 100% or the 16 . 7 % probability condition ( p < 10−6 in each bin ) but no significant effect of location probability within set-size 6 ( p > 0 . 2 in each time bin ) . Thus neural responses , like behavioral performance , were related to the number of display elements independent of variations in target location probability .
The LIP has been proposed to encode a priority representation of the visual world , a sparse topographic representation in which only objects that are likely to be attended are strongly represented . Two principal factors are known to activate LIP neurons—the automatic orienting of attention toward a salient but task-irrelevant stimulus [9] and the voluntary selection of a saccade target [11 , 12 , 14 , 15 , 22–24] . Bisley and Goldberg have shown that presaccadic sustained activity in LIP is related to the deployment of covert attention that precedes an overt saccade [22] . Here we go a step further in linking the LIP with covert attention independently of saccades: in our task neurons were strongly active even though monkeys were explicitly trained to withhold saccades throughout the task . It remains in principle possible that monkeys formed covert plans to make a saccade toward the attended target ( although we , like others [25] , failed to find direct behavioral evidence for this idea ) . However , the correlations between LIP activity and performance of the covert search itself strongly implicate this area in covert target selection independent of eye movements . This conclusion is supported by the findings of Wardak et al . that reversible inactivation of the LIP impairs performance on visual search tasks whether these are performed with free or fixed gaze [26 , 27] . While quantitative models have speculated on the contributions of the LIP to overt saccade decisions [28] , accounting for its contributions to covert attention is considerably more challenging . A common working hypothesis is that visual-oculomotor areas such as the LIP and the frontal eye field provide topographic , top-down feedback to feature-selective visual areas including V4 , the middle temporal area , and inferior temporal cortex , which boosts visual responses to the attended object [29–31] . The biased competition theory proposes that attentional feedback is especially important in environments containing multiple distractors , where feedback biases neuronal competition in favor of the attended object , allowing neurons representing this object to “win” the competition and filter out the effects of distractors [32] . Our findings are consistent with the idea that LIP plays a specific role in selecting targets and overcoming distractor interference in cluttered visual environments . In our task , as in saccade-based search tasks , neurons selectively encode the location of the search target , and their responses correlate with the efficiency of target selection [11 , 12 , 14] . In contrast , responses to nontargets has been shown to correlate with the distracting effects of these objects on performance [22 , 24 , 33] . Here we show a novel mechanism of distractor interference: adding distractors to the display suppresses LIP activity through competitive visual interactions , producing a neuronal set-size effect that correlates with the effect of set size on performance . Together with the finding that deficits following LIP inactivation are larger at higher set sizes [26] , the findings suggest that the LIP plays a special role in overcoming distractor interference in complex environments . In light of these considerations , the dissipation of the set-size effect during the reaction time may represent an active process through which the brain suppresses distractor interference . It may be argued that the decline of the set-size effect reflected mere disengagement of the LIP from the task toward the end of the reaction time , as at this time there was a general decline ( though not a complete disappearance ) in target location selectivity ( Figure 3 ) . This possibility is unlikely , however , as high location selectivity was not , in and of itself , necessary for seeing a set-size effect: robust effects were found in the presearch epoch and on error trials , when firing rates were low and there was no location selectivity . Thus , it is more likely that the dissipation of distractor effects reflected an active search-related process . One such process could be selection of the search target . Target-related responses peaked between 100 and 200 ms after search onset , slightly before the time when the set-size effect was filtered out at the population level ( 200–250 ms ) . It is therefore possible that the elevated target-related activity suppressed distractor competition , consistent with a biased competition model [8 , 32] . In addition , feedback about limb motor planning , which reaches the LIP [13] , may have helped render responses stereotyped and independent of set size [16] . Because the LIP receives strong input from extrastriate cortical areas , one must consider to what extent the competitive interactions that we report reflect properties of this bottom-up input . However , while competitive effects in extrastriate cortex are based on visual features , those in the LIP are based on spatial location . In areas V2 and V4 , the middle temporal area , and the middle superior temporal area , visual competition is triggered when two stimuli are presented in close proximity within an individual RF so that one stimulus has a preferred feature ( e . g . , orientation or motion direction ) while the other is nonpreferred [8 , 21 , 34] . These results support a model in which competition ( mutual inhibition ) arises between neurons with overlapping RFs but different preferred features [8] . In the LIP , in contrast , competitive effects arise among physically identical stimuli ( the placeholders during the search epoch ) and are triggered even when the competing stimuli are outside the classical RF . Thus competition in the LIP engages neurons that have nonoverlapping RFs but similar ( or no ) feature selectivity . These considerations suggest that our findings are more closely related with location-based competitive interactions in the superior colliculus and the frontal eye field [35 , 36] and reflect the internal organization of all three structures in topographic , nonfeature-selective representations . Thus visual clutter appears to affect multiple levels of representation through both space- and feature-based competition . Increasing set size in our task also increased the uncertainty about target location or , conversely , lowered the probability that the target appears at any given location . However , our data suggest that under the present conditions monkeys did not explicitly compute and represent location probability . In a control condition in which set size was kept constant but target location probability varied between 100% and 16 . 7% , we found no effect of location probability on either behavior or neural responses . We note , however , that some models represent the effect of distractors as a broadening of the underlying noise distribution ( i . e . , distractor-related firing rate distribution ) [37] . This type of increase in noise , which may be interpreted as an implicit representation of uncertainty , may indeed be an appropriate mathematical description of our data . The lack of a probability effect in our task may appear puzzling given prior reports that monkeys are sensitive to manipulations of location or reward probability in saccade-based tasks [18–20] . We suggest , however , that this result is explained by the precise task conditions that we used . Whereas previous studies manipulated the probability of an overt , rewarded saccade , here we used location probability to bias attention , a variable that is by definition covert and cannot be directly rewarded ( in our task , reward was linked to a manual response ) . In addition , the visual conditions in our task provided little incentive to guide attention to the likely target location based on probability information . Because the target was suprathreshold and was present until the manual response , monkeys could simply ignore the prior targets and find the target when it appeared with little loss of accuracy . Indeed , probability effects on covert attention were previously found in monkeys by using brief , near-threshold targets [38] but not in a detection task with suprathreshold stimuli [39] . Thus , our findings do not preclude the possibility that LIP neurons reflect location probability in conditions in which probability is computed and used; they show , however , that neurons are strongly affected by competitive visual interactions independently of target location probability . The target-related activity in the LIP correlated with both performance accuracy and reaction time . Target location selectivity was entirely absent in error trials , suggesting that many errors reflected failures in locating the target ( possibly along with failure in target discrimination or response selection ) . In addition , target-related firing rates showed trial-by-trial covariation with reaction time so that higher firing rates were associated with shorter reaction times both within and across set sizes . While the correlation coefficients that we find are modest ( −0 . 2 to −0 . 3 in Figures 7 and 8 ) , existing evidence suggests that such weak correlations are to be expected in cortical association areas . Because correlations between neural activity and motor output increase along the sensory–motor continuum [40] , an area such as the LIP , which represents a nonmotor processing stage , would a priori be expected to covary only weakly with trial-by-trial reaction time . The task that we used is complex and is likely to have engaged multiple areas in addition to the LIP , including extrastriate visual areas and areas related to limb motor planning , procedural memory , motivation , and reward evaluation , all of which have high trial-to-trial variability and weak interneuronal correlations [41–43] . In this regard it is remarkable that the correlations that we report are slightly larger than the average correlation coefficient of −0 . 09 reported in a saccade-based task [44] . Computational models show that reliable information may be extracted from ensembles of as few as 10–100 task-related neurons with highly variable , weakly correlated firing [42 , 45] , suggesting that our findings reflect significant contributions of the LIP to covert search . A puzzling aspect of our data is the finding that set size lowered firing rates but did not modify neuronal selectivity for target location ( the discrimination between target and distractors in the RF ) as indexed by the ROC analysis ( Figure 5 ) . We found that increasing set size reduced target- and distractor-related activity by similar amounts , leaving the dynamics of target location selectivity constant across set sizes ( Figures 2–5 ) . This appears to be at odds with two prior reports that found a consistent relationship between the time of onset of the ROC and saccade reaction time during visual search [12 , 14] . It should be noted , however , that variations in ROC dynamics in these studies may have been driven primarily by variations in target-related firing rates , as was the case in the within set-size analysis in Figure 7 ( see also [16 , 44] ) . Thus , while both the ROC signal and the target response itself can covary with reaction times , the latter may show a more consistent relation across task conditions . Resolving this question will require more detailed understanding how activity is read out from the entire LIP map under different task conditions .
Electrode penetrations were aimed at the posterior half of the lateral bank of the intraparietal sulcus as guided by structural MRI . Upon isolation each neuron was first tested with the memory-saccade task on which , after the monkey fixated a central point , a small target annulus ( 1° diameter ) was flashed for 100 ms . After a 1000–1250 ms delay the fixation point was extinguished , and monkeys were rewarded for making a saccade to the remembered location of the target within 100–500 ms . Neural responses were tested at 8–12 locations circularly distributed at a constant eccentricity around fixation , including the location estimated to be the center of the RF ( that eliciting the strongest visual response ) . The search task was conducted in randomly interleaved blocks of trials . An array containing 2 , 4 , or 6 figure-8 placeholders remained on the screen from the beginning to the end of a block , including intertrial intervals . A trial ( and data collection ) began with presentation of a fixation point ( a 0 . 5° red square ) at the center of the array . After monkeys grabbed two response bars and maintained central fixation for a 500 ms period , the search display was presented by removing two line segments from each placeholder . The line segments to be removed were selected randomly with the constraints that ( 1 ) a single item became the search target ( a right- or left-facing letter “E” ) , ( 2 ) each of the remaining shapes was continuously connected , and ( 3 ) no shape was presented at more than one location in one trial . The location and orientation of the target were selected randomly with uniform probability and independently of each other . Monkeys were rewarded for continuously maintaining fixation within a window of 2° × 2° and indicating the orientation of the target by releasing the right bar for the right-facing E or the left bar for the left-facing E within 100-1000 ms after removal of the line segments . A correct response was followed by removal of the fixation point and delivery of reward 250 ms later . Incorrect responses ( including fixation breaks and early , late , or inaccurate bar releases ) were followed by removal of the fixation point without reward delivery . The target array remained on the screen for an additional 500 ms , allowing us to collect eye movement data following bar release . All trials terminated with restoration of the missing line segments , reinstating the placeholder display . The radius of the placeholder array and its rotation around the center were varied for each training and recording session . During neural recording care was taken that one placeholder fell in the center of the neuron's RF as determined with the memory-guided saccade task . Stimuli were scaled with eccentricity and ranged from 1 . 5° to 3 . 0° in height and from 1 . 0° to 2 . 0° in width . Median RF eccentricity was 10 . 6° ( range , 2 . 6–14 . 3° ) . Median separations between adjacent array elements at set sizes 2 , 4 , and 6 were , respectively , 20° , 15 . 0° , and 10 . 7° ( ranges , 4 . 0–28 . 6° , 3 . 7–20 . 2° , and 4 . 6–14 . 3° , respectively ) . The ratios between interstimulus distance and eccentricity were 2 . 0 , 1 . 4 , and 1 . 0 at set sizes 2 , 4 , and 6 , far exceeding the critical ratio of 0 . 5 that defines the critical distance for crowding [1] . Comparisons across samples were made with t-tests or paired t-tests and one-way and two-way ANOVAs as specified below and evaluated at p < 0 . 05 . Regressions were calculated with weighted least-squares algorithms [46] . The 95% CI of the slope and intercept were calculated and used for statistical testing . In addition , we verified the results of the parametric tests using one-way nonparametric ANOVA ( the Kruskal–Wallis test ) and two-way nonparametric ANOVA ( the Friedman test ) . In all cases the results were equivalent , and , for simplicity , we adopted the convention of reporting only the outcomes of the parametric statistics in the text . In addition , we used ROC analysis , which is a nonparametric measure of the separation between two firing rate distributions and hence the likelihood that an ideal observer can distinguish between the two [47] . Results are shown as mean ± standard error unless otherwise stated . Data were collected from 107 neurons that had significant spatial selectivity during both memory-saccade and covert search tasks . However , the bulk of the analysis concentrates on 50 of those neurons ( 24 in monkey 1 ) that were tested at all three set sizes . Data from additional subsets tested at set-sizes 2 and 4 ( n = 73 ) , set-sizes 2 and 6 ( n = 55 ) , and set-sizes 4 and 6 ( n = 53 ) were used for the error trial analysis and in additional analyses shown in Text S1 . Approximately 8% of all trials were discarded from the analysis , as they terminated in fixation breaks or in short- or long-latency bar release . Reaction times were measured as the time from presentation of the search display ( measured by means of a light-sensitive diode mounted in the upper left corner of the screen ) to the time of the bar release ( measured by a transistor–transistor logic pulse emitted upon the onset or termination of contact with the bar ) . Accuracy was measured as the fraction of correct out of the total number of correct and incorrect bar releases . Reaction times ( RTs ) were analyzed separately for correct and erroneous responses . To assess sensitivity to set size , reaction times were fit with the regression model where SS is the set size ( 2 , 4 , or 6 ) and ε is random error distributed as multivariate normal ( Figure 2A and 2B ) . The slope b1 is an estimate of sensitivity to set size in milliseconds per item . This analysis was carried out on a neuron-by-neuron basis , where each data point represented one trial , and across the population , where each data point represents the average RT for a single session . Accuracy ( percent correct ) was fit using the population data . All analyses were conducted on raw ( unsmoothed ) spike counts . Firing rates on the memory-saccade task were measured in the baseline ( 200 ms before target presentation ) , visual ( 50–250 ms after target onset ) , delay ( 400–900 ms after target onset ) , and presaccadic epochs ( 200 ms before saccade onset ) . A neuron was tested on the search task only if it showed significant spatial tuning during the visual , delay , or presaccadic epochs ( p < 0 . 05 , one-way ANOVA ) . Nearly all neurons ( 96% ) had significant spatial tuning during the delay period . Visual responses on the memory-saccade task were highly correlated , across locations , with those during the search task ( r = 0 . 94 ) , showing that neurons preserved a constant spatial RF in both tasks . Several analyses were performed for the covert search task . To measure the sensitivity of firing rates to set size we fit firing rates to the linear model where FR is the trial-by-trial firing rate in the time bin indicated in the text . The coefficient b1 represents sensitivity to set size in units of spikes per second per item . Fits were obtained separately for trials in which the target or a distractor was in the RF . For the time-course analysis ( Figure 4B ) our choice of time bin ( 50 ms nonoverlapping windows ) represented the best compromise between the need for temporal resolution and the need to use larger bins for more reliable estimation of firing rates and regressions . To analyze the relationship between firing rates and reaction time , we first computed the Spearman correlation coefficient between firing rate on a trial-by-trial basis within each neuron ( Figure 7B ) . To compute the coefficient across the population we pooled all trials after normalizing each neuron's data by subtracting the average in the appropriate time bin ( Figure 7B , left panel ) . In a second step ( Figure 8 ) , we fit the data using analysis of covariance ( ANCOVA ) , which simultaneously fits separate regression lines of the form to data from set-sizes 2 , 4 , and 6 . The slope parameter b1 indicates the sensitivity in units of milliseconds per spike per second , whereas any difference in the intercept indicates the component of RT that depends on set size independently of LIP firing rates . The ANCOVA was computed on normalized data pooled across all neurons . Reaction times were normalized by subtracting the average reaction time across all three set sizes . Firing rates were normalized by subtracting the average neuronal response across all time bins , set sizes , and target/distractor trials . ROC indices comparing selectivity for target versus distractor in the RF were calculated for each neuron in 10 ms bins aligned on search display onset [47] . We found that ROC analysis was relatively robust to firing rate variations resulting from small bin sizes , allowing analysis with higher temporal resolution . Confidence intervals were obtained by a permutation test with 1 , 000 repetitions , and a value was deemed significant if its 95% confidence interval did not include 0 . 5 . The onset of significant selectivity was defined as the start of the first four consecutive bins with ROC values significantly different than 0 . 5 [13] . Each of the above analyses was evaluated for each monkey separately . In no instance did we find significant differences between monkeys , and thus the pooled data are presented throughout the paper . Data are also pooled across right and left bar release; analysis of data segregated according to bar release is included in Text S1 .
|
It is well known that the brain is limited in the amount of sensory information that it can process at any given time . During an everyday task such as finding an object in a cluttered environment ( known as visual search ) , observers take longer to find a target as the number of distractors increases . This well-known phenomenon implies that inputs from distractors interfere with the brain's ability to perceive the target at some stage or stages of neural processing . However , the loci and mechanisms of this interference are unknown . Visual information is processed in feature-selective areas that encode the physical properties of stimuli and in higher-order areas that convey information about behavioral significance and help direct attention to individual stimuli . Here we studied a higher-order parietal area related to attention and eye movements . We found that parietal neurons selectively track the location of a search target during a difficult visual search task . However , neuronal firing rates decreased as distractors were added to the display , and the decrease in the target-related response correlated with the set-size-related increase in reaction time . This suggests that distractors trigger competitive visuo-visual interactions that limit the brain's ability to find and focus on a task-relevant target .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neuroscience"
] |
2008
|
Neuronal Correlates of the Set-Size Effect in Monkey Lateral Intraparietal Area
|
Severe dengue disease is associated with high viral loads and overproduction of pro-inflammatory cytokines , suggesting impairment in the control of dengue virus ( DENV ) and the mechanisms that regulate cytokine production . Vitamin D3 has been described as an important modulator of immune responses to several pathogens . Interestingly , increasing evidence has associated vitamin D with decreased DENV infection and early disease recovery , yet the molecular mechanisms whereby vitamin D reduces DENV infection are not well understood . Macrophages represent important cell targets for DENV replication and consequently , they are key drivers of dengue disease . In this study we evaluated the effect of vitamin D3 on the differentiation of monocyte-derived macrophages ( MDM ) and their susceptibility and cytokine response to DENV . Our data demonstrate that MDM differentiated in the presence of vitamin D3 ( D3-MDM ) restrict DENV infection and moderate the classical inflammatory cytokine response . Mechanistically , vitamin D3-driven differentiation led to reduced surface expression of C-type lectins including the mannose receptor ( MR , CD206 ) that is known to act as primary receptor for DENV attachment on macrophages and to trigger of immune signaling . Consequently , DENV bound less efficiently to vitamin D3-differentiated macrophages , leading to lower infection . Interestingly , IL-4 enhanced infection was reduced in D3-MDM by restriction of MR expression . Moreover , we detected moderate secretion of TNF-α , IL-1β , and IL-10 in D3-MDM , likely due to less MR engagement during DENV infection . Our findings reveal a molecular mechanism by which vitamin D counteracts DENV infection and progression of severe disease , and indicates its potential relevance as a preventive or therapeutic candidate .
During the last decades , there has been an expansion in the geographic range and incidence of dengue virus ( DENV ) due to spreading of its mosquito vectors , globalization and the lack of a protective tetravalent dengue vaccine [1–3] . It is estimated that nearly a third of the world population is at risk of infection with an annual incidence of 96 million symptomatic cases and high economic burden in countries where active DENV transmission has been identified [4 , 5] . Infection with DENV may result in a self-limiting febrile illness known as dengue fever with or without warning signs that can progress to severe dengue . Disease severity is hallmarked by hemodynamic compromises that can lead to organ failure , hypovolemic shock and ultimately death [6] . While only a small percentage of cases evolve to severe dengue , progression and severity of dengue disease can differ depending on eco-epidemiology , host genetic factors , age , and virus virulence [7 , 8] . Additionally , complex interactions between the host immune response and the virus have been proposed as critical factors contributing to the pathogenesis of the disease [9–12] . In general , upon biting by a dengue-infected mosquito , dermal dendritic cells ( DC ) and macrophages are the main targets of DENV . In the skin , these cells host viral replication and facilitate further dissemination to peripheral tissues [13 , 14] , therefore becoming key drivers in the regulation of DENV-induced immune responses . To initiate infection , DENV has been shown to interact with C-type lectin receptors such as the mannose receptor ( MR ) , Dendritic Cell-Specific Intercellular adhesion molecule-3-Grabbing Non-integrin ( DC-SIGN ) and C-type lectin domain family 5 member A ( CLEC5A ) [15–18] . Although MR and DC-SIGN can bind DENV with high avidity facilitating attachment of the virus to the cell , in macrophages , only MR is thought to play a predominant role in virus binding and signaling [16 , 18] . Ligation of MR to DENV facilitates spatial interaction of the virus with a lower avidity receptor , CLEC5A [18] , which in turn initiates signaling pathways that aim at the secretion of cytokines that potentiate immediate local response and priming of the immune system [17 , 18] . DENV-induced activation of target cells is generally believed to induce excessive production of pro-inflammatory cytokines such as TNF-α and IL-1β that affect endothelial integrity and consequently enhance capillary permeability [9 , 10 , 19 , 20] . Furthermore , early in the infection , components of mosquito saliva and local tissue damage at the site of infection cause neutrophils , basophils and mast cells to elicit a Th2 cytokine response with predominant production of IL-4 [21] . Notably , the presence of IL-4 induces up-regulation of MR expression on dermal macrophages and recruits monocyte-derived macrophages ( MDM ) , thereby boosting further infection and pro-inflammatory events that may lead to disease progression [13 , 16 , 21] . Currently , no treatment for clinical improvement of dengue disease symptoms is available [3 , 22]; however , antiviral and immunomodulatory factors such as vitamin D may have the necessary potential . Apart from its classical role in maintaining calcium homeostasis , vitamin D3 is a potent modulator of the immune system [23] . Indeed , binding of the biologically active form of vitamin D , 1 , 25-dihidroxyvitamin D3 ( vitamin D3 ) , to the vitamin D receptor ( VDR ) allows the VDR to act as a transcription factor that modulates the gene expression of proteins involved in calcium absorption , cell proliferation and differentiation [24] . Vitamin D3 modulates the immune response to several pathogens including DENV [25–28] . In fact , epidemiological studies have associated genetic variants in the VDR with disease progression and vitamin D supplementation with early disease recovery [29–31] . In vitro , vitamin D3 treatment of myelo-monocytic cell lines reduces DENV infection and to modulates the cytokine response [32–34]; yet the underlying mechanism remains elusive [35] . Therefore , we here investigated the phenotypic features , susceptibility and innate responses to DENV infection of monocyte-derived macrophages differentiated in the presence of D3 ( D3-MDM ) .
Protocols for sample collection and written informed consent were approved by the Committee of bioethics Research of the Sede de Investigación Universitaria , Universidad de Antioquia ( Medellín–Colombia ) . The DENV-2 New Guinea C ( NGC ) strain was provided by the Center for Disease Control ( CDC , CO , USA ) and was propagated in C6/36 cells . Briefly , monolayers of C6/36 HT cells in 75-cm2 tissue culture flasks were inoculated with DENV at a MOI of 0 . 05 in 1 mL of L-15 medium supplemented with 2% Fetal Bovine Serum ( FBS ) . After 3 h , 10 mL of L15 medium containing 2% FBS were added and the cells were cultured for 5 days at 34°C without CO2 . The supernatants were obtained by centrifugation for 5 min at 1800 rpm to remove cellular debris and were stored at -70°C . Virus titration was performed by flow cytometry as described [36] . Briefly , C6/36 HT cells were seeded in 12-well plates and cultured overnight at 34°C without CO2 . The cells were infected with 10-fold serial dilutions of the virus and harvested at 24 h post-infection ( hpi ) . Indirect intracellular staining of DENV E protein with the monoclonal antibody 4G2 ( Millipore , Darmstadt , Germany ) and the secondary antibody goat anti-mouse IgG-FITC ( Invitrogen , Life Technologies , CA , USA ) was performed as described later below . The cells were analyzed by FACScanto flow cytometry using the FACSdiva software . The percentage of infected cells in each sample and the total number of cells seeded per well were used to calculate the final titer of the virus . Isolation of viral RNA from cell lysates and supernatants was performed according to manufacturer’s instructions using the RNeasy mini kit and the QIAamp Viral RNA Mini Kit ( Qiagen , Hilden , Germany ) , respectively . The number of genome equivalent copies ( GEc ) was determined by RT-qPCR using DENV-2 specific primers ( forward: 5’CAATATGCTGAAACGCGAGAGAAA 3’ , and reverse: 5’ CCCCATCTATTCAGAATCCCTGCT 3’ ) . The calculation of the GEc was performed based on a standard curve , as previously reported [37 , 38] . The mosquito C6/36 HT cell line was obtained from ATCC and cultured in Leibovitz L-15 medium ( L-15 ) supplemented with 10% v/v heat-inactivated FBS , 4 mM L-glutamine , and 10 units/ml penicillin/0 . 1 mg/ml streptomycin ( Sigma-Aldrich Chemical Co , MO , USA ) , at 34°C in an atmosphere without CO2 . This study was conducted according to the principles expressed in the declaration of Helsinki . All samples of venous peripheral blood were obtained in Medellin-Colombia from an equal proportion of healthy women and men that were not vaccinated against yellow fever virus , were seronegative for the DENV NS1 antigen and DENV IgM/IgG , and were between 20 and 33 years old . Peripheral blood from donors was mixed with EDTA 4% v/v and human peripheral blood mononuclear cells ( PBMCs ) were isolated by Ficoll-Histopaque ( Sigma-Aldrich ) gradient at 650 g during 30 min as described [39] . Platelet depletion was performed by washing with PBS ( Sigma-Aldrich ) three times at 250 g during 10 min . Monocytes were obtained from total PBMCs by plastic adherence as described [39] . Briefly , 5x105 CD14+ cells from total PBMCs were plated in 24 well plastic plates and were allowed for to adhere during 4 h in 1640 RPMI medium ( Sigma-Aldrich ) supplemented with 0 . 5% of heat inactivated human serum pool ( HSP ) at 37°C and 5% of CO2 . Non-adherent cells were removed by washing twice with PBS , and MDM differentiation was allowed for 144 h in RPMI medium 10% HSP at 37°C and 5% of CO2 . Additionally , MDMs were differentiated in presence of 1 , 25 di-hydroxyvitamin D3 to obtain D3-MDMs . For this , 1 , 25 di-hydroxyvitamin D3 ( Sigma-Aldrich ) was added to the culture media at a final concentration of 0 . 1 nM and replenished every 48 h until the final time point of differentiation was reached . Kinetics and concentration of the vitamin D dose was determined on the basis of cytotoxicity levels lower than 5% ( measured by the MTT and trypan blue exclusion assays , of transcriptional induction of vitamin D signaling targets ( VDR and CYP24A ) and of modulation of immune responses , as previously described [40 , 41] . For each experiment , two equal fractions of PBMCs were used from the same donor for MDM and D3-MDM differentiation . At the indicated experiments , after MDM and D3-MDM differentiation , the cells were stimulated with IL-4 ( 100 ng/mL ) ( PeProtech ) or mock-treated ( culture medium ) and incubated for an additional 48 h previous to DENV challenge . Expression of macrophage surface and intracellular molecules was evaluated by flow cytometry . Intracellular staining using anti-CD68-PE ( BD Pharmingen ) and surface staining using anti-CD14-FITC ( BD Pharmingen ) , CD83-PeCy-7 ( eBioscience ) and anti-CD206-PE ( BD Pharmingen ) were performed on MDMs and D3-MDMs . For each experiment , unstained and isotype controls were included . DENV infection was measured by intracellular detection of the E protein using the murine monoclonal 4G2 antibody ( Millipore , Darmstadt , Germany ) and a secondary antibody , the goat anti-mouse IgG-FITC ( Invitrogen ) . Unstained cells , mock-infected cells plus secondary antibody only , mock-infected plus detection pair and infected cells plus secondary antibody were included as controls for every experiment . The percentage of infected cells was expressed as number of 4G2 positive cells over the total number of cells analyzed . Surface and intracellular staining samples were read on a FACScanto flow cytometer ( BD Biosciences ) and data were analyzed using the FACSdiva ( BD Biosciences ) and Kaluza ( Beckman Coulter ) softwares . After differentiation , macrophage monolayers were washed with warm PBS and were infected with DENV at the indicated MOI in 300 uL of RPMI medium and 2% HSP per well . In some experiments , macrophage monolayers were pre-treated with α-Methyl D-Mannoside ( MM ) ( Sigma-Aldrich ) at 10 mM for 1 h before the virus was added to the cells . For these experiments , mock vehicle controls of cells pre-treated with the same volume of MM diluent alone were included . At 3 hpi , the cells were washed with warm PBS to remove unbound virus and were incubated at 37°C with 5% CO2 in RPMI medium and 10% PSH for another 21 h . At 24 hpi , macrophage monolayers were harvested and used for flow cytometry staining and analysis as mentioned above . Additionally , in some experiments cell lysates and supernatants were used to determine the GEc titer . Since low temperature can not only affect the dynamics of C-type lectin receptors and the fluidity of the cell membranes to facilitate endocytosis , but also the DENV conformational structure and attachment to target cells [42 , 43] , we conducted our binding experiments at 37°C , as previously described [44] . Macrophages were infected with DENV at a MOI of 10 for 1 h . Thereafter the cells were delicately washed twice with PBS to remove the unbound virus and lysed to determine the GEc titers as described above . Since rapid internalization may occur during this time , we report here percentage of bound and/or internalized GEc . The number of GEc titers detected in the inoculum served as 100% binding and/or internalization control . mRNA expression of some macrophage genes was measured by quantitative real-time PCR . For total RNA isolation , an RNeasy Mini Isolation Kit ( Qiagen , Valencia , CA , USA ) was used . For cDNA synthesis , the RevertAid Minus First Strand cDNA Synthesis Kit ( Thermo Scientific , Wilmington , DE , USA ) was used according to the manufacturer’s instructions . Quantitative real-time PCR reactions were performed using the following primers . For CD206 ( Mannose receptor ) , forward: 5’-CACGATCCGACCCTTCCTTG-3’ and reverse: 5’GCTTGCAGTATGTCTCCGCT-3’ . For CLEC5A forward: 5’-CTTCCAGGGAGAAAGAGGCCC-3’ and reverse: 5’-CTGGTGGTGGTGAAACCATCG-3’ . For CD209 ( DC-SIGN ) forward: 5’-GAGTTCTGGACACTGGGGGAG-3 and reverse: 5’-CAAGACACCCTGCTAAGCTCTTG-3’ . For CYP24A1 , forward 5'-ACCAGGGGAAGTGATGAAGC-3' and reverse 5'-GTACAAGTCTTCAACGTGGC-3' . For VDR , forward: 5’-TGCTATGACCTGTGAAGGCTG-3’ and reverse: 5’-AGTGGCGTCGGTTGTCCTT-3’ . For β2M , forward 5'-GAGTATGCCTGCCGTGTG-3’ and reverse 5’-AATCCAAATGCGGCATCT-3’ . The Bio-Rad CFX manager was used to obtain the cycle thresholds ( Ct ) that were determined in each sample using a regression fit in the linear phase of the PCR amplification curve . Duplicate assays were performed for each sample and relative transcript units ( RTU ) and fold-change values were calculated in relation to β2M ( define ) expression by using the ΔCt and the ΔΔCt method , respectively . Levels of TNF-α , IL-6 , IL-1β and IL-10 in cell supernatants were measured by ELISA ( BD Biosciences , San Jose , CA , USA ) according to the manufacturer’s instructions . Statistical comparisons were performed using the non-parametric Mann-Whitney test using the software GraphPad Prism 5 ( GraphPad Prism , CA , USA ) . In addition , the Wilcoxon signed-rank test was used to compare paired MDMs and D3-MDM data . The critical value for statistical significance used for the analysis in the present study was p<0 . 05 , denoted as * , p<0 . 01 denoted as ** , and p<0 . 001 denoted as *** .
First , we assessed the effect of D3-MDM differentiation on the main phenotypic features of MDMs . To this end , CD14+ monocytes were obtained from human PBMCs by plastic adherence as described [39] , and were cultured during 144 h in the absence ( MDM ) or presence of 0 . 1nM vitamin D3 ( D3-MDM ) . Cell yield , viability and morphology were not affected by the presence of vitamin D3 during MDM differentiation ( S1A Fig ) . Likewise , both MDMs and D3-MDMs were positive for CD68 and showed a similar low expression of CD83 ( S1B Fig ) . As expected , D3-MDM differentiation induced an increase in the mRNA levels of two vitamin D inducible target genes , the VDR and hydroxylase CYP24A [45] ( S1C Fig ) . This confirmed the functional induction of vitamin D3 signaling in D3-MDMs . Next , we compared the susceptibility of MDMs and D3-MDMs towards the DENV-2 NGC strain . Infection was evaluated 24 hpi by intracellular detection of the viral E glycoprotein using flow cytometry . In line with an earlier publication [34] , we observed that DENV-2 infection in MDMs depends on the MOI . At a MOI of 10 , 9 . 5% of the cells were infected ( S1D Fig and Fig 1A ) . Importantly , and in line with studies in immortal cells , infection of D3-MDMs at a MOI of 10 resulted in a significantly lower ( p = 0 . 01 ) percentage of DENV-positive cells ( ~4% ) ( Fig 1A and 1B ) . To verify these results , we measured the number of DENV ( GEc present in cell lysates and supernatants . In line with the infection data , we found that both , intracellular and secreted GEc titers were significantly lower ( p = 0 . 03 ) in D3-MDMs when compared to MDMs ( Fig 1C ) . Interestingly , whereas the percentage of infected cells was reduced only by 2 fold , the viral progeny release was reduced by 200 fold , suggesting that D3-MDM differentiation may contribute to viral restriction during early and late stages of the viral life cycle . To assess whether DENV infectivity in D3-MDMs is restricted at early stages of infection , we evaluated the number of bound and/or internalized DENV-2 particles in MDMs and D3-MDMs at 1 hpi . The number of bound and/or internalized DENV-2 GEc was measured by RT-qPCR as described in the Methods section . Fig 2 shows the percentage of bound and/or internalized DENV-2 GEc in relation to the number of added GEc titers . For MDMs , the percentage of bound and/or internalized DENV-2 GEc ranged from 15% to 80% depending on the blood donors . Notably , only for D3-MDMs , was the percentage of bound and/or internalized DENV-2 GEc comparable between the donors and significantly ( p = 0 . 03 ) lower in comparison to their MDM counterparts . This led us to propose that macrophage differentiation in the presence of vitamin D3 reduces expression of receptors required for DENV to gain entry into macrophages . Among several C-type lectins expressed on macrophages that can facilitate virus attachment , MR is thought to represent a key binding receptor [15 , 16 , 18 , 46] . Thus , we next tested whether the decreased binding in D3-MDMs was due to the limited accessibility of MR on these cells . To this end , surface expression of MR by FACS detection of CD206 in MDMs and D3-MDMs was evaluated . As shown in Fig 3A , both MDMs and D3-MDMs stained positive for CD206 in all 4 donors tested . Yet , the percentage of CD206 positive cells was significantly lower ( p = 0 . 04 ) in D3-MDMs ( ~36% ) when compared with MDMs ( ~43% ) for all donors ( Fig 2B ) . In addition , the mean fluorescence intensity ( MFI ) values of CD206 were lower in D3-MDMs ( ~750 ) as compared with those observed in MDMs ( ~1400 ) ( Fig 3A and 3B ) . Interestingly , we observed a positive correlation between the percentage of CD206 positive cells and the percentage of infection in MDMs and D3-MDMs ( p = 0 . 001 , r = 0 . 761 ) ( Fig 3C ) . Moreover , and in line with our flow cytometry data , we also found that the transcriptional activity of MR was significantly ( p = 0 . 01 ) lower in D3-MDMs than in MDMs ( S2A Fig ) , suggesting an overall reduction in MR availability . Likewise , the same pattern of transcriptional activity was observed for CLEC5A ( S2B Fig ) , however , for DC-SIGN , it was not detected in either MDMs or D3-MDMs ( S2C Fig ) . This is in line with previous reports where transcriptional and surface expression of DC-SIGN has not been detected in MDMs obtained under the same conditions as reported here [47] . Since MR is critical for the MR/CLEC5A complex function , the reduced susceptibility of D3-MDMs to DENV could be attributed to the lower expression of MR on the cell surface . To confirm this result , MR and other C-type lectins were blocked , reasoning that this treatment would have no effect on DENV infection in D3-MDMs . Accordingly , cells were pre-treated with α-Methyl-D-mannoside ( MM ) , a mannose binding site competitor and inhibitory sugar of C-type lectins [48] . As shown in Fig 3D , pre-treatment of MDMs with 10 mM MM reduced the intracellular DENV-2 GEc titers by approximately ~1 Log as compared with those observed in mock pre-treated control cells . Of note , the intracellular GEc titers in MDMs treated with MM were virtually identical to those found in D3-MDMs . Importantly , MM treatment had no effect on the GEc titers observed in D3-MDMs thereby confirming that low accessibility of C-type lectins indeed restricts infection of D3-MDMs . To substantiate the contribution of limited MR accessibility to reduced infection in D3-MDMs , we sought to up-regulate MR expression by treatment with IL-4 , a well-known inducer of MR expression in macrophages [16 , 46] . To this end , MDMs and D3-MDMs were treated with IL-4 ( 100 ng/mL ) for 48 h , as described [16] . Analysis of MR expression by flow cytometry showed that IL-4 treatment induced a significant ( p = 0 . 007 ) increase in the percentage of MR positive cells in MDMs but not in D3-MDMs ( Fig 4A and 4B ) . In addition , no differences were observed in the MFI values of CD206 positive cells . This indicates that D3-MDMs are refractory to the canonical effects of IL-4 on MR expression . In line with this , we observed that in MDMs , IL-4 treatment induced an 8-fold increase in the transcriptional activity of the MR gene when compared to mock-treated , whereas in IL-4-treated D3-MDMs , the fold-change induction was only 2-fold ( S2D Fig ) . Likewise , induction of CLEC5A transcriptional activity was lower in D3-MDMs when compared with MDMs ( S2E Fig ) , indicating that other IL-4-induced c-type lectins may also be less responsive to the effect of vitamin D . We then tested whether this limited IL-4-induced increase of MR expression in D3-MDMs was also linked to reduced susceptibility to DENV-2 . Thus , IL-4-treated MDMs and D3-MDMs were infected with DENV-2 at a MOI of 10 and 24 hpi the number of DENV GEc was determined by RT-qPCR in cell lysates ( Fig 4C ) . IL-4 treatment increased the susceptibility of both MDMs and D3-MDMs as compared with their control , the mock-treated cells . As expected , for the mock-treated control D3-MDMs , the number of DENV GEc titers were ~1 . 5 log lower as compared with mock-treated control MDMs . However , the number of intracellular DENV GEc observed in IL-4-treated D3-MDMs was significantly lower than that found in IL-4-treated MDMs ( p = 0 . 007 ) . In spite of this , there was no detectable increase in MR expression in D3-MDMs , IL-4 did increase the susceptibility to DENV infection . This rather surprising finding prompted us to investigate the role of MM in controlling DENV infectivity in IL-4-treated cells . To this end , the number of DENV GEc was determined in IL-4-treated cells at 24 hpi ( Fig 4D ) . MM treatment lowered GEc production by ~2 Log ( p = 0 . 01 ) and ~1 Log ( p = 0 . 01 ) in MDMs and D3-MDMs , respectively , as compared with mock-treated control cells . All these data suggest the importance of MR during DENV infection of MDMs and D3-MDMs and provide insights of in the mechanism by which vitamin D limits DENV infection in human macrophages . Besides the role of macrophages as target cells for DENV infection , they also play an important role in the production of pro-inflammatory cytokines that can enhance the pathogenesis of dengue disease [12] . Since vitamin D3 has been previously suggested to act as an important modulator of immune responses [41] , we next evaluated changes in cytokine response of MDMs and D3-MDMs following DENV infection . To this end , we measured by ELISA in culture supernatants , the production of several cytokines related to DENV pathogenesis , including TNF-α , IL-1β and IL-10 . As shown in Fig 5A , baseline levels ( mock-infected cells ) of the cytokines are similar between MDMs and D3-MDMs . However , upon DENV infection , the induced cytokine levels were significantly lower ( p<0 . 01 and p<0 . 05 ) in D3-MDMs as compared with MDMs ( Fig 5A ) . This observation could be a direct consequence of the lower infection or poor engagement of C-type lectin receptors , such as MR , due to the overall down-regulation of these molecules observed in D3-MDMs . DENV attachment to MR allows spatial interaction and cooperation of the MR/DENV complex with CLEC5A , that interacts with the virus , and triggers the DENV-induced pro-inflammatory signaling pathways [18] . Since we observed down-regulation of MR and CLEC5A in D3-MDMs , we hypothesized reduced participation of these receptors during DENV-induced cytokine response in these cells . Accordingly , we evaluated the DENV-induced cytokine response after blockade of C-type lectin ligation to DENV using MM . As shown in Fig 5B , pre-treatment of the cells with 10 nM MM had no effect on the levels of TNF-α , as compared with mock-treated control cells . The MM pretreatment also did not affect TLR4/CD14-mediated signaling indicating that incubation with MM does not interfere with the secretion of pro-inflammatory mediators induced by C-type lectin receptor-independent pathways . Of note , and in line with previous reports [32 , 41] , LPS did not induce TNF-α production in D3-MDMs , substantiating the reduction of inflammatory responses in D3-MDMs . Additionally , MM treatment itself did not alter the secretion of TNF-α and IL-1β in MDMs and D3-MDMs . Importantly , upon DENV infection , MM treatment reduced the secretion of TNF-α and IL-1β in MDMs but not in D3-MDMs . On the other hand , in D3-MDMs , these cytokine levels were significantly ( p<0 . 05 ) lower than in MDMs and showed the same levels as those found in MM-treated cells . Taken together , these data suggest that in MDMs and D3-MDMs , C-type lectin ligation to DENV triggers cytokine release and supports the hypothesis that down-regulation of these molecules can contribute in modulating the DENV-induced cytokine response in D3-MDMs .
Containment of DENV infection and controlled secretion of inflammatory cytokines by macrophages are crucial events needed to avoid progression of dengue disease [12 , 49] . Our findings show that differentiation in the presence of vitamin D3 restricts DENV infection in human MDMs by affecting DENV binding to cells . We found that MR is reduced in D3-MDMs and given the importance of this receptor for DENV attachment , we argue its accessibility as a limiting factor for less virus binding . Although we cannot rule out the participation of other C-type lectins such as CLEC5A , engagement of the MR receptor during DENV infection depends on the formation of the MR-DENV complex [18] , attributing to MR an essential role for binding and signaling . Indeed , we show that reduced expression of MR and likely of CLEC5A contributes to a lower secretion of DENV-induced TNF-α , IL-1β and IL-10 . The expression of MR is under the control of the pro-adipogenic peroxisome proliferator-activated receptor γ ( PPARγ ) [50] , a transcription factor that is responsible for the induction of MR expression via IL-4/IL-13 signaling and polarization of inflammatory responses in macrophages [51] . Interestingly , vitamin D activity was recently found to downregulate PPARγ expression and co-regulate PPARγ-induced transcriptional activity in macrophages [52 , 53] . This could explain the reduced expression of MR in D3-MDMs and a limited effect of IL-4 on these macrophages . Importantly , this may also have in vivo relevance in the context of dengue pathogenesis , as since inflammatory skin conditions caused by mosquito bites are accompanied by secretion of IL-4 that leads to enhancement of infection and cytokine response [16 , 46] . It is important to note , that the effects of vitamin D on MR expression reported here may contrast previous reports [54 , 55] , where vitamin D3 concentrations and target cells were different to from those used in the present study . Furthermore , we showed here that secretion of TNF-α , IL-1β and IL-10 was significantly lower in DENV-infected D3-MDMs than in MDMs . This observation is in line with a previous report in the human cell line U937 [56] and can be a direct consequence of either lower cell activation due to the decreased viral load and less availability of viral antigens , or to the immunomodulatory features of vitamin D3 . In the latter scenario , it is recognized that vitamin D can control cytokine responses by indirect modulation of NF-κB activity or by direct regulation of VDR-dependent genes [32 , 57–59] . Both mechanisms may contribute to variations in the expression of several PRRs that are important to trigger cytokine responses . Indeed , our data shows that MDM differentiation in the presence of vitamin D modulates several DENV-relevant PRRs , such as MR and CLEC5A . MR orchestrates the MR/DENV/CLEC5A complex functions to provide an essential connection between binding and triggering of downstream signaling pathways that aim at cytokine secretion [18] . Since we found diminished MR expression in D3-MDMs , it is likely that a lower participation of the MR/CLEC5A complex during DENV infection can occur . Certainly , blockade of MR ligation to DENV in MDMs decreased secretion of TNF-α and IL-1β to the same levels as observed in D3-MDMs , showing the effect of MR availability in the induction of cytokine responses during DENV infection . This study provides a mechanism underlying the resistant phenotype of human MDMs differentiated in the presence of vitamin D3 to DENV infection in vitro . Interestingly , we also demonstrated that the anti-DENV effect provided by vitamin D3 was retained after IL-4 stimulation . Since this cytokine can enhance C-type lectin receptor-mediated-DENV infection and the cytokine response [46 , 60] , our observations indicate a potential in vivo role of vitamin D3 in down-tuning the immune response during infection by DENV . Interestingly , susceptibility to severe dengue disease has been associated with variations on in VDR and Fc receptors [30] , that are crucial for antibody-dependent enhancement of infection during secondary encounters with the virus . However , future and critical studies are required to assess the clinical importance of our findings and the potential role of vitamin D as a preventive or therapeutic target to treat disease severity . Interestingly , several reports have already anticipated VDR genetic variants with clinical outcomes of dengue disease and oral vitamin D supplementation with disease recovery and moderate inflammation [29–31] .
|
Dengue represents a major worldwide concern for public health . Clinical complications rely on vascular leak of fluids and molecules from the bloodstream that leads to a potentially fatal hemodynamic compromise . Disease progression has been related to poor control of dengue virus ( DENV ) dissemination and excessive production of pro-inflammatory mediators that affect the endothelial function . Vitamin D has been shown to modulate immune responses and to alleviate dengue disease . Here , we studied how addition of vitamin D during macrophage differentiation modulates the functional features of these cells in the context of DENV infection . We observed that vitamin D reduced susceptibility of these cells to DENV infection and down-regulated the virus-induced cytokine response . This phenotype was attributed to downregulation of MR , a molecule hijacked by the virus to gain entry into the cells and a key receptor of the MR/CLEC5A complex that links binding and immune activation during DENV infection . Our study sheds light on the mechanism by which vitamin D can restrict DENV dissemination and the cytokine response in macrophages , indicating the potential relevance of this hormone as a preventive and therapeutic candidate .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"dengue",
"virus",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"cytokines",
"pathology",
"and",
"laboratory",
"medicine",
"chemical",
"compounds",
"pathogens",
"immunology",
"microbiology",
"organic",
"compounds",
"cell",
"differentiation",
"viruses",
"developmental",
"biology",
"rna",
"viruses",
"molecular",
"development",
"white",
"blood",
"cells",
"animal",
"cells",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"chemistry",
"vitamins",
"immune",
"response",
"immune",
"system",
"biochemistry",
"lectins",
"cell",
"biology",
"flaviviruses",
"organic",
"chemistry",
"physiology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"macrophages",
"vitamin",
"d",
"organisms"
] |
2017
|
Human macrophages differentiated in the presence of vitamin D3 restrict dengue virus infection and innate responses by downregulating mannose receptor expression
|
Most studies evaluating epidemiologic relationships between helminths and HIV have been conducted in the pre-ART era , and evidence of the impact of helminth infections on HIV disease progression remains conflicting . Less is known about helminth infection and clinical outcomes in HIV-infected adults receiving antiretroviral therapy ( ART ) . We sampled HIV-infected adults for eight gastrointestinal parasites and correlated parasitic infection with demographic predictors , and clinical and immunologic outcomes . Contrasting with previous studies , we measured parasitic infection with a quantitative , highly sensitive and specific polymerase chain reaction ( PCR ) method . This cohort study enrolled HIV-infected Ugandans from August-September 2013 in Mbale , Uganda and collected stool and blood samples at enrollment . Real-time PCR quantified stool: Ascaris lumbricoides , Ancylostoma duodenale , Necator americanus , Strongyloides stercoralis , Trichuris trichiura , Cryptosporidium spp . , Entamoeba histolytica , and Giardia intestinalis infection . Generalized linear models assessed relationships between parasitic infection and clinical or demographic data . 35% of participants ( 71/202 ) tested positive for ≥1 helminth , mainly N . americanus ( 55/199 , 28% ) , and 4 . 5% ( 9/202 ) were infected with ≥2 stool parasites . Participants with hookworm infection had lower average CD4+ cell counts ( -94 cells/mcL , 95%CI: -141 , -48 cells/mcL; p<0 . 001 ) after adjustment for sex , CD4+ nadir at clinic entry , and time on ART . The high prevalence of parasitic infection and correlation with decreased CD4+ concentrations highlight the need to re-examine the effects of invasive helminth co-infection in rural , HIV-infected populations in the era of widely available ART . Elucidating the relationship between hookworm infection and immune recovery could provide opportunities for health optimization , e . g . integrated deworming , in these vulnerable populations .
Five soil transmitted helminth species Ascaris lumbricoides , Trichuris trichiura , hookworm species Necator americanus and Ancylostoma duodenale , and Strongyloides stercoralis infect over a billion people worldwide . [1 , 2] The burden of parasitic infection is greatest in low-income areas , particularly in certain areas of sub-Saharan Africa , where human immunodeficiency virus ( HIV ) is also highly prevalent . Studies of African adults living with HIV have shown helminth co-infection rates that range from 10% to upwards of 45% . [3–9] To date , the majority of research investigating the impact of intestinal helminth infection on HIV disease progression has occurred prior to widely available antiretroviral therapy ( ART ) . [3 , 4 , 9–13] The effect of helminth and HIV co-infection in the presence of ART is less well characterized . Indeed , to our knowledge , only two studies to date have examined the impact of deworming on CD4+ recovery in persons receiving ART . [14 , 15] The current literature examining the relationship between soil transmitted helminth infections and HIV in the pre-ART era presents an inconclusive picture . The large body of observational data is mixed . Two observational cohort studies found no beneficial effect of deworming on HIV viral loads and CD4+ T-cell concentrations , [4 , 5] while another suggested the possibility of a protective effect of helminths on decreasing HIV viral replication . [3] Of the three randomized experiments evaluating the impact of deworming on markers of HIV disease progression without ART , two found an improvement in either CD4+ T-cell concentrations or HIV viral load after anthelmintic therapy . [9 , 13] Another larger , reflexive randomized deworming trial failed to show a statistically significant benefit of empiric deworming treatment versus standard of care in preventing HIV progression to either a CD4+ count of <350 cells/mcL , first reported use of antiretroviral treatment , or death due to a non-traumatic cause ( 44 . 0 versus 49 . 8 events per 100 person-years; hazard ratio = 0 . 88 , 95%CI 0 . 74 to 1 . 04 , P = 0 . 10 ) . [10] However , it is possible that there was a less extreme benefit to presumptive therapy , which they were underpowered to detect . Interactions between soil-transmitted helminths and their human hosts are complex , and helminth infection may influence a patient’s relationship with other pathogens . A recent review discusses not only the links between selected parasites and HIV susceptibility and disease progression , but also the relationship between soil transmitted helminths and the potential for increased susceptibility to malaria and tuberculosis . [16–18] It is also important to recognize that soil transmitted helminths , through their potent and systemic T helper cell type 2 ( Th2 ) cytokine and regulatory responses , [19] may induce Th2 protective effects that could benefit long-term HIV survivors , e . g . protection against conditions associated with chronic inflammation . [20–23] However , this same Th2 immune response may mediate increased susceptibility for Th1-related infections . [24] At present , integrated presumptive anthelmintic therapy in the context of HIV care is neither recommended by Ugandan National Guidelines , [25] nor is it recommended by WHO . [26] While WHO does recommend periodic so called “preventive chemotherapy” for high risk groups , including women of child bearing age and adults with occupational exposures , [27] these guidelines have generally not been integrated into any type of standard care , nor has particular emphasis been given to HIV-infection . Given the frequency and consistency with which HIV-diagnosed persons interact with their care providers , the integration of adult deworming programs into HIV care may be a logical conclusion . [28] However , given the dearth of high quality and adequately powered species-specific studies , dramatic increases in ART availability , and incomplete understanding of biological mechanisms that are impacted by helminths during HIV infection implies that research questions focused on soil transmitted helminths and HIV have not been exhausted . Our current study evaluated the prevalence and burden of the five most common soil-transmitted helminths and three protozoal species in adults living with HIV enrolled in outpatient HIV care in peri-urban Uganda . We also evaluated the relationship between helminth infection and clinical and immunologic outcomes , and examined risk factors for helminth infection in this population .
Written informed consent was provided by all participants . The University of Minnesota , The AIDS Support Organisation ( TASO ) , and the Uganda National Council of Science and Technology institutional review boards approved this protocol . From August through September 2013 , we screened HIV-infected adults engaged in outpatient care at the TASO HIV clinic in Mbale , Uganda , during their normal clinical visit for a one-time stool sample analysis , and longitudinal follow-up via chart review . This study was powered to estimate overall parasitic infection prevalence among patients with a recent CD4+ T cell count <500 cells/mcL , a population of approximately 600 . We estimated a sample size of 210 , based on a true population prevalence of 30% , an alpha level of 0 . 05 , 5% precision estimate , and a finite population size of 600 . Inclusion criteria were age ≥18 and most recent CD4+ count <500 cells/mcL . We excluded persons who reported or had a record of taking albendazole or other anthelmintics in the past three months , and persons with known albendazole allergy . Pregnant women were also excluded due to potential albendazole teratogenicity . Participant data were collected via participant interview and chart review . We collected data on age , sex , weight , village of residence , and occupation . We also collected data on date of HIV diagnosis , date of enrollment into HIV care , World Health Organization ( WHO ) clinical stage at clinic enrollment , CD4+ at enrollment into clinical care ( “nadir CD4+” ) , 12-month history of opportunistic infections , and ART history ( regimen , duration ) through review of medical records by a medical officer . Participants underwent a physical examination at study enrollment for assessment of current WHO clinical stage , weight , and presence of current opportunistic infections . Finally , we collected follow-up CD4+ T helper cell values that were gathered as part of TASO’s routine clinical practice in the 24 months since study enrollment . Study follow-up occurred in a passive fashion , and no attempts were made beyond standard clinical practice to return patients to care if they stopped attending clinic . We collected blood and stool from participants during their study visit , which was also a participant’s normal clinic visit . We performed a single blood draw to evaluate CD4+ T cell count via the FACSCalibur flow cytometer ( BD Biosciences , San Jose , CA ) per routine TASO laboratory protocol . Parasitic infection status was only evaluated at one time-point: study enrollment . Participants provided a single stool sample , which we froze without fixatives on site at -20°C within 1–2 hours of collection . Stool specimens were transported on a weekly basis to Kampala , Uganda for long-term -80°C storage during enrollment . At the Translational Research Laboratory of the Infectious Diseases Institute in Kampala , Uganda , we used a modified version of a validated quantitative PCR described previously in Mejia et al . to assess participant stool for Ascaris lumbricoides , Ancylostoma duodenale , Necator americanus , Strongyloides stercoralis , Trichuris trichiura , Cryptosporidium spp . , Entamoeba histolytica , and Giardia intestinalis infection . [29] This PCR assay was modified to increase the total volume of each reaction from 7μL to 10μL to accommodate the minimum settings on the Applied Biosystems 7900HT Fast Real-Time PCR System . Reagent concentrations of the 10μL reaction matched those of the 7μL reaction concentrations . [29]DNA was extracted from approximately 50mg of stool via the FastDNA SPIN Kit for Soil ( MP Biomedicals , Santa Ana , CA ) using a low reagent method developed by Mejia et al . for resource-limited contexts , which has been included as S1 DNA Extraction Protocol . An additional step was required to extract T . trichiura DNA , whereby the remaining insoluble pellet from one DNA extraction was re-suspended in 200μL DNA-free water , heated at 90°C for 10 minutes , and centrifuged at 14 , 000g for 10 minutes . We then repeated the above-described DNA extraction method to process the resulting soluble portion of the sample . Sequences for the species-specific primers and probes and methods for the qPCR analysis are found in Mejia et al . [29] All control standards were tested in triplicate , and all unknown samples were tested in duplicate . A PCR cycle threshold ( Ct ) value >38 was considered a negative result . To help ensure that false positives were not driving our results , we conducted a post hoc experiment to bind the N . americanus primers and probes to the pBR322 internal control plasmid . [30] We did not observe any evidence of binding between the N . americanus primers or probes and the pBR322 control plasmid . Parasite burden quantification was performed by interpolating against parasite specific sequences standards and reported as DNA fg/μl . [29 , 31] Briefly , egg counts were estimated from McMaster microscopy techniques of subjects infected with N . americanus and/or A . duodenale and compared directly to qPCR results . Estimated egg counts from qPCR were calculated using Yova/g feces = 0 . 03472*Xfg/μl per correlation studies . [31] Similar calculations were used to estimate Trichuris trichiura egg counts: Yova/g feces = ( 1 . 095 x 10−5 ) *Xfg/μl , which was derived by comparing qPCR to Kato-Katz results in infected individuals . [29] Parasite infection prevalence was estimated overall , by species , and by species type ( protozoa or nematoda ) . Infection intensity was summarized by species for helminth worms . [32 , 33] Statistical analyses focused on hookworm infection a posteriori , due to its unique immunologic and clinical features , and overwhelming prevalence relative to other species of helminths . We used generalized linear models with a binomial distribution and log link , and a robust covariance estimator , to estimate associations between parasitic infection ( overall helminth infection , hookworm infection only , and protozoa infection ) and clinical and demographic characteristics , specifically occupation ( farming as primary profession versus any other ) , sex , age ( scaled to 5-year increments ) , weight ( scaled to 5-kg increments ) , WHO Clinical Stage ( 3 or 4 versus 1 or 2 ) , ART status ( receiving or not receiving ) . We also estimated the association between parasitic infection and CD4+ T cells/mcL at study enrollment , and the potential effect of parasitic infection on over CD4+ T cell concentrations over follow-up . Age- , and sex-adjusted linear regression models estimated the mean difference in CD4+ T cells/mcL at study enrollment by parasitic infection status ( any protozoa , any helminth , hookworm only ) . Restricted maximum likelihood linear mixed models , which included participant-specific random intercepts , and an identity covariance matrix , evaluated change in CD4+ T cell concentrations over time across hookworm infection status among participants who were ART-initiated at baseline . These longitudinal models were adjusted for sex , age , time on ART , and weight at baseline . Additional exploratory sub-analyses of change in CD4+ T cell concentrations by hookworm infection status were performed among 1 ) participants who had initiated ART <1 year before enrollment , and 2 ) participants who had initiated ART for ≥1 year before enrollment . Time on ART , weight , sex , and age were a priori included as covariates given their relationship with either the outcome , or to control for potential confounding , e . g . age . We attempted to evaluate the relationship between CD4+ T cell count and parasite burden ( light , moderate , and heavy intensity infections per WHO classification ) . However , because all infections were classified as light intensity ( <2 , 000 eggs/gram feces ) , we were unable to create any clinically meaningful exposures beyond presence or absence of hookworm infection . No imputation was performed for missing data , which occurred in <2% of participants . All analyses were performed in Stata/IC 13 . 1 ( StataCorp , College Station , Texas ) and results were evaluated against an alpha level of 0 . 05 .
Multi-parallel quantitative PCR results indicated that 35 . 2% ( 71/202 ) of participants were infected with at least one species of helminth or protozoa . Of these 71 participants , 10 were infected with two species . Most parasitic infections were caused by N . americanus ( 27 . 6% , 55/199 ) . Giardia had the next highest prevalence ( 6 . 1% , 12/197 ) , followed by Strongyloides ( 4 . 0% , 8/202 ) . Prevalence and infection intensity of parasitic organisms are described in Table 2 . Calculated egg burdens for N . americanus infections had a median of 0 . 72 eggs per gram of stool ( IQR: 0 . 53 , 6 . 34; maximum: 275 ) and 527 eggs/gram of stool for the single Ancylostoma duodenale infection . These are considered light infections by the World Health Organization . [34] An estimated 52 , 694 eggs/gram of stool was calculated for the single heavy Trichuris trichiura infection . Strongyloides stercoralis eggs generally hatch and mostly larvae are seen in stool samples , there are no current categories for intensity of larvae in infected patients . Results of generalized linear models analyses indicated that each 5-year increase in age was inversely related with a composite outcome of either Giardia , Cryptosporidium , or E . histolytica infection ( Prevalence Ratio ( RR ) = 0 . 71 , 95%CI: 0 . 55 , 0 . 92 , p = 0 . 01 ) ; 11 . 5% ( 6/52 ) of participants <30 years of age , 6 . 7% ( 6/90 ) of participants 31 to 40 years of age , and 1 . 7% ( 1/60 ) of participants ≥40 years of age were infected with ≥1 protozoal species . Protozoal infection was more prevalent in farmers than other occupations , although this relationship was unstable and not statistically significant in an age- and sex-adjusted model ( PR = 3 . 96; 95%CI: 0 . 89 , 17 . 60; p = 0 . 07 ) . Other factors–sex , CD4+ count at enrollment , ART status–were not associated with protozoa infection . ( See Table 3 ) There were no statistically significant associations between demographic and clinical characteristics and any helminth infection , i . e . either A . lumbricoides , A . duodenale , N . americanus , S . stercoralis , or T . trichiura from univariable analyses . Adjusting sex and/or age in multivariable models did not change these results . There were no statistically significant relationships between age , sex , occupation and ART status and prevalent helminth infection . Helminth infection was slightly more prevalent in women ( PR = 1 . 71; 95%CI: 1 . 00 , 2 . 91 ) and less prevalent in people currently receiving ART ( PR = 0 . 66; 95%CI: 0 . 41 , 1 . 07 ) , but neither were statistically significant . ( Table 3 . ) Prevalent infection with hookworm species A . duodenale or N . americanus was positively associated with age , and female sex in univariable analyses . Current receipt of ART was inversely associated with prevalent hookworm infection . These relationships were exaggerated in multivariable models that adjusted for sex and/or age . Increasing age ( PR5-years = 1 . 20; 95%CI: 1 . 01 , 1 . 42; p = 0 . 04 ) was associated with hookworm infection . Women were more likely to be infected with hookworm ( PR = 1 . 82; 95%CI: 1 . 10 , 3 . 24; p = 0 . 04 ) , even after adjustment for age , versus male participants . Participants receiving ART were less likely to have prevalent hookworm infection ( PR = 0 . 57; 95%CI: 0 . 36 , 0 . 93; p = 0 . 02 ) ; 52 . 6% ( 10/19 ) of participants not receiving ART were infected with hookworm , and 25 . 7% ( 46/179 ) of participants receiving ART were infected with hookworm . We assessed the relationship between hookworm infection and CD4+ T helper cell concentrations at study enrollment . Participants with hookworm infection demonstrated consistently lower concentrations of CD4+ cells/mcL when compared to hookworm-uninfected peers ( Table 4 ) . Unadjusted analyses indicated an average difference of -70 cells/mcL ( 95%CI: -113 , -26 , p = 0 . 002 ) in participants with hookworm infected relative to those without detectable hookworm infection . This relationship became more pronounced when adjusting for participant age , sex , and time on ART; participants with hookworm infection had 94 fewer CD4+ cells/mcL on average ( 95%CI -133 , -55 , p = <0 . 001 ) than those without hookworm . Stratified analyses on ART status ( receiving or not currently receiving ART ) indicate a similar relationship among those persons receiving ART at enrollment ( n = 171 ) ( mean: -102 cells/mcL; 95%CI -145 , -58; p = <001 ) . An additional stratified analysis among those persons who were ART naïve was limited by a small sample size ( n = 19 ) , but did not show a statistically significant relationship between hookworm infection and CD4+ T cell concentrations ( mean: -43 cells/mcL; 95%CI: -118 , 32; p = 0 . 24 ) . Among participants who had initiated ART at enrollment , results from the longitudinal analyses demonstrate that participants with hookworm infection and participants without hookworm infection had a similar rate of CD4+ T cell immune recovery in the 24-months post-enrollment ( βhookworm-time = 0 . 55; 95%CI: -0 . 35 , 1 . 45; hookworm-time interaction term p-value = 0 . 23 ) . Participants with hookworm did , however , have consistently lower CD4+ concentrations relative to their hookworm-uninfected peers over the 24 months of follow-up ( -87 cells/mcL; 95%CI -151 , -22; p = 0 . 009 ) , based on an average of 2 . 4 CD4+ measurements ( min = 1 , max = 5 ) ( Fig 1 ) . The mean number of measurements over time across hookworm-infected versus uninfected participants was similar ( 2 . 2 and 2 . 4 , respectively ) . At 12 months post-enrollment , predicted means from sex- , age- , time on ART- and baseline weight-adjusted linear mixed models estimated that participants with hookworm infection had an average of 361 CD4+ cells/mcL ( 95%CI 310 , 412 ) versus an average of 419 cells/mcL ( 95%CI 391 , 448 ) amongst those without hookworm infection at baseline . Participants with hookworm infection at baseline had , on average , 58 fewer CD4+ cells/mcL ( 95%CI: -117 , 1 ) relative to their uninfected peers at 12-months of follow-up . At 24 months post-enrollment , participants with hookworm had predicted mean 438 CD4+ cells/mcL ( 95%CI 365 , 514 ) versus mean 469 cells/mcL ( 95%CI 429 , 509 ) among those uninfected with hookworm at baseline . While participants with hookworm infection still had lower CD4+ cell concentrations than their uninfected peers at 24 months of follow-up , the difference in average CD4+ cells/mcL between hookworm infected versus uninfected across groups was attenuated ( -30 cells/mcL; 95%CI -114 , 55 ) and not statistically significant . Furthermore , participants who had initiated ART ≥1 year prior to study enrollment ( n = 108 ) demonstrated a similar relationship to our overall cohort; there was no difference in rate of CD4+ cell recovery during the study period , but those with hookworm were at an immunologic deficit relative to their uninfected peers ( 77 fewer CD4+ cells/mcL in those with hookworm versus those without; 95%CI -154 , -1 ) . Among persons who had initiated ART <1 year prior to study enrollment ( n = 68 ) , the effect was less pronounced , with only 26 ( 95%CI: -121 , 70 ) fewer CD4 cells/mcL in those with hookworm versus those without , and no difference in change over time , like other analyses .
Our results are comparable with much of the available literature regarding parasite infection prevalence in adults . Other studies conducted in Uganda have found similar prevalences of hookworm infection ( 24 to 52% ) , [35–39] and Strongyloides ( 4 to 8% ) [40–42] in adults with and without HIV . In persons with HIV in Nigeria , Senegal , and Ethiopia , Giardia prevalence has been observed at approximately 5% . [43–45] Partial immunity to most parasitic infections is acquired over the life course , leading to an increased rate of parasite destruction and worm expulsion with increasing age and re-infection . Hookworm species , however , do not induce the same adaptive immunity in humans as the other soil transmitted helminths , and consequently , may continue to infect adults with high frequency and intensity . [46] In the context of frequent and repeated infection , this lack of adaptive immunity may have important implications for host response to co-infections like HIV , and important Th1-moderated HIV co-morbidities , such as tuberculosis and cryptococcal meningitis . Our results found that participants who were infected with hookworm were at a significant CD4+ T-helper cells/mcL deficit , relative to participants who were not infected with hookworm . CD4+ T-helper cells are critical in mediation of the immune system’s response to various pathogens , and commonly used to monitor HIV disease progression and response to ART . [47] The inverse relationship between hookworm infection and CD4+ T cell concentrations was qualitatively and statistically consistent across various analyses , from unadjusted to adjusted regression , analyses restricted to persons receiving ART , and over time . We did not observe a difference in CD4+ cells/mcL among persons who were ART-naïve; however , the small proportion of persons not receiving ART in this cohort ( n = 19 ) renders these analyses relatively uninformative . Ample evidence demonstrates that soil-transmitted helminths are potent immunomodulators , and infection with soil-transmitted helminths involves many major body systems , from the gastrointestinal and circulatory systems , to soft tissues . [19] Multiple biologic mechanisms could be driving our observed relationship; and these results are likely multifactorial for any given participant . Hookworm infection in HIV-uninfected persons with celiac disease has been shown to decrease expression of interferon ( IFN ) -γ on intestinal T cells , and increase in CD4+FoxP3+ regulatory T cells , which could contribute to decreased differentiation to CD4+ T helper cells . [48] Other research has demonstrated that hookworm antigens induce cytotoxic and pro-apoptotic activity in Jurkat T Cells , contributing to an increase in CD4+ , CD8+ , and CD19+ lymphocytes that were in an early and/or late stage of programmed cell death . [49] Cuellar et al . found that commonly excreted hookworm protein Ac-TMP-1 , a Tissue Inhibitor of Metalloproteases , induced murine splenic T cells to differentiate to CD4+ and CD8+CD25+FoxP3+ regulatory T cells that expressed interleukin ( IL- ) 10 and suppressed naïve and activated CD4+ T cells differentiation . [50] Other human studies , however , have not found similar increases in T regulatory responses to hookworm infection . [51] Other human studies have not found differences in CD4+ T cell concentrations between hookworm-infected and -uninfected groups of HIV-uninfected participants . In a quasi-experimental study by George et al . , which measured the impact of deworming on microbial translocation ( a contributor to chronic immune activation linked to decreased concentrations of CD4+ T helper cells ) , observed that hookworm was associated with elevated levels of pro-inflammatory markers , e . g . lipopolysaccharide , soluble CD14 . [52] They did not , however , observe differences in T cell subsets among naturally infected , HIV-uninfected participants at baseline . [52] The authors postulate that lack of difference in T cell subsets is mediated by a counterbalancing , anti-inflammatory effect of hookworm infection , e . g . elevated levels of IL-10 , and decreased C-reactive protein , IL-17 and haptoglobin . [52] From the standpoint of clinical endpoints , results from clinical trials conducted in ART-naïve persons remain mixed . Results from the HEAT trial , which evaluated the impact of reflexive and repeated deworming on a patient’s risk for ART eligibility , i . e . a drop below 350 CD4+ cells/mcL , found no difference between the reflexive deworming group ( 400 mg albendazole every 3 months plus 25 mg/kg praziquantel annually ) versus the standard of care group ( no empiric deworming ) . [10] The trial had 80% power to detect a hazard ratio of 0 . 775 , which could be considered a large , albeit clinically important , difference in treatment groups . That said , actual CD4+ T cell concentrations at study completion were very similar across randomization groups , supporting the idea that deworming may not dampen CD4+ decline in the absence of ART . Other studies support this conclusion . [4 , 5 , 53] However , still other studies and meta-analyses demonstrated reductions in plasma viral loads and increases in CD4+ T helper cells with deworming in persons living with HIV . [13 , 54] The differences in these results may in part be explained by differences in methodology , and in particular the need to pool species due to limited species-specific sample sizes . Repeated and long-term exposure to hookworm and other helminth species may cause fibrosis of the gut associated lymphatic tissues ( GALT ) . IL-13 , in particular , is increased in the presence of hookworm infection[46] and a dominant mediator of fibrotic tissue , which induces fibrosis independently and via simulation and activation of transforming growth factor ( TGF ) -β . In the case of chronic and repeated helminth infections , and corresponding Th2 type immune responses , IL-13 production can become pathological . Fibrosis of the GALT has been linked to the dysregulation of immune cells , including CD4+ T cells , and impaired CD4+ recovery . [55 , 56] Chronic and repeated exposure to helminths and subsequent GALT fibrosis may have impacted the results of this and other studies . Indeed , the longitudinal analyses in this study demonstrate that hookworm-infected versus uninfected participants have a similar rate of CD4+ recovery over the 24-month follow-up period , but that those infected with hookworm remained at a significant immunologic deficit relative to their uninfected peers over time . Fibrosis is not reversible with deworming or other therapy . This bears mentioning because while hookworm and other helminths are still causally implicated in the decrease in CD4+ T cell concentrations , there are important implications for public health intervention design , e . g . increased deworming frequency targeting all stages of the human life course . It is possible that our results are spurious , either due to confounding , a misunderstanding of the directionality of the hookworm-CD4+ relationship , or a type I statistical error . The primary limitations of this study arise from its observational nature . First , the temporal relationship between parasite infection and immune status remains undetermined; it is conceivable that being immunocompromised would increase the likelihood of persistent infection . Research on this topic remains mixed and parasite dependent . [57–60] However , most research to date suggests no difference in hookworm risk between immunocompromised and immunocompetent persons . [61–64] There is potentially one exception to this pattern . A cross-sectional study by Sanyaolu et al . found that 4 . 6% ( 3/65 ) of HIV-infected Nigerians had hookworm infection , versus 1 . 8% ( 18/1015 ) of HIV-uninfected peers . [65] However , we were unable to duplicate their results based on the data provided in the paper . Our results are based on a single stool sample , and we did not use any concentration techniques prior to DNA extraction . Diagnostic sensitivity for hookworm–and other species–when evaluating a single stool sample is lower than sensitivity when using multiple stool samples . For example , Knopp et al . found that a single stool sample yielded a 7 . 1% prevalence , while 2 samples yielded a prevalence of 15 . 6% via Kato-Katz . [66] These authors found that Strongyloides prevalence with 1 stool sample versus 2 samples was similar , 3 . 5% and 5 . 3% , respectively . It is unlikely that our observed hookworm prevalence would have doubled had we analyzed >1 stool sample . However , we may have misclassified parasite-infected participants as uninfected , particularly among those with a low burden . Finally , our results may be confounded by data that would have been useful in these analyses but were not available . Hookworm , and other intestinal parasites , are considered diseases of poverty . The relationship between increased infection incidence among economically disadvantaged persons is well established . [38 , 67–69] Additionally , being economically disadvantaged could have impacted health outcomes in this study , e . g . CD4+ T helper cells concentrations , as it has in other research . [70–72] While we collected information on place of residence , this information could ultimately only be dichotomized into participants who lived in Mbale town , where the clinic is located , versus all others , which could represent varied levels of development and corresponding hookworm exposure . Also , ART adherence data were routinely collected and reflected uniformly high adherence levels . Past research on adherence at TASO ART clinics report similarly high levels of ART adherence , with ~90% of patients reporting no missed pills in the past 30 days . [73–75] However , data from prior TASO adherence research–like adherence data for this study–are limited by the fact that they are self-reported , which consistently over-reports adherence relative to pill counts , pharmacy refill information , and/or drug concentrations in blood . [76 , 77]
Despite the limitations of this study , we feel that these results are generalizable to other adults receiving outpatient HIV therapy in low-income , peri-urban areas . The results presented herein point to a high prevalence of helminth infection in this vulnerable population , and that hookworm infection is associated with sub-optimal health outcomes , i . e . lower CD4 . Therefore , further examination of these questions via a randomized trial is warranted , especially how systematic deworming may impact the immune status of this vulnerable population in the presence of ART .
|
Intestinal parasites , diseases of poverty that infect low-income populations and decrease school attendance and earning potential , infect more than 1 billion people worldwide . Current international guidelines focus deworming campaigns on high-risk populations of preschool- and school-aged children , and women of childbearing age . Intestinal parasitic infections also overlap significantly with human immunodeficiency virus ( HIV ) , which is highly concentrated in Sub-Saharan Africa; and some research indicates that certain intestinal parasitic infections have contributed to the spread of HIV in Sub-Saharan Africa . While there is a significant body of research examining the intersection between intestinal parasites and HIV disease progression , most of this work has occurred before HIV therapy was widely available in the region . This is the first study , to our knowledge , that focuses on a mixed sex , HIV treatment-initiated adult population , and directly measures intestinal parasitic infection via highly sensitive and specific molecular techniques . The results of this study indicate that adults who are living with HIV and parasitic hookworm infection are at an immunologic disadvantage when compared to adults who are not infected with hookworm . Our results also suggest that integrating deworming medications into HIV care could be an effective way to maximize the health status of this vulnerable population .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Conclusions"
] |
[
"blood",
"cells",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"helminths",
"antiviral",
"therapy",
"pathogens",
"immunology",
"hookworms",
"microbiology",
"parasitic",
"diseases",
"animals",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"preventive",
"medicine",
"rna",
"viruses",
"antiretroviral",
"therapy",
"vaccination",
"and",
"immunization",
"public",
"and",
"occupational",
"health",
"white",
"blood",
"cells",
"animal",
"cells",
"medical",
"microbiology",
"hiv",
"microbial",
"pathogens",
"t",
"cells",
"protozoan",
"infections",
"helminth",
"infections",
"cell",
"biology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"lentivirus",
"organisms"
] |
2017
|
Hookworm infection is associated with decreased CD4+ T cell counts in HIV-infected adult Ugandans
|
Dengue virus ( DENV ) causes a spectrum of diseases ranging from self-limiting dengue fever to severe conditions such as haemorrhagic fever and dengue shock syndrome . Antibody-dependent enhancement ( ADE ) is thought to explain the occurrence of severe dengue whereby pre-existing binding but non-neutralising antibodies enhance DENV infection . The ADE phenomenon is supported by epidemiological findings that infants that born to dengue immune mothers are at greater risk to develop severe dengue upon primary infection . The role of maternally acquired dengue-specific antibodies in disease enhancement was recently recapitulated in a mouse model where mice born to DENV1-immune mothers experienced enhanced disease severity upon DENV2 infection . Here , this study investigates the relative contribution of maternal dengue-specific antibodies acquired during gestation and breastfeeding in dengue disease . Using a surrogate breastfeeding mother experimental approach , we showed that majority of the maternal dengue-specific antibodies were acquired during breastfeeding and conferred an extended enhancement window . On the other hand , in the context of homologous infection , breastfeeding conferred protection . Furthermore , measurement of dengue-specific antibody titres over time in mice born to dengue immune mothers revealed a biphasic pattern of antibody decay as reported in humans . Our work provides evidence of the potential contribution of breast milk-acquired dengue-specific IgG antibodies in enhancement and protection against dengue . Should such contribution be established in humans as well , it may have important implications for the development of guidelines to dengue-immune breastfeeding mothers .
Dengue is a mosquito-borne viral disease responsible for an estimated 390 million annual dengue infections in the tropical and sub-tropical regions [1] . While most infected individuals manifest as asymptomatic or self-limiting dengue fever ( DF ) , a significant proportion progresses to more severe conditions—dengue haemorrhagic fever ( DHF ) and dengue shock syndrome ( DSS ) –characterised by symptoms such as vascular leakage and haemorrhage , which could be fatal [2 , 3] . The lack of effective vaccine and treatment against the potentially life-threatening dengue poses a serious public health concern . Dengue virus ( DENV ) , the etiological agent responsible for dengue , consists of four distinct serotypes ( 2 ) . Infection with one serotype confers long-term protection against the same serotype but only short-term protection against the other serotypes [4] . On the other hand , antibodies generated during primary infection may cause enhancement of dengue disease , a phenomenon coined as antibody-dependent enhancement ( ADE ) [3 , 5–8] . ADE develops due to the presence of pre-existing sub-neutralising antibodies that opsonise but do not effectively neutralise the virus . This results in the binding and endocytosis of virus-antibody immune complexes to Fcγ receptors ( FcγR ) -bearing cells such as monocytes and macrophages . However , instead of being degraded within the endosome , the virus escapes and replicates within the cells , thereby making the FcγR-mediated virus entry an efficient way to produce virus progeny . Furthermore , the antibody-mediated internalisation of DENV was shown to suppress innate antiviral responses , which further enhanced viral production [9] . This ADE hypothesis may explain the occurrence of DHF/DSS in secondary heterotypic infections as well as in primary infections of infants who passively acquired homologous or heterologous maternal antibodies [5 , 7] . Whereas the role in disease severity of other immune cells such as T cells during a secondary heterotypic infection remains a matter of debate , such possibility is clearly excluded in primary infections of infants born to dengue immune mothers , which exclusively relies on the maternal antibodies . This scenario was recently reproduced in two mouse models whereby young mice born to DENV1-immune mothers experienced enhancement of disease severity upon DENV2 infection [10 , 11] . Transfer of maternal antibodies transplacentally and via breastfeeding has been known to help protect infants against pathogens during early life [12 , 13] . Infants born to mothers immunised during pregnancy were protected against respective infections [14–16] , indicating the role of transplacentally acquired pathogen-specific IgG antibodies in protection . On the other hand , it has been well established that breastfeeding provides IgA-mediated mucosal immunity and was shown to protect infants against pathogen-associated diarrhoea [17–20] . While protection afforded by transplacentally acquired IgG antibodies has been well studied , information on the role of IgG antibodies acquired from breastfeeding is limited [21] . Neonatal Fc receptor ( FcRn ) mediates maternal IgG transfer across the placenta and small intestines [22–25] , allowing transfer into the circulation of IgG antibodies during gestation and breastfeeding respectively . Here , using A129 mice ( type I interferon-deficient ) , we studied the relative contribution of dengue-specific IgG antibodies that were acquired transplacentally or through breastfeeding in mediating dengue disease enhancement and protection . We report that in this mouse model breastfeeding represents the main route of maternal IgG transfer and that dengue antibodies present in breast milk play a critical role in enhancement or protection against dengue infection .
All the animal experiments were carried out in accordance with the guidelines of the National Advisory Committee for laboratory Animal Research ( NACLAR ) . Animal facilities are licensed by the regulatory body Agri-Food and Veterinary Authority of Singapore ( AVA ) . The described animal experiments were approved by the Institutional Animal Care and Use Committee ( IACUC ) from National University of Singapore ( NUS ) under the protocol number R14-97 . DENV1 ( Dengue 1 05K3903DK1 ) was isolated during dengue outbreak in Singapore in 2005 . DENV2 ( Dengue 2 D2Y98P-PP1 ) was derived from a clinical strain isolated in Singapore in 2000 . All DENV stocks were generated in C6/36 Aedes albopictus cell line ( American Type Culture Collection ( ATCC ) #CRL-1660 ) , maintained in Leibovitz’s L-15 medium ( GIBCO ) supplemented with 2% fetal bovine serum ( FBS ) as previously described [10] . Viruses were stored at -80°C . Virus titres were determined via plaque assay in BHK-21 cells ( See section below ) . Concentrated virus stocks were also prepared by precipitating viruses with 14% ( w/v ) polyethylene glycol ( PEG ) and subsequently used for IgG antibody quantification . Plaque assay was performed using BHK-21 cells ( ATCC #CCL-10 ) as previously described ( 10 ) . BHK-21 cells were cultured in 24-well plates ( NUNC ) . Cells monolayers reaching 80% confluency were infected with serially diluted viral suspensions . After 1 hour incubation at 37°C , 1% ( w/v ) carboxymethyl cellulose-containing RPMI-1640 medium supplemented with 2% FBS was added . After incubation for 4 ( DENV2 ) or 5 ( DENV1 ) days , cells were fixed with 4% paraformaldehyde and stained with 1% crystal violet . Plaques were counted and expressed as the number of plaque forming unit per mililiter ( PFU/mL ) . Adult A129 females were infected with 106 PFU per mouse of DENV1 or DENV2 via intravenous ( iv ) route . One week post-infection after virus clearance , the females were mated with naïve adult males . Age-matched naïve females were also mated concurrently . At birth , pups born to either dengue-immune or naïve mothers were switched and nursed by naïve or dengue-immune mothers respectively ( Fig 1 ) . Control groups comprised of mice nursed by birth mothers were also included . All pups were breastfed by the respective mothers and were weaned out 21 days later . At 5 and 10 weeks of age , these mice were challenged with either 106 PFU ( sub-lethal dose ) or 107 PFU ( lethal dose ) of DENV2 via iv route and monitored for survival and bled at specific time point for viremia determination . The infected animals were monitored daily for clinical manifestations . The scoring system used was: 0: Healthy; 1: Ruffled fur; 2: Hunched back; 3: Diarrhoea; 4: Lethargic; 5: Moribund . Survival rate was derived from the number of mice that were euthanized at moribund stage as evidenced by severe diarrhoea and extreme lethargy as described previously [11] . Serum collection was performed at various time points after birth for antibody measurement , or at day 4 post-infection to measure virus titres by plaque assay . Levels of systemic IgG and IgA antibodies specific to DENV1 or DENV2 were quantified via indirect enzyme-linked immunosorbent assay ( ELISA ) . UV-inactivated viruses ( 150 ng ) were coated onto 96-well ELISA plates ( Corning costar ) and incubated at 4°C overnight . Serially diluted serum samples were added to wells and incubated at 37°C for 1 hour . HRP-labelled goat anti-mouse IgG ( H+L ) ( Bio-rad ) or anti-mouse IgA ( Thermo-Scientific ) was subsequently added and incubated at 37°C for 1 hour . Detection was done with the addition of o-phenylenediamine dihydrochloride substrate SigmaFast ( Sigma Aldrich ) . The reaction was stopped upon addition of 1 M H2SO4 . Absorbance was read at 490 nm , and titre was calculated by nonlinear regression as the reciprocal of the highest serum dilution with absorbance corresponding to 3 times the blank absorbance . Heat-inactivated serum samples were serially diluted with RPMI-1640 medium supplemented with 2% FBS containing approximately 500 PFU/mL DENV1 or DENV2 . After 1 hour incubation at 37°C , plaque assay was performed in BHK cells . The percentage neutralization at each dilution was calculated based on the reduction in plaque number compared to that of the positive control . PRNT50 was determined by nonlinear regression as the reciprocal of the highest serum dilution that gave 50% reduction in number of plaques . Data were analysed using Mann-Whitney statistical test using Graphpad Prism . Difference were considered significant ( * ) at p value <0 . 05 .
In a previous work , we showed that 5-week old A129 mice born to DENV1-immune mothers displayed enhancement of disease severity upon heterologous DENV2 infection [11] . To investigate the relative contribution of dengue-specific IgG antibodies acquired during gestation and breastfeeding in disease enhancement , mice born to DENV1-immune were switched at birth and nursed by surrogate breastfeeding dengue naïve mothers ( Fig 1 ) . Similarly , mice born to naïve dams were nursed by DENV1-immune mothers . Thus , DENV1 specific-IgG antibodies circulating in mice born to DENV1-immune mothers but nursed by naïve mother ( DN ) were only acquired during gestation . On the other hand , DENV1-IgG antibodies circulating in mice born to naïve mother but nursed by DENV1-immune mother ( ND ) were only acquired during breastfeeding . Control groups included mice born to and nursed by DENV1-immune mothers ( DD ) , and mice born to and nursed by naïve mothers ( NN ) . Measurement of DENV1-specific IgG levels by ELISA showed that the systemic level of DENV1 IgG antibodies in DN mice ( acquired maternal antibodies from gestation only ) was significantly lower than that measured in DD mice ( acquired antibodies from both gestation and breastfeeding ) ( Fig 2A ) . Instead , mice which acquired antibodies from breastfeeding only ( ND ) displayed DENV1 IgG levels that were comparable to those measured in DD control mice , suggesting that breastfeeding is the major route for maternal antibody transfer ( Fig 2A ) . ELISA results using DENV2 as coating antigen to assess levels of cross-reactive IgG antibodies revealed similar trends ( Fig 2B ) . Measurement of DENV1-specific IgG subclass indicated that IgG2a is the main subclass of DENV1-specific IgG acquired by pups born to and/or nursed by DENV1-immune mothers ( S1 Fig ) . Furthermore , measurement of neutralising activity ( PRNT50 ) of sera showed similar trend as that of ELISA where mice from DD and ND groups had similar neutralising titres against DENV1 and DENV2 , whereas DN mice displayed lower titres ( Table 1 ) . Together , these data support that in this mouse model , DENV1-specific maternal IgG antibodies are mainly acquired from breastfeeding . Since secretory IgA antibodies represent the main class of antibodies in breast milk [26] , the level of DENV1-specific IgA antibodies was measured in the sera from mice born to and/or nursed by DENV1-immune mothers . ELISA results showed that the level of DENV1-specific IgA antibodies was below the detection limit in 5-week old mice born to and/or nursed by DENV1-immune mothers ( Fig 2C ) . Measurement of pooled sera samples from DENV1-immune mothers also revealed levels of DENV1-specific IgA near the detection limit ( Fig 2D ) . Thus , in this mouse model of ADE mediated by maternal antibodies , DENV1-sepcific IgA antibodies , if any , are unlikely to play a role in influencing disease outcome . To assess the contribution in disease severity of maternal antibodies acquired during gestation or breastfeeding , 5-week old mice from DD , DN , ND and NN groups as described above , were challenged with a sub-lethal dose of DENV2 . As previously observed [11] , mice born to and nursed by DENV1-immune mothers ( DD ) displayed enhanced disease severity compared to control mice born to and nursed by their naïve mothers ( NN ) whereby DD mice were moribund by day 4 post-infection while majority of NN mice remained healthy throughout the observation period ( Fig 3A and 3B ) . Interestingly , both ND mice ( born to naïve mothers but nursed by DENV1-immune mothers ) and DN mice ( born to DENV1-immune mothers but nursed by naïve mothers ) displayed disease kinetic and severity similar to that observed with DD mice ( Fig 3A and 3B ) . Furthermore , measurement of viral loads at day 4 post-infection in serum samples from DENV2-infected DD , DN and ND groups all showed increased viremia titres compared to NN controls ( Fig 3C ) . Together , these data demonstrate the enhancing potential of maternal DENV1-IgG antibodies acquired during either gestation or breastfeeding . They also indicate that the lower level of anti-DENV1 IgG antibodies measured in DN mice ( acquired during gestation only ) was sufficient to enhance DENV2 infection . Maternal antibodies in infants are catabolised with time . Dengue disease enhancement contributed by maternally acquired antibodies is thus likely to be lost over time when these antibodies decrease to non-enhancing levels [8] . Here , using the same switching strategy ( Fig 1 ) , we monitored the decay of maternal anti-DENV1 IgG antibodies acquired during gestation , breastfeeding or both . Results showed that the level of DENV1-specific IgG antibodies measured at birth in mice born to DENV1-immune mothers was comparable to the level measured in 3-week old pups born to and nursed by DENV1-immune mothers ( DD ) as well as in mice born to naïve dams and nursed by DENV1-immune mothers ( ND ) ( Fig 4A ) . In contrast , antibody titres in 3-week old mice born to DENV1-immune dams but nursed by naïve mothers ( DN ) dropped by >1log ( Fig 4A and Table 2 ) . This observation thus indicated that the antibody titer measured at 3 weeks of age is mainly contributed by breastfeeding . Upon weaning at 3 weeks of age , antibody decay in all the groups was monitored over time and the antibody half-life was determined . In mice born to and/or nursed on DENV1 immune dams ( DD and ND groups ) , the antibody titers dropped by one log between week 3 and week 8 leading to a half-life of 11 . 5 and 10 . 5 days , respectively ( Fig 4A and Table 2 ) . Between week 8 and week 16 , the antibody decay in these groups was slower with antibody half-life of 16 . 5 and 14 . 5 days , respectively . These observations thus indicated a biphasic decay pattern of the maternal DENV1 IgG antibodies in pups born to and/or nursed on DENV1 immune dams . Similarly , a biphasic decay pattern was observed in mice born to DENV1 immune mothers and nursed on naïve mothers ( DN group ) . A first phase of rapid decay between birth and 3 week of age was observed with antibody half-life of 5 days , followed by a slower decay phase between week 3 and week 10 with antibody half-life of 16 . 5 days ( Fig 4A and Table 2 ) . When coating the ELISA plates with DENV2 virus to study the cross-reactivity of the anti-DENV1 maternal antibodies , similar trends were observed ( Fig 4B ) . Altogether , these observations thus suggested that i ) majority of DENV1-specific IgG antibodies measured in 3-week old mice originates from breast milk , ii ) placentally acquired dengue-specific IgG decay rapidly during the first 3 weeks after birth , and iii ) decay of both placentally acquired and breast milk-derived IgG antibodies display a biphasic pattern . Based on the ELISA results , 10-week old DN mice had negligible levels of anti-DENV1 IgG whereas 10-week old ND mice still displayed significant levels of these antibodies ( Fig 4A ) . To test whether such difference may translate into differential disease outcomes , 10-week old mice from the different groups were challenged with a sub-lethal dose of DENV2 . Mice in ND and DD groups displayed disease enhancement with majority of the mice being moribund by day 4–5 post-infection ( Fig 5A ) . In contrast , majority of the mice from DN group survived and displayed moderate clinical manifestations that were comparable to the NN control group ( Fig 5B ) . Together , these data confirm that majority of DENV1-specific IgG are acquired during breastfeeding and indicate a greater enhancement window in mice nursed by DENV1-immune mothers ( regardless of the immune status of their birth mothers ) compared to mice born to DENV1-immune mothers but nursed by naïve mothers . Besides being implicated in ADE , DENV-specific antibodies may also be involved in protection against infection . The protective role of dengue antibodies acquired from breast milk was thus examined . For this , a dengue homotypic model was set up where mice born to and nursed by DENV2-immune mothers were subsequently challenged with a lethal dose of the same DENV2 strain . The relative amounts of maternal DENV2 IgG acquired during gestation or from breastfeeding were again determined by switching pups at birth and measuring their systemic levels of DENV2 IgG at 5 weeks of age . Similar to the heterotypic ADE model , mice nursed by DENV2-immune mothers regardless of the immune status of their birth mothers ( DD and ND ) had comparable systemic levels of DENV2-IgG antibodies that were significantly higher than those measured in DN mice ( born to DENV2-immune mothers but nursed by naïve mothers ) ( Fig 6 ) . The neutralising activity ( PRNT50 ) of sera against DENV2 displayed similar trends ( Table 3 ) . The comparable levels of DENV2-IgG antibodies in DD and ND groups again indicated that the main route of maternal dengue-IgG transfer is during breastfeeding . Upon lethal DENV2 challenge , all mice in DD and ND groups survived ( Fig 7A ) and remained asymptomatic throughout the course of the experiment ( Fig 7B ) . These mice also had undetectable viremia levels ( Fig 7C ) . In contrast , more than 50% of mice from NN group were moribund by day 4 post-infection ( Fig 7A and 7B ) . Interestingly , mice in DN group displayed enhanced disease severity with 100% mortality ( Fig 7A and 7B ) and significantly higher virus titres compared to NN mice ( Fig 7C ) . Together these data suggested that 5-week old DD and ND mice , which acquired maternal DENV2-specific IgG antibodies during breastfeeding , had protective levels of maternal DENV2 antibodies . On the other hand , 5-week old DN mice , which acquired these antibodies during gestation only , had sub-neutralizing levels of maternal anti-DENV2 IgG that led to enhancement of disease severity .
The present study demonstrates the importance of dengue-specific IgG antibodies acquired during breastfeeding in disease enhancement and protection in a symptomatic mouse model . While maternal dengue-specific IgG antibodies acquired during either gestation , breastfeeding or via both routes enhanced dengue disease severity upon heterotypic infection , breastfeeding resulted in an extended window of disease enhancement . On the other hand , breastfeeding protected mice upon homotypic challenge , whereas the level of transplacentally acquired IgG antibodies led to enhancement of disease due to sub-neutralizing concentrations . Human challenge studies performed by Sabin in 1952 showed long-term protection against homotypic dengue infection , but only short-term protection against heterotypic infection , after which susceptibility to disease was observed [4 , 5] . Furthermore , epidemiological reports suggested increased risk of DHF/DSS development upon secondary heterotypic infection associated with pre-existing dengue antibodies [27 , 28] . It was also noted that dengue disease severity is influenced by the time interval in between two sequential heterotypic infections whereby a longer time interval was associated with greater disease severity [3 , 29] . In infants born to dengue immune mothers , it was observed that the risk of developing severe disease upon primary dengue infection was significantly increased between 5–9 months of age , correlating with maternal antibodies waning from protective to enhancing levels [5 , 8 , 18–22] . However , the relative contribution of antibodies acquired during gestation or during breastfeeding has never been investigated . Here , using our maternal antibody transfer mouse model , we found that high levels of maternal dengue-specific IgG antibodies were present in neonates at birth thus reflecting efficient IgG translocation across the placenta during gestation . However , a rapid decay of those placentally acquired IgG was observed as evidenced by >1 log drop in titres and an antibody half-life of 5 days only in 3-week old pups born to DENV1-immune dams but nursed by naïve mothers . In contrast pups born to naïve dams and breastfed for 3 weeks on DENV1-immune mothers had IgG titres that were comparable to the levels measured at birth in mice born to DENV1-immune mothers . This observation thus supported that breastfeeding provides sustained high levels of dengue-specific IgG . Interestingly , the half-life of serum DENV-IgG antibodies after 3 weeks of age ranged between 10 . 5 and 16 . 5 days , which is greater than the average half-life of 6–8 days reported for IgG antibodies in rodents [30] . This finding replicates well observations made in humans where maternal DENV-IgG antibodies persist for 24 to 150 days [31–33] whereas the average half-life of human IgG has been reported to be around 21 days [34] . Furthermore , a biphasic decay pattern was observed in all the groups characterized by an initial phase during which the maternal IgG half-life ranges from 5 to 11 . 5 days , followed by a second slower decay phase where the antibody half-life ranges from 14 . 5 to 16 . 5 days . Such biphasic antibody decay pattern has been previously reported in infants who maternally acquired dengue-specific IgG antibodies , and was characterized by a first decay phase with antibody half-life of 24–29 days , followed by a slower decay phase with antibody half-life of 44–150 days [31–33] . Similar observations were made in vaccination studies against pertussis and malaria , indicating that this phenomenon of biphasic antibody decay pattern is not specific to dengue [35 , 36] . The importance of breastfeeding as a major maternal IgG transfer route is in line with previous literature reporting efficient intestinal FcRn-mediated IgG translocation in rodents [22 , 23 , 37 , 38] . In humans , maternal IgG antibody transfer is believed to occur mainly through placental FcRn-mediated translocation during gestation [22–24] . However , there are very limited studies on maternal IgG antibodies in human milk and their impact on infants [21] . In addition , the role of human intestinal FcRn in facilitating IgG translocation is not widely documented . Nevertheless , the existence of FcRn in human intestinal cells has been reported [39 , 40] and it was shown to be functional in bidirectional transport of IgG antibodies [41] . In addition , intestinal FcRn in human foetus was suggested to mediate maternal IgG uptake from ingested amniotic fluid during gestation [42–44] . Thus , the same intestinal FcRn may also function in postnatal IgG translocation during breastfeeding . Further investigation on human FcRn is thus necessary and may reveal a greater role and importance of breastfeeding-mediated maternal IgG transfer . Studies of breast milk antibodies indeed mainly focused on secretory IgA involved in mucosal defence [26] . However , in the context of dengue , IgA antibodies are likely to play a minimal role , as evidenced by undetectable levels of dengue-specific IgA in dengue-infected mice . Besides maternal antibodies , immunologically active components such as lactoferrin , lysozyme and immune cells including macrophages , neutrophils and lymphocytes have also been reported to be transmitted maternally via breastfeeding [45 , 46] . Transfer of maternal T- and B-lymphocytes was proposed to play a role in immune tolerance against maternal HLA [47] , and to provide IgG-secreting B-lymphocytes that could possibly be involved in the maintenance of detectable levels of serum IgG in the offspring [48] , respectively . However , trans-epithelial migration of maternal lymphocytes to the systemic circulation in infants remains controversial with conflicting literature , indicating that more experimental work needs to be done in order to fully assess the contribution of these maternally transferred lymphocytes [46 , 49] . Nevertheless , the possibility that immunologically active components present in the breast milk from DENV1-immune mothers other than free DENV1-specific IgG antibodies may play a role in disease enhancement cannot be completely ruled out . However , the fact that passive administration of purified monoclonal enhancing IgG antibodies recapitulated ADE strongly support that IgG antibodies play an important role in triggering this phenomenon [50–52] . Current guidelines neither preclude nor encourage breastfeeding by dengue-immune mothers as this area has been unexplored [46] . Our work thus provides the first evidence of the possible role of breast milk acquired dengue-specific IgG antibodies in both dengue disease enhancement and protection in infants . However , since the four DENV serotypes co-circulate in most of the dengue endemic countries , mothers are likely to be immune to more than one serotype . In fact , majority of infants who displayed disease enhancement were found to be born from mothers immune to more than one DENV serotype [32; 53 , 54] . Thus , using the same mouse model of maternal IgG transfer , it will be interesting to investigate the infection outcome in pups born to mothers immune to more than one DENV serotype . Also , a prospective study enrolling breastfeeding and non-breastfeeding dengue-immune mothers , and their babies in dengue endemic regions would provide further insight on the contribution of breast milk acquired antibodies in dengue disease and protection .
|
Epidemiological observations showed that 5–9 month old infants born to dengue immune mothers have increased risk of developing severe disease upon primary dengue infection . This disease enhancement has been associated with the presence of binding but non-neutralizing maternal dengue antibodies . The recent development of experimental dengue mouse models involving maternal antibodies supports their role in both disease enhancement and protection . Here , we examined the contribution of maternal antibodies acquired during gestation and breastfeeding in disease enhancement and protection . Our findings support that majority of maternal IgG antibodies circulating in mice born to dengue immune mothers are acquired from breast milk . As such , we showed that breastfeeding conferred extended window of enhancement or protection . These findings provide the first experimental evidence for a role of breast milk dengue antibodies in mediating dengue infection outcome . This may help develop guidelines to dengue immune breastfeeding mothers .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"neonatology",
"medicine",
"and",
"health",
"sciences",
"milk",
"radiochemistry",
"immune",
"physiology",
"enzyme-linked",
"immunoassays",
"body",
"fluids",
"maternal",
"health",
"breast",
"milk",
"animal",
"models",
"of",
"disease",
"immunology",
"tropical",
"diseases",
"microbiology",
"pediatrics",
"nuclear",
"decay",
"animal",
"models",
"model",
"organisms",
"women's",
"health",
"neglected",
"tropical",
"diseases",
"antibodies",
"immunologic",
"techniques",
"research",
"and",
"analysis",
"methods",
"animal",
"models",
"of",
"infection",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"animal",
"studies",
"proteins",
"dengue",
"fever",
"immunoassays",
"chemistry",
"mouse",
"models",
"physics",
"biochemistry",
"breast",
"feeding",
"anatomy",
"nuclear",
"physics",
"physiology",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"physical",
"sciences"
] |
2016
|
Relative Contribution of Dengue IgG Antibodies Acquired during Gestation or Breastfeeding in Mediating Dengue Disease Enhancement and Protection in Type I Interferon Receptor-Deficient Mice
|
One of the key immunological characteristics of active visceral leishmaniasis ( VL ) is a profound immunosuppression and impaired production of Interferon-γ ( IFN-γ ) . However , recent studies from Bihar in India showed using a whole blood assay , that whole blood cells have maintained the capacity to produce IFN-γ . Here we tested the hypothesis that a population of low-density granulocytes ( LDG ) might contribute to T cell responses hyporesponsiveness via the release of arginase . Our results show that this population is affected by the anticoagulant used to collect blood: the frequency of LDGs is significantly lower when the blood is collected with heparin as compared to EDTA; however , the anticoagulant does not impact on the levels of arginase released . Next , we assessed the capacity of whole blood cells from patients with active VL to produce IFN-γ and IL-10 in response to antigen-specific and polyclonal activation . Our results show that whole blood cells produce low or levels below detection limit of IFN-γ and IL-10 , however , after successful treatment of VL patients , these cells gradually regain their capacity to produce IFN-γ , but not IL-10 , in response to activation . These results suggest that in contrast to VL patients from Bihar , India , whole blood cells from VL patients from Gondar , Ethiopia , have lost their ability to produce IFN-γ during active VL and that active disease is not associated with sustained levels of IL-10 production following stimulation .
Visceral leishmaniasis ( VL ) is a neglected tropical disease caused by parasites of the Leishmania ( L . ) donovani complex . An estimated 200 , 000 to 400 , 000 new cases of VL with an incidence of 50 , 000 deaths occur each year , however these numbers are widely acknowledged to be a gross underestimation of the real burden [1 , 2] . In global estimates , Sudan , South Sudan , Ethiopia , Kenya and Somalia account for the second largest number of annual VL cases , after South Asia [1] VL inflicts an immense toll on the developing world and impedes economic development , with an estimated loss of 2 . 3 million disability-adjusted life years . There is no effective vaccine; currently used chemotherapy is toxic and increasing drug resistance is reported [3] . VL can be asymptomatic or can manifest as a progressive disease characterised by hepatosplenomegaly , fever , weight loss , hyperglobulinemia and pancytopenia . In Ethiopia , VL is caused by L . donovani and it is one of the most significant vector-borne diseases; Ethiopia has the second largest number of VL cases in sub-Saharan Africa with an estimated annual burden of 4500 to 5000 new cases [2] . VL is worsened by malnutrition and HIV co-infection , and treatment access is often difficult because of the remote location of areas endemic for VL . Non-healing VL in humans has been associated with increased levels of IL-10 , a potent immunosuppressive cytokine ( reviewed in [4] ) and indeed , there is also ample evidence in the literature that patients with active VL are severely immunosuppressed and do not respond to the Leishmanin skin test . In addition , their PBMCs fail to produce IFN-γ and to proliferate in response to Leishmania antigen; this impaired capacity to respond to antigenic challenge is restored following successful chemotherapy [5] and reviewed in [4 , 6 , 7] . The mechanisms leading to these impaired T cell responses during symptomatic VL remain to be fully identified . We have recently shown that L-arginine depletion contributes to this lack of T cell responses: arginase-induced L-arginine metabolism has been identified as a potent mechanism of immune suppression [8–10] . We have shown previously in both experimental and human leishmaniasis that arginase activity is significantly increased in non-healing disease . In human leishmaniasis , we identified the phenotype of arginase-releasing cells as low-density granulocytes ( LDGs ) , as these cells were collected in the PBMCs fraction , but not in the erythrocytes fraction . We have also established that LDGs are activated neutrophils that have degranulated and released arginase[11 , 12] . The subsequent elevated arginase in the microenvironment efficiently depletes L-arginine , an amino acid that is essential for efficient T cell responses , and this reduction in L-arginine results in impaired T cell responses [12 , 13] . Recent studies in Bihar , India , have challenged our current view on the apparent hyporesponiveness of PBMCs from patients with active VL to antigen-specific stimulation . Using a whole blood assay ( WBA ) , the authors showed that whole blood cells produce IFN-γ in response to antigenic activation [14 , 15] and identified CD4+ T cells as the main type of IFN-γ producing cells [16] . The levels of IFN-γ were similar before and after successful treatment , suggesting that the inability of these patients to control the disease was not due to a defect in Th1 responses . In contrast , IL-10 production was elevated in the group with active VL and significantly reduced in cured patients . In the present study , we evaluated the responsiveness of whole blood cells from a cohort of patients with active VL in Gondar , North West of Ethiopia . We first assessed whether the frequency of immunomodulatory LDGs in the blood of patients with active VL was affecting the levels of arginase activity in a WBA and might therefore affect the levels of cytokines . In the next step , we evaluated the antigen-specific production of IFN-γ and IL-10 over time in these patients .
This experimental study was approved by the Institutional Review Board of the University of Gondar ( IRB , reference SBMLS/1199/07 ) and informed written consent was obtained from each patient and control . For this cross-sectional study , a cohort of 23 patients with active visceral leishmaniasis ( VL patients ) , whose diagnosis of VL was based on positive serology ( rK39 ) and presence of amastigotes in spleen or bone marrow aspirates [17] was recruited from the Leishmaniasis Treatment and Research Center of Gondar University Hospital before treatment . All the patients in this study presented with fever , hepatosplenomegaly , pancytopenia and low BMI . Their age , duration of illness , parasite grade and treatment are summarized in Table 1 . In addition , 16 VL patients after the end of successful treatment ( 17 days = TOC ( Test Of Cure ) ) , 20 patients 3 months following successful treatment; 10 patients 6 months following successful treatment ( from different groups of patients at each time point ) ; and 10 non-endemic healthy age- and sex-matched individuals ( controls ) coming from the city of Gondar , which is non-endemic for visceral leishmaniasis were recruited at the Leishmaniasis Treatment and Research Center of Gondar University Hospital . TOC was defined as follows: at the end of successful treatment , patients look improved , afebrile , and usually have a smaller spleen size than on admission and have an increased haemoglobin ( Hgb ) level . No women presented with visceral leishmaniasis during our study , all patients were male migrant workers . Patients < 18 years old or presenting with tuberculosis or malaria were excluded from the study . All VL patients were routinely screened for HIV using the following tests: KHB Shanghai Kehua Bio-engineering Co . Ltd and Chembio HIV 1/ 2 STAT-PAK; Uni-Gold ( Trinity Biotech PLC ) was used to resolve ambiguous results; all patients enrolled in our study were HIV negative . 4–8 ml of blood was collected in EDTA and/or heparin tubes . Patients were treated with a combination of sodium stibogluconate ( SSG , 20mg/kg body weight/day ) , and paromomycin ( PM , 15mg/kg body weight/day ) injections , given intramuscularly for 17 days or with Ambisome ( max of 30mg/kg body weight , with 6 injections of 5mg/kg body weight /day ) and showed an initial clinical cure rate of 100% after treatment ( TOC ) . Antibodies used were as follows: anti-CD15 ( Clone H198 , BD Pharmingen ) , anti-arginase I ( HyCult Biotechnology: clone 6G3 ) and the isotype control ( BD Pharmingen: clone MOPC21 ) coupled with Alexa Fluor 647 ( Molecular Probes ) . Cells were washed with PBS , the fixation step was performed with 2% formaldehyde in PBS and the permeabilization step with 0 . 5% saponin in PBS . The determination of intracellular arginase was performed as described in [18] . The percentages for the isotype controls were <1% . Acquisition was performed using a FACSCalibur ( BD Biosciences ) and data were analyzed using Summit v4 . 3 software . Arginase activity was measured as described in [18] . To determine arginase activity in plasma from stimulated blood samples , urea concentrations were first determined without the activation and hydrolysis steps; these values were subtracted from those obtained by measuring the urea levels as described in [11] . One unit of enzyme activity is defined as the amount of enzyme that catalyzes the formation of 1 μmol of urea per min . Three x 2ml of blood were collected in EDTA tubes and 3 x 2ml in heparin tubes ( BD ) . Soluble Leishmania antigen ( SLA ) was prepared from stationary-phase L . donovani promastigotes isolated from 5 patients as described in [14] , and was added immediately after blood collection at a concentration of 5μg/mL and phytohaemagglutinin ( PHA , Sigma ) at 10 μg/mL . For the stimulation in the presence of L-arginine , two x 2ml tubes of blood were collected and 1mM L-arginine ( Sigma ) was added directly in the tubes . Unstimulated blood was used as negative control ( nil ) . Plasma from activated blood samples and negative controls was collected after 24 hours of incubation at 37°C and stored at -20°C for further analysis . The level of IFN-γ and IL-10 in the plasma from stimulated blood was measured using Human IFN-γ and IL-10 ELISA Ready-SET-Go ! kit using the manufacturer’s instructions and procedure ( eBioscience ) . Antigen-specific IFN-γ and IL-10 levels ( expressed in pg/mL ) produced in response to SLA and PHA stimulation were determined by subtracting background levels measured in the non-stimulated samples ( nil ) . The detection limit for IFN-γ and IL-10 was 2 pg/ml . Data were analyzed for statistical differences using nonparametric two-sided Mann-Whitney , Wilcoxon or Kruskal-Wallis tests ( GraphPad Prism 6 ) when appropriate and differences were considered statistically significant at p < 0 . 05 . The Bonferroni method was used for multiplicity correction whenever it was needed . Unless otherwise specified , results are expressed as median ± SEM .
We have previously shown that the frequency of LDGs is significantly increased in patients with active VL [12] . Since these cells have been shown to have immunomodulatory properties , we proposed that LDG-mediated T cell suppression is a key element in the outcome of VL . Our preliminary data also suggested that the anticoagulant used to collect blood had a remarkable effect on the survival of LDGs ex vivo [12] . Here , we tested the impact of two commonly used anticoagulants , EDTA and heparin , on the frequencies of LDGs: blood was collected from VL patients in EDTA and heparin tubes , their PBMCs were isolated by Ficoll gradient and the frequencies of CD15+arginase+ cells ( LDGs [12] ) were determined by flow cytometry . Results presented in Fig 1 show that the frequency of LDGs is significantly lower in PBMCs isolated from blood from VL patients collected in heparin as compared to EDTA tubes ( Fig 1A , Table 2 ) . Similar results were obtained with blood from controls collected in EDTA and heparin ( Fig 1B , Table 2 ) , indicating that heparin affects the frequency of LDGs not only in patients with active VL but also in controls . It has been reported recently that whole blood cells from patients with active VL maintain the capacity to produce IFN-γ and IL-10 following activation of whole blood with soluble Leishmania antigen ( SLA ) [15] . Here we first tested the impact of heparin and EDTA on the production of IFN-γ in the whole blood assay ( WBA ) . To obtain the levels of IFN-γ produced following stimulation of whole blood cells with SLA or PHA , the background levels measured in the non-stimulated samples ( nil ) were substracted from the activated samples . The production of IFN-γ in the WBA in response to SLA was low or below detection limit , independently of the anticoagulant used ( EDTA: 54 . 1±23 . 9 vs heparin: 61 . 3±54 . 0 , p = 0 . 3125 , data not illustrated ) . Similar results were obtained following polyclonal activation with PHA ( EDTA: 42 . 9±34 . 8 vs heparin: 41 . 8±19 . 9 , p = 0 . 6523 , data not illustrated ) . We have previously shown that arginase-induced L-arginine depletion can suppress T cell activation and cytokine release [13] . To determine whether the poor IFN-γ response observed in the WBA could be due to increased levels of arginase activity released in the microenvironment , we measured the activity of arginase in the plasma after the 24 hours incubation . Results presented in Fig 2 show that the levels of arginase activity were not significantly affected by the anticoagulant used ( summarized in Table 3 ) . The levels of arginase activity were similar in all groups tested ( Fig 2 , p = 0 . 949 ) , suggesting that the anticoagulants used did not impact on the production of IFN-γ via increased release of arginase . Our results show that the IFN-γ production is low or below detection limit in the WBA and that this is unlikely to be due to arginase-induced L-arginine depletion . To exclude any technical problem with the IFN-γ ELISA assay , we stimulated whole blood cells from non-endemic controls with PHA in exactly the same conditions and in the same laboratory as the VL patients . Results presented in Fig 3 show that following polyclonal activation , IFN-γ was clearly detectable when the blood was collected with heparin , indicating that the unresponsiveness of some VL patients to PHA in the WBA is not due to technical issues . Since EDTA chelates the Ca+ needed for cellular activation , IFN-γ was low or below detection limit when the blood was collected with EDTA ( 630 . 4±166 . 1 vs 13 . 3±20 . 0 , p = 0 . 0020 , Fig 3 ) . As expected [19] , these results show that collecting blood in EDTA prevents the production of IFN-γ . IFN-γ was low or below detection limit in the supernatant of whole blood from non-endemic healthy controls activated with SLA ( 2 . 10 ±1 . 19 pg/ml , data not illustrated ) . IFN-γ is elevated in the plasma of VL patients [4] and indeed our data show elevated levels of IFN-γ in plasma from active VL patients as compared to controls ( 123 . 0±27 . 5 pg/ml vs 31 . 1±23 . 2pg/ml , p = 0 . 0030 , Table 4 ) . Recent studies in Bihar , India , have shown that the production of antigen-specific IFN-γ in the WBA from patients with active VL was similar to that in cured VL patients , suggesting that T cells from these patients have the capacity to respond to antigenic challenge by producing significant amounts of IFN-γ [15] . These results were contrary to previous studies showing that PBMCs from VL patients do not produce IFN-γ after stimulation with Leishmania antigen ( reviewed in [4 , 6] ) . To evaluate the capacity of whole blood cells from VL patients in Gondar , Ethiopia , to respond to antigenic challenge over time by using the WBA , we performed a cross-sectional study and collected blood in heparin before the start of the treatment , at the end of successful treatment ( for 17 days , TOC ) and 3 and 6 months after treatment ( from different groups of patients at each time point ) and activated blood cells with SLA and PHA . In contrast to the results obtained previously in India , the levels of antigen-specific IFN-γ were increased significantly and gradually after successful treatment ( for 17 days , TOC ) , 3 months and finally 6 months ( Fig 4A , Table 5 ) . Similar results were obtained in response to polyclonal activation ( Fig 4B , Table 5 ) : the production of IFN-γ in response to PHA was low or below detection levels before treatment . To further determine whether the lack of clear response was due to low levels of L-arginine in the plasma of these patients [12] , whole blood from active VL patients was activated with PHA in the presence or absence of L-arginine and the resulting levels of IFN-γ produced were similar in both groups ( 123 . 9±84 . 2 vs 119 . 8±42 . 9 , respectively , p = 0 . 6667 , data not illustrated ) . In the study conducted in India , IFN-γ levels detected in the WBA were similar before and after treatment , in contrast , IL-10 levels were elevated during active VL , but reduced significantly in cured patients , suggesting that IL-10 is associated with active disease [15] . Furthermore , IL-10 is also elevated in the plasma from patients with active VL , and our results also show that plasma IL-10 is significantly increased in VL as compared to healthy controls ( 88 . 9±12 . 5 vs 4 . 0±5 . 5pg/ml , respectively , p = 0 . 0002 , Table 4 ) . Here we compared the levels of IL-10 in the plasma of the WBA over time . Unexpectedly , results presented in Fig 5A show that antigen-specific IL-10 levels are low or below detection limit during active VL , after the 17 days of treatment ( TOC ) and 3 and 6 months after the end of successful treatment ( summarized in Table 6 ) . These results suggest that active VL is not associated with high IL-10 production in the WBA in the cohort of VL patients from North West Ethiopia . However , blood cells from these patients have the capacity to produce IL-10 in response to polyclonal activation after treatment , as results in Fig 5B show gradually increasing levels of IL-10 after successful treatment ( for 17 days , TOC ) , 3 months and finally 6 months ( Fig 5B , Table 6 ) . Our results presented in Table 7 also show that whole blood cells from VL patients cannot be induced to secrete these cytokines in vitro since the levels of both cytokines in response to SLA and nil or PHA and nil were similar in the plasma of the WBA . However , significant differences in IFN-γ levels between SLA and nil and PHA and nil were observed after successful treatment ( for 17 days , TOC ) , 3 and 6 months after the end of treatment , demonstrating that in this setting , production of IFN-γ can be induced after successful treatment , but not at time of active VL . A similar conclusion can be made with the production of IL-10 in response to PHA . These results demonstrate that in patients from Gondar , whole blood cells are hyporesponsive during the active phase of the disease , however , this is progressively reversed after successful treatment . Furthermore , our results show that increased IL-10 production by whole blood cells is not a hallmark of active VL in patients from Gondar .
Recently , the existing dogma on T cell hyporesponsiveness during active VL was challenged by data showing that VL patients from India maintain the capacity to produce IFN-γ in the WBA [15]; these studies suggest that the inability of these patients to control the disease was not due to a defect in the Th1 response . We first tested whether the anticoagulant used in the WBA might impact on the frequency of LDGs , a subpopulation of neutrophils with immunomodulatory properties that can suppress the production of cytokines . Indeed , our preliminary data suggested that the use of heparin , the anticoagulant used in the WBA discussed above , results in a sharp reduction of the frequency of LDGs [20] . Our results show that the frequency of LDGs is drastically reduced when the blood is collected in heparin , suggesting that the use of heparin as anticoagulant could result in underestimated frequencies of LDGs . Whereas EDTA has been shown to impact on biological functions of neutrophils [21] , their viability does not seem to be affected by different anticoagulants [21 , 22] . EDTA chelates the free calcium needed as a cofactor to activate enzymes responsible for coagulation , whereas heparin blocks coagulation by activating antithrombin . LDGs are a distinct subpopulation of highly activated neutrophils; thus , it is possible that LDGs isolated from blood harvested in EDTA lack Ca+ required to undergo cell death and therefore survive longer in EDTA than LDGs harvested in the presence of heparin . Similarly , the chelation of Ca+ by EDTA explains the lack of IFN-γ production in whole blood assays [19] . We cannot formally exclude the possibility that EDTA may results in activation and thereby degranulation of neutrophils , however , we and others have shown that the frequency LDGs is increased in several conditions , such as HIV[23] , SLE [24] , visceral leishmaniasis[12] , pregnancy[18] , asthma[25] and cancer[26] , supporting our conclusions that this is specific to inflammatory conditions in the affected individuals . Furthermore , our samples are processed immediately and it is therefore unlikely that neutrophils from VL patients are "more" activated then those from controls in this short period of time . Release of arginase in the microenvironment results in the depletion of extracellular L-arginine , that in turn prevents T cell activation [10] . We have previously shown that in patients with active VL , the frequency of activated LDGs in their PBMCs is significantly increased , that these cells express significantly less intracellular arginase and that the levels of arginase in the plasma is significantly increased [12] . We considered the possibility that the lack of IFN-γ response in the WBA could be due to increased levels of released arginase: however , our results show that the levels of arginase activity are similar in all plasma samples harvested 24 hours after activation of whole blood cells showing that enhanced arginase release is not accounting for the hyporesponsiveness of the cells in the WBA . But we cannot exclude that LDGs might still account for the observed hyporesponsiveness via other mechanisms such as cell-cell contact or release of molecules . However , we can conclude from our results that cells from whole blood cells collected from patients with active VL produce low or no IFN-γ in response to antigenic or polyclonal activation , suggesting that blood cells from patients with active VL are hyporesponsive . We can exclude technical problems with our assay , as IFN-γ was clearly detectable in the supernatant of whole blood cells from healthy controls activated with PHA . Our results are in agreement with ample evidence from the literature showing that one of the key immunological characteristics of active VL is profound immunosuppression , as demonstrated by the failure of PBMCs to produce IFN-γ and proliferate in response to Leishmania antigen ( reviewed in [4 , 6] ) . Whereas ex vivo , PBMCs from patients with active VL cannot be induced to produce IFN-γ and IL-10 in response to antigenic or polyclonal stimulation , it is still possible that responsive cells remain at the site of pathology; indeed , IFN-γ and TNF-α have been detected in the supernatant of spleen cells from VL patients [27] . Of note , no significant correlations were found between the levels of IFN-γ and IL-10 detected between PHA and SLA stimulation , nor with the levels of arginase . The results presented in the current study are in apparent contradiction with recent studies showing that IFN-γ is produced by whole blood cells from patients with active VL and that their IFN-γ levels were similar to those levels detected in cured patients [15] . Furthermore , our results also show that antigen specific IL-10 production in the WBA is not associated with active VL in Ethiopia , nor that it is produced by whole blood cells from cured patients in response to antigenic stimulation . Thus , our results clearly demonstrate that blood cells from patients with active VL are hyporesponsive , as activation with PHA resulted in significantly increased levels of IL-10 only in cured patients . Taken together our cytokine results show that blood cells from patients with active VL are hyporesponsive to both antigen-specific ( as summarized in the review by Kumar et al . [7] ) and polyclonal activation ( as previously shown in [28–30] ) . The discrepancies between the results presented here and the above mentioned study might be explained by several factors: We cannot exclude that the low or undetected levels of IFN-γ might be due to the severe lymphopenia in VL patients ( 2 . 1±0 . 2 in VL patients vs 5 . 7±0 . 6 white blood cells ( x103 ) ; normal range = 4 . 5–10 . 5 white blood cells ( x103 ) ) , however , there was no correlation between the levels of IFN-γ in the supernatant of the WBA and the WBC counts ( p = 0 . 1052 , data not illustrated ) , suggesting that the low frequency of cells is unlikely to account for the observed levels of IFN-γ . Despite the fact that we find little or no production of IFN-γ and IL-10 in the plasma harvested from the WBA at time of acute disease , these cytokines are clearly detected in the plasma of these patients directly ex vivo ( summarized in [4] ) demonstrating that these cytokines have been produced in vivo . Indeed , whereas the levels of these cytokines in the WBA in response to antigenic or polyclonal activation were below or barely above the levels of cytokines detected in the absence of stimulation , IFN-γ and IL-10 were detectable in the supernatant of the non-activated sample of whole blood cells ( 185 . 9±178 . 7 and 61 . 5±12 . 6 pg/ml , respectively ) as well as in the plasma of these patients ( 123 . 0±27 . 5 and 88 . 9±12 . 5 pg/ml , respectively ) ( E . Adem , F . Tajebe , M . Getahun and P . Kropf , data not illustrated ) . This demonstrates that these cytokines are produced in vivo , but that whole blood cells cannot be induced to secrete these cytokines in vitro in response to activation . It is tempting to speculate that other cells , such as neutrophils , monocytes and spleen cells produce these cytokines . In the current study , we show that whole blood cells from patients with active VL are hyporesponsive as no or low IFN-γ was released in response to activation . Since a recent study [16] showed that IFN-γ produced by antigen-specific CD4+ T cells contributes to the control of parasite replication in VL patients , it is possible that the lack of appropriate Th1 response might be responsible for the uncontrolled parasite replication in the patients in Gondar . The use of the WBA has many advantages , such as being easy to perform , not requiring the use of sophisticated equipment and using only a small amount of blood . In addition , it is likely to contain all the factors necessary for cell activation and should mimic the in vivo conditions as closely as possible . The WBA provides a simple tool for determining cytokine profiles that may be useful laboratory predictors of early disease , aiding the evaluation of new interventions and offering insights into disease pathogenesis .
|
The leishmaniases , a group of diseases caused by Leishmania parasites , belong to the most neglected tropical diseases: they are mainly found in low-income countries and affect the poorest populations . These parasites infect cells of the immune system called macrophages , which can kill the intracellular parasites in response to soluble mediators they receive from other cells of the immune system , the lymphocytes . Visceral leishmaniasis is the most severe form of the leishmaniases and is characterized by enlarged liver and spleen , fever , weight-loss and anaemia and represents a major public health problem in Ethiopia . Currently there is no vaccine available , the existing treatment has many severe side effects and drug-resistance is increasing . In the present study , we worked with patients suffering from visceral leishmaniasis . This form of the disease is fatal if the patients are not treated . We studied the ability of lymphocytes isolated from their blood to produce soluble mediators before and at different times after the end of treatment . Our results show that the lymphocytes have an impaired capacity to produce the soluble mediator required to instruct infected cells to kill the intracellular parasites , but that this lack of response is gradually restored with time after successful treatment .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"kala-azar",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"immune",
"physiology",
"cytokines",
"immune",
"cells",
"drugs",
"immunology",
"tropical",
"diseases",
"parasitic",
"diseases",
"developmental",
"biology",
"molecular",
"development",
"neglected",
"tropical",
"diseases",
"pharmacology",
"heparin",
"infectious",
"diseases",
"white",
"blood",
"cells",
"zoonoses",
"animal",
"cells",
"t",
"cells",
"protozoan",
"infections",
"blood",
"plasma",
"hematology",
"immune",
"system",
"blood",
"anatomy",
"cell",
"biology",
"physiology",
"leishmaniasis",
"biology",
"and",
"life",
"sciences",
"cellular",
"types"
] |
2016
|
Successful Treatment of Human Visceral Leishmaniasis Restores Antigen-Specific IFN-γ, but not IL-10 Production
|
In many bacteria , including Vibrio cholerae , cyclic dimeric guanosine monophosphate ( c-di-GMP ) controls the motile to biofilm life style switch . Yet , little is known about how this occurs . In this study , we report that changes in c-di-GMP concentration impact the biosynthesis of the MshA pili , resulting in altered motility and biofilm phenotypes in V . cholerae . Previously , we reported that cdgJ encodes a c-di-GMP phosphodiesterase and a ΔcdgJ mutant has reduced motility and enhanced biofilm formation . Here we show that loss of the genes required for the mannose-sensitive hemagglutinin ( MshA ) pilus biogenesis restores motility in the ΔcdgJ mutant . Mutations of the predicted ATPase proteins mshE or pilT , responsible for polymerizing and depolymerizing MshA pili , impair near surface motility behavior and initial surface attachment dynamics . A ΔcdgJ mutant has enhanced surface attachment , while the ΔcdgJmshA mutant phenocopies the high motility and low attachment phenotypes observed in a ΔmshA strain . Elevated concentrations of c-di-GMP enhance surface MshA pilus production . MshE , but not PilT binds c-di-GMP directly , establishing a mechanism for c-di-GMP signaling input in MshA pilus production . Collectively , our results suggest that the dynamic nature of the MshA pilus established by the assembly and disassembly of pilin subunits is essential for transition from the motile to sessile lifestyle and that c-di-GMP affects MshA pilus assembly and function through direct interactions with the MshE ATPase .
Vibrio cholerae , the causative agent of the human intestinal disease cholera , is a natural inhabitant of aquatic ecosystems [1] . Cholera infection results from consumption of food and water contaminated with V . cholerae . Subsequently the bacteria turn on regulatory networks that facilitate bacterial growth and survival during the infection process [2] . They also activate production of virulence factors including the toxin-coregulated pilus ( TCP ) , essential for intestinal colonization , and the cholera toxin ( CT ) , responsible for production of massive watery diarrhea that results in dissemination of V . cholerae back to aquatic ecosystems [3] . V . cholerae’s ability to cause epidemics is tied to its dissemination and survival in aquatic habitats and its transmission to the human host . One critical factor for dissemination , environmental survival and transmission of the pathogen is its ability to form matrix enclosed surface-associated communities termed biofilms [4–6] . V . cholerae forms biofilms on the surfaces of phytoplankton and zooplankton [7] , and exists in the surface waters of cholera endemic areas in matrix-enclosed aggregates thought to arise from biofilm-like populations of V . cholerae present in human stools [5] . Removal of particles >20 μm in diameter from water can reduce cholera incidence by 48% [8] . Additionally , growth in a biofilm induces a hyper-infectious phenotype [9] . Collectively , these studies highlight the importance of the biofilm growth mode in both the intestinal and aquatic phases of the V . cholerae life cycle . Biofilm formation by V . cholerae begins with surface attachment , and subsequent development of microcolonies and mature biofilm structures [10–16] . The biofilm matrix is primarily composed of Vibrio exopolysaccharide ( VPS ) [17] , extracellular DNA [18] , and biofilm matrix proteins ( RbmA , RbmC , and Bap1 ) [13 , 14 , 16 , 19] , which are required for cell-cell and cell-surface interactions and development of mature V . cholerae biofilms [15] . Two cell-surface structures , a single polar flagellum and type IVa mannose-sensitive hemagglutanin pili ( MshA ) , are critical for initial attachment and biofilm formation [11 , 20] . Type IVa pili have the ability to rapidly extend and retract , contributing to twitching and swarming motility in many bacteria [21] . Though V . cholerae produces the type IVa MshA pilus , twitching motility has not been reported . Genes required for biogenesis of MshA pilus are clustered into 16 . 7 kb region in V . cholerae chromosome-I and organized into two operons: the first operon harbors 9 genes from mshI-mshF predicted to encode proteins required for assembly and secretion and the second harboring 7 genes from mshB—mshQ encoding pilus structural components mshB-Q [22] . The pilus is comprised of repeats of the major pilin , MshA . Predicted function of the proteins encoded by MSHA gene locus are included in Table 1 . MshA pili and flagellum are also crucial for two distinct near-surface motility trajectories of V . cholerae: ‘roaming’ and ‘orbiting’ , [10] . Low curvature ‘roaming’ trajectories meander over large distances and result from weak MshA-surface interactions . In contrast , orbiting trajectories repeatedly trace out tight circular tracks over the same region , and are the result of strong MshA-surface interactions . Cells that attach to the surface come from the orbiting subpopulation , while the roaming subpopulation pass over the surface without attaching . That orbiting motility is ablated when a mannose derivative is added to the medium to saturate MshA pilus binding further indicates that interactions between MshA pili with the surface are important [10] . The second messenger cyclic dimeric guanosine monophosphate ( c-di-GMP ) is an important promoter of the switch from motile planktonic growth mode to biofilm growth mode [23–25] . c-di-GMP synthesis is catalyzed by diguanylate cyclases ( DGC ) harboring a GGDEF domain while degradation is catalyzed by phosphodiesterases ( PDE ) harboring an EAL or HD-GYP domain . Subsequently , c-di-GMP is sensed by different classes of receptor proteins or RNAs and thereby converted to specific phenotypic outputs affecting motility , biofilm formation , and virulence . Elevated intracellular levels of c-di-GMP inhibit motility both by post-transcriptional and transcriptional mechanisms . In Salmonella enterica and Escherichia coli , the PilZ class of c-di-GMP receptor protein YcgR affects flagellar motor functions through interaction with FliG and FliM subunits of the flagellar rotor or the stator subunit MotA [26 , 27] . In Pseudomonas aeruginosa and V . cholerae , c-di-GMP inhibits motility by repressing transcription of flagellar genes through the AAA+ ATPase enhancer binding class of c-di-GMP receptor and transcriptional regulators FleQ [28] and FlrA [29] , respectively . In addition to the regulation of flagellum production and activity by c-di-GMP , there are reports of c-di-GMP regulating the assembly and activity of Type IV pili . In Klebsiella pneumoniae , c-di-GMP is bound by PilZ class of c-di-GMP receptor protein MrkH , which upregulates the transcription of the fimbrial subunit mrkA [30–32] . In P . aeruginosa , the degenerate GGDEF-EAL domain class of c-di-GMP receptor FimX modulates Type IV pili production in an intracellular c-di-GMP concentration-dependent manner [33 , 34] . The V . cholerae genome encodes 31 GGDEF domain , 12 EAL domain and 10 dual GGDEF/EAL domain proteins [35] . Systematic analysis of in-frame deletion mutants of all V . cholerae genes encoding proteins with GGDEF and/or EAL domains for motility phenotypes revealed that four DGCs ( CdgH , CdgK , CdgL , and CdgD ) and two PDEs ( CdgJ and RocS ) affect motility in an LB soft agar motility assay [36] . Though deletion of the PDE cdgJ affected motility , no difference in intracellular c-di-GMP concentration was observed between the ΔcdgJ mutant and WT [36] . This measurement was conducted from a population , so there may be subcellular localized differences or population differences that affect motility . The molecular mechanism of c-di-GMP mediated motility repression and contribution of these DGCs and PDEs to switch from motile to surface-associated lifestyle remains elusive . In this study , we demonstrate that c-di-GMP inversely regulates motility and biofilm formation through direct regulation of the assembly and activity of the MshA pilus . Swimming motility is impaired in strains lacking the phosphodiesterase cdgJ , and disruption of the assembly or disassembly of the MshA pilus restores motility to WT levels by reducing the interactions with surfaces . Quantitative measurements indicate that c-di-GMP leads to increased production of MshA pili , which in turn bind surfaces and reduce motility . We demonstrate that the ATPase responsible for pilus polymerization , MshE , functions as a c-di-GMP receptor thereby providing an input for the c-di-GMP signal into the assembly of the MshA pilus . Collectively , this study elucidates how type IV pili and swimming motility are regulated by c-di-GMP in V . cholerae by presenting the first characterization of the complex involved in the assembly and disassembly of the MshA pilus and how c-di-GMP regulates the production and function of this complex .
CdgJ is a PDE and a cdgJ mutant displays a decrease in motility , enhanced VPS production , and increased biofilm formation compared to WT [36] . To begin investigating the mechanism by which CdgJ impacts motility , we performed transposon mutagenesis in a cdgJ deletion mutant ( ΔcdgJ ) and screened the resulting mutants for enhanced motility phenotype using LB soft agar motility assay . We screened 7054 transposon mutants and identified 42 extragenic suppressor mutants with increased motility and mapped the transposon insertions site to 22 different genes ( Table 1 , Fig 1 ) . As previously reported , we found that mutations in the DGC encoding genes cdgH and cdgK in a ΔcdgJ strain enhance motility [36] . We also found that insertion in the gene encoding a pilus retraction motor , PilT , and insertions into different genes predicted to be required in mannose-sensitive hemagglutinin type IV pilus ( MshA ) biogenesis restored swimming motility . To further investigate the suppression of a ΔcdgJ motility phenotype , we generated in-frame deletions of several of the genes identified in the transposon screen in wild-type and ΔcdgJ strains and analyzed the mutants for motility phenotype using LB soft agar motility assay . We focused on mshA ( encoding major pilin subunit ) , mshE ( encoding putative polymerizing ATPase ) , and pilT ( encoding putative depolymerizing ATPase ) as they are crucial for production and function of MshA pili . As previously reported , ΔcdgJ has a significant motility defect compared to the parental WT strain [36] ( Fig 2A ) . Deletion of mshA , mshE , and pilT in a ΔcdgJ background restored motility similar to the WT strain , confirming that those mutations mediate suppression of the flagellar motility defect in the ΔcdgJ mutant . Deletion of pilU ( predicted to encode a second copy of putative depolymerizing ATPase ) in ΔcdgJ had no effect on the motility compared to the parental ΔcdgJ mutant . Mutants of mshA , mshE , and pilT in a WT background were assayed for motility to determine if their enhanced motility in the ΔcdgJ background is dependent on this mutation , or if this could be a case of bypass suppression . The ΔmshA , ΔmshE , and ΔpilT mutants exhibited enhanced motility compared to the WT strain ( Fig 2A ) . The ΔpilU strain had a similar motility phenotype to the WT strain , suggesting that the functions or expression profiles of pilT and pilU are different . These data demonstrate that deletion of mshA , mshE , and pilT enhances motility regardless of the presence of a wild-type copy of cdgJ . However , we could not rule out the possibility that cdgJ could directly or indirectly control the production or function of the MshA pilus . To determine if mshA-mediated suppression of the flagellar motility phenotype is specific to the cdgJ mutation or if it occurs in other PDE deletion backgrounds , we mutated mshA in a ΔrocS strain . RocS has both GGDEF and EAL domains and is predicted to function mainly as a PDE as ΔrocS mutants have reduced motility along with enhanced VPS production and biofilm formation [36–38] . We generated a ΔrocSmshA double mutant and determined that this mutant exhibited a wild-type motility phenotype ( S1 Fig ) . These findings suggest that MshA negatively impacts V . cholerae flagellar motility and is involved in general c-di-GMP mediated repression of motility . To evaluate further the ability of MshA to repress motility , we analyzed the effect of the expression of a WT copy of mshA provided in trans in an expression plasmid with an IPTG-inducible promoter . Motility assays confirmed that expression of mshA , upon induction with IPTG , significantly reduced motility in a ΔmshA strain ( Fig 2B ) . Additionally , expression of mshA in the WT strain reduced motility , suggesting that overproduction of MshA impairs motility . IPTG had no effect on motility in strains harboring an empty vector control . Since the MshA pilus is critical for initial stages of surface attachment and subsequent biofilm formation [10 , 11 , 20] , we hypothesized that the ΔmshE and ΔpilT mutants would phenocopy the reduced biofilm phenotype of a ΔmshA mutant . Biofilms of these mutants were grown using a flow cell system , imaged using confocal microscopy , and analyzed using the COMSTAT image analysis software package to evaluate biofilm structural properties . As expected , the ΔmshA mutant attached poorly to the substrate and formed biofilms with low biomass ( Fig 3A and 3B ) . The ΔmshE and ΔpilT mutants grew biofilms with significantly less biomass , thickness , and substrate coverage than WT . COMSTAT analysis revealed that while biofilm biomass of ΔpilT and ΔmshA was similar , the biomass of ΔmshE was significantly less than the ΔpilT strain ( Fig 3B ) . Additionally , surface coverage of ΔpilT mutant was greater than that of ΔmshA and ΔmshE . These differences suggest that though the ΔmshA , ΔmshE , and ΔpilT mutants are deficient at forming biofilms , there are subtle differences in the phenotypes of the strains . The ΔpilU mutant produced biofilms that were indistinguishable from WT . Since our initial interest in the MshA pilus was sparked by the discovery that mutations in pilus genes can suppress the motility defect in mutants of the PDE cdgJ , we determined whether the MshA pilus mutations were epistatic to the cdgJ mutation . As previously reported , the ΔcdgJ strain produced thicker biofilms than the WT strain ( Fig 3C and 3D ) [36] . The ΔcdgJmshA , ΔcdgJmshE , and ΔcdgJpilT strains formed biofilms with significantly less biomass , thickness , and surface coverage than the ΔcdgJ strain . These findings are consistent with the hypothesis that these proteins are involved in formation of the MshA pilus and that a functional MshA pilus is required for biofilm formation , surface attachment , and the inhibition of flagellar motility . Biofilms formed by the ΔcdgJmshE mutant had a significant reduction in biomass , thickness , and surface coverage compared to the ΔcdgJmshA or ΔcdgJpilT strains . In addition to using bulk differences in biofilm formation as an index of the transition between motile and sessile behavior , we also directly monitored surface attachment of V . cholerae with single cell resolution using high-speed microscopy and cell tracking ( at 5 ms frame rate ) to elucidate how genes involved in MshA pilus production impact microscopic outcomes such as initial surface attachment . ΔmshA mutants do not attach to the surface in significant numbers , with no attached cells observed in the first 15min after inoculation . In contrast , well over 100 cells of the WT strain attach to the surface over the same time interval ( Fig 4A ) . Using the same metrics , the ΔmshE mutant was also unable to attach to the surface , exhibiting binding profiles that were similar to the ΔmshA strain . The ΔpilT strain was able to attach to the surface more than the ΔmshA; however , it is unable to attach with the same efficiency as the WT strain . This suggests the following hierarchy of behavioral categories for comparison with non-WT backgrounds: the strong binding strain ( WT ) , the intermediate binding strain ( ΔpilT ) , and the weak binding strains ( ΔmshA and ΔmshE ) . To dissect the origin of this aggregate statistical behavior of surface attachment , we needed more single-cell metrics for bacterial behavior near a surface . We examined the temporal aspects of single bacterium interactions with the surface in the form of single cell residence time ( number of seconds that each stationary cell remained associated with the surface , Fig 4B ) . The WT strain formed prolonged associations with the surface , with a mean residence time of 2 . 6 seconds and a maximum of 78 . 9 seconds . The ΔmshA , ΔmshE , and ΔpilT mutant strains demonstrated more transient interactions with the surface , with reduced mean residence times compared to WT ( 0 . 69 , 0 . 58 , and 0 . 58 seconds , respectively ) . This is most evident in the inset of Fig 4B , where the entire adherent populations of these mutants have residence times of less than 6 sec , while the WT residence times extend out to nearly 80 seconds . We used high-speed microscopy and near-surface cell tracking to record the trajectories of cells within one micrometer of the coverslip surface in a microscopy chamber ( Fig 5 ) . All strains with flagella are capable of swimming motility in 3D , and can exhibit trajectories that come in and out of focus . Consistent with previous reports , the WT strain exhibits orbiting and roaming behavior with regards to near surface motility [10] . All of the surface attached cells come from the orbiting subpopulation . The ΔmshA mutant does not show orbiting or roaming behavior , consistent with the model that MshA pili-surface interactions are responsible for these near-surface motility phenotypes . Moreover , also consistent with the model , the mutant shows greatly reduced surface attachment ( Fig 4A ) [10] . The ΔmshE strain tracks phenocopy ΔmshA , exhibiting predominantly a swimming phenotype with little observable attachment . The ΔpilT mutant has an intermediate phenotype; a few cells appear to exhibit behavior similar to WT orbiting , but has greatly reduced attachment compared to WT . These data are consistent with the biofilm and attachment data described in Figs 2 and 3 . To investigate the role of the PDE CdgJ on initial surface attachment , we observed mshA , mshE , pilT mutants in a ΔcdgJ background with high-speed microscopy using similar experiments . We determined that the ΔcdgJ mutant exhibits strong attachment to the surface , with a sharp increase in the number of attached cells as a function of time during the initial few minutes compared to WT . ( Fig 4C , time = 1 ) . This strain rapidly associated with the surface , with nearly all cells binding within the first two minutes of observation . As predicted from the biofilm and motility data ( Figs 1 and 2 ) , the ΔcdgJmshA , ΔcdgJmshE , and ΔcdgJpilT strains exhibited reduced surface attachment compared to the parental ΔcdgJ strain and WT ( Fig 4C ) . As in the WT background , the ΔcdgJpilT strain exhibited an intermediate level of attachment and the ΔcdgJmshA and ΔcdgJmshE strains were poor at attachment . The ΔcdgJmshA , ΔcdgJmshE , and ΔcdgJpilT strains exhibited short residence times with mean values of 0 . 43 , 0 . 53 , and 0 . 57 s , respectively ( Fig 4D ) . By contrast , the ΔcdgJ strain exhibits much longer mean surface residence times ( 3 . 59 s with and a maximum of 75 . 8 seconds ) , which surpass even WT ( Fig 4D , red vs black bars ) . These longer residence times indicate a strong tendency of the ΔcdgJ strain to associate with surfaces , which is also evident in the rapid decrease in density of tracks of the ΔcdgJ over time ( Fig 5 ) : since only motile cells are displayed in each image , the density of tracks diminishes over time in the ΔcdgJ strain due to the increasing proportion of adherent , nonmotile cells . These results demonstrate the ΔcdgJ strain adheres to surfaces more rapidly than WT , and that mshA is epistatic to cdgJ . The motility and biofilm phenotypes of the ΔmshE and ΔpilT mutants , combined with homology to known Type IV pilus motor proteins suggests that these genes are involved in the production of a functional MshA pilus . MshE shares 75% amino acid similarity with the Type IV extension ATPase PilB of P . aeruginosa and 77% similarity to the Type II extension ATPase EpsE of V . cholerae , suggesting that MshE is the extension ATPase of the MshA pilus ( S2 and S3 Figs ) . We investigated the role of these genes in the production of MshA pili using a surface MshA pilin ELISA ( Fig 6A ) , which detects only assembled pili . ΔmshA , ΔmshE , and ΔpilT mutants produced significantly less surface MshA pili than the WT strain . This supports the hypothesis that MshE and PilT are involved in the production of a functional MshA pilus . The ΔpilT strain produced significantly more surface pili than ΔmshA or ΔmshE , correlating with the intermediate phenotype of this strain observed in biofilm and near surface motility assays ( Fig 3A and 3A ) . Surface MshA pilus production was determined in deletion mutants of other genes in the secretory operon with multiple transposon insertions ( S4 Fig ) . Deletion of mshL , which encodes the putative outer membrane pore protein , resulted in no surface pili . While deletion of mshM ( predicted to encode an ATPase ) or mshN ( predicted to encode a tetratricopeptide repeat domain ) resulted in similar pilus production to the WT strain , pilus production was increased in a ΔmshI mutant , The specific mechanisms by which lack of these genes results in suppression of a motility defect in a ΔcdgJ strain is yet to be determined . The ΔcdgJ mutant produced significantly more surface MshA pili than WT . Surface MshA pili were reduced in ΔcdgJmshA , ΔcdgJmshE , and ΔcdgJpilT mutants to the level of the ΔmshA strain , indicating that none of these strains could produce a functional pilus . Whole cell western analysis indicated that all of the strains except ΔmshA and ΔcdgJmshA produced similar amounts of MshA protein , suggesting that the altered surface MshA is not a result of altered MshA production ( Fig 6B and 6C ) . These data support the hypothesis that a functional MshA pilus inhibits motility and enhances surface attachment and biofilm formation . This effect is likely due to the interactions between the pili and surfaces , however , there may be additional mechanisms affecting motility and surface attachment , as discussed below . Additionally , the elevated production of MshA pili by the ΔcdgJ mutant could explain the motility and biofilm phenotypes observed in this mutant . To confirm further that mshE was responsible for the lack of pili in a ΔmshE mutant , we generated chromosomal replacements at the native locus with either the WT mshE or a sequence encoding a mutation in the Walker A ATPase active site ( K329A ) ( S5 Fig ) . Pilus production was restored to WT levels when the WT sequence was inserted , however the K329A produced no detectable surface pili . These data confirm that MshE , and specifically an intact ATPase domain , are required for MshA pilus production . We also utilized transmission electron microscopy ( TEM ) to assess presence of MshA on the cell surface ( Fig 7 ) . We determined that WT produces several pili along the cell body ( range 2–5 ) that were about one half to one cell body length . These pili were not present in the mshA mutant , suggesting that the pili observed are in fact MshA pili . Similarly , the ΔmshE strain produced no visible pili , which is consistent with the prediction that MshE is the motor protein responsible for extension of the MshA pilus . These images reveal that there were no differences observed between pili produced by WT and the ΔpilT , ΔpilU and ΔcdgJ mutants . Due to the fragile nature of MshA pili , quantitative measurements were not possible with TEM though these images observing the presence or absence of MshA pili support the quantitative measurements observed by ELISA . The pilus motor proteins MshE and PilT contain Walker A ATPase domains , which are utilized to energize the assembly and disassembly of pili . Baraquet et al . demonstrated that the activity of the P . aeruginosa regulator FleQ is modulated by binding c-di-GMP at its Walker A site [39] . We hypothesized that one or both of the Msh pilus motor proteins could function as a c-di-GMP receptor . Isothermal calorimetry was utilized to investigate the interaction of these proteins with c-di-GMP . We determined that MshE binds c-di-GMP , while PilT and PilU were unable to bind c-di-GMP ( Fig 8A ) . ATPase activity of purified MshE , PilT , and PilU confirmed that these preparations contain functional protein ( S6 Fig ) . VpsT was purified and included as a positive control , as it has been demonstrated to bind c-di-GMP [40] . Fitting the data to a single binding site model indicates that MshE has a slightly lower affinity for c-di-GMP ( K = 1 . 14x105 ± 3 . 24x104 M-1 ) than VpsT ( K = 9 . 9x104 ± 2 . 06x104 M-1 ) . These data suggest that c-di-GMP does affect MshA pilus production through interactions with MshE . Since MshE , PilT , and PilU contain conserved Walker A ATPase domains , but only MshE bound c-di-GMP , we purified the N terminal domain of MshE ( amino acids 1–180 ) for analysis of the interactions with c-di-GMP . This domain lacks homology with either PilT or PilU and therefore is a likely candidate for binding c-di-GMP . Fluorescence thermal shift assays were utilized to determine the interaction between MshE or MshE-N terminal domain and nucleotide . Purified proteins are stabilized by a bound ligand; therefore the temperature midpoint of unfolding ( Tm ) of the protein in the presence of ligand is higher than the Tm of the protein in buffer [41 , 42] . We observed that the Tm of full-length MshE in buffer was 38 . 43 ± 0 . 52°C ( Fig 8B ) . When full-length MshE was incubated with ATP or c-di-GMP , the Tm increased to 40 . 15 ±1 . 02°C and 42 . 77±0 . 70°C , respectively . This indicates that full-length MshE binds both ATP and c-di-GMP . The N-terminal domain of MshE had a Tm of 66 . 36 ± 0 . 79°C in buffer . There was no increase in Tm in the presence of ATP ( Tm = 62 . 72 ±0 . 46°C ) . In contrast , incubation of the MshE N-terminal domain with c-di-GMP increased the Tm to 70 . 18 ±0 . 55°C , indicating that this domain binds c-di-GMP ( Fig 8C ) . Additionally , these interactions are specific , as neither the full-length nor the N-terminus of MshE bind to cAMP , c-di-AMP , GTP , or cGMP in thermal shift assays ( Fig 8B and 8C ) . To further evaluate the link between c-di-GMP production and pilus assembly , we performed a surface pilin ELISA to determine the production of MshA pili over a range of c-di-GMP concentrations ( Fig 9 ) . To generate a range of c-di-GMP concentrations , we introduced a construct harboring an IPTG-inducible copy of the DGC VCA0956 ( Ptac0956 ) into the chromosome of the V . cholerae O1 El Tor strain A1552 . This strain was grown in the presence of varying concentrations of IPTG , followed by detection of extracellular MshA with the surface pilin ELISA . We observed that upon induction of the expression of VCA0956 with IPTG , there was an increase in intracellular c-di-GMP ( Fig 9 , Grey bars ) . Additionally , the surface pilin ELISA detected increased amounts of extracellular MshA that coincided with the increased c-di-GMP ( Fig 9 , Black bars ) , though the total MshA production was not increased ( S7B Fig ) suggesting that the increased surface pilin is due to enhanced assembly . There was a significant correlation between elevated c-di-GMP and increased surface MshA ( Pearson correlation , p = 0 . 0025 , R2 = 0 . 9199 ) . Increased c-di-GMP resulted in reduced motility and increased biofilm maximum thickness compared to WT ( S7 Fig ) . Collectively , these data indicate that c-di-GMP promotes assembly of the MshA pilus . Induction of VCA0956 in a ΔmshE strain does not increase the production of surface MshA pili , further supporting that MshE is required for the c-di-GMP-dependent increase of MshA pili ( S8 Fig ) .
High c-di-GMP levels inhibit motility of bacteria and studies have highlighted some of the mechanisms involved . c-di-GMP can repress transcription of flagellar genes , or can act post-transcriptionally to regulate flagellar reversals by interactions with particular flagellar motor proteins or by altering the chemotaxis signal transduction system [26–29 , 43–45] . Our work revealed a role for c-di-GMP in the regulation of MshA pilus production with effects on near surface motility , motile to sessile transition , and biofilm formation via a post-translational mechanism . c-di-GMP also promotes the assembly and activity of Type IV pili in P . aeruginosa [34 , 46] . In this example , ectopic expression of a DGC results in elevated amounts of c-di-GMP and increased surface pili . We demonstrate that production of the MshA pilus in V . cholerae is increased in response to high concentrations of c-di-GMP . These similar results in two organisms suggest that c-di-GMP may promote type IV pili assembly and activity as a more general mechanism of pilus regulation than previously identified . Many bacteria rely on type IV pili for motility and attachment , so integration of c-di-GMP in post-translational control of these structures could be a conserved mechanism across species . The V . cholerae genome encodes 31 GGDEF domain proteins but we have found that only a subset of these impact motility , biofilm formation , or both . We previously reported that four DGCs CdgH , CdgK , CdgL , and CdgD and two PDEs CdgJ and RocS affect motility . Furthermore , a strain lacking all four DGC-encoding genes ( ΔcdgDΔcdgHΔcdgKΔcdgL ) has a markedly high motility phenotype , suggesting the effect of these proteins on motility is additive [36] . Screening for suppressor mutants of ΔcdgJ that restored swimming motility identified CdgH and CdgK , suggesting that c-di-GMP produced by these DGCs could be a substrate for CdgJ . To test if CdgJ and DGCs that control motility physically interact , we analyzed interactions of CdgJ with CdgH , CdgK , CdgL , and CdgD using the commercially available Bacterial Adenylate Cyclase Two-Hybrid System ( BACTH ) ( Euromedex Strasbourg , France ) . We did not observe any interaction between DGCs and CdgJ or between CdgJ and MshE or PilT , suggesting that these proteins do not need to be in physical contact , or the interaction is too weak or brief to be detected with this method . Pili are dynamic structures that are generated by the assembly and disassembly of pilin subunits . The motor proteins responsible for this process have been characterized in many other systems . The two putative motor proteins of interest in this report , MshE and PilT , were identified based on homology to the PilT motor protein in P . aeruginosa . These proteins belong to AAA+ ATPase family proteins and are responsible for energizing the addition and disassembly of pilin subunits . Previous studies revealed presence of a bacterial AAA+ ATPase enhancer binding class of c-di-GMP receptors [28] . Here , we demonstrate that MshE binds to c-di-GMP . This is an important finding , as it establishes that ATPases beyond enhancer binding proteins are also capable of binding c-di-GMP . We note that PilT and PilU , which both harbor AAA+ ATPase domain , are unable to bind to c-di-GMP under the conditions tested . We also showed that N-terminal domain of MshE , which is not present in PilT and PilU is capable of binding to c-di-GMP . Future studies will elucidate the effect of c-di-GMP binding by MshE , as well as the specific mechanisms of this interaction . We have also characterized several proteins necessary for the production of a functional MshA pilus . Although the function of members of the MshA operons were predicted based on homology to proteins of known function in other bacteria , function of these genes in MshA biogenesis were not characterized [22] . We have demonstrated that MshE is responsible for the assembly of MshA pilin subunits into a functional pilus . These data indicate that the decreased attachment and biofilm phenotypes , as well as the enhanced motility of the mshA , mshE and pilT mutants relies on the dynamic nature of the MshA pilus . If the presence of a pilus could enhance attachment and biofilm formation , a ΔpilT strain would in principle produce biofilms that had similar , or even greater biomass than the WT strain . The observation that the ΔpilT strain phenocopies ΔmshA and ΔmshE , which lack extracellular pili , suggests that both extension and retraction of the pili are critical for normal substrate attachment . The ELISA indicates that the ΔpilT strain produces fewer pili than WT , though the TEM images confirm that there are pili on the surface . Future studies will further investigate the role of PilT in the retraction of the MshA pilus . It is important to note that PilT has already been characterized as the retraction pilus of the ChiRP pilus ( chitin-induced competence ) , suggesting that PilT can function in more than one type IV pilus system [47] . This could be a mechanism for genomic conservation , where one promiscuous retraction ATPase is encoded , while several extension ATPases allow for specificity of the system in regulation . Future studies will address the mechanism of co-regulation of these systems to determine whether pilT is expressed constitutively while the specific extension ATPases are regulated , or if there is overlap in the regulation between the extension and retraction ATPases . Several studies have investigated how bacteria sense and respond to surfaces , often times by rapidly upregulating production of adhesins and polysaccharides [48–50] . This regulation is typically mediated by c-di-GMP [23 , 28 , 40 , 50 , 51] . Many bacteria also regulate motility in response to surfaces . E . coli uses the resistance to flagellar rotation as a mechanosensor and adapt by adding force-generating motor subunits to the stator complex [43] . This allows the bacterium to adjust the force of flagellar rotation to match the viscosity of the environment . An additional example of “stator swapping” to modulate flagellar force was recently published for P . aeruginosa [44] . In this example , c-di-GMP represses motility by excluding the swarming proficient MotC/D proteins from the stator complex in favor of the swarming deficient MotA/B proteins . P . aeruginosa also utilizes the altered chemotaxis protein WspA as a surface sensor , which results in production of c-di-GMP by the DGC WspR upon interaction with a surface [52 , 53] . In B . subtilis , production of flagella , and therefore swarming motility , is regulated by Lon-dependent proteolysis of the master regulator of flagellar biosynthesis SwrA upon surface contact [54]; and flagellar function is modulated by EpsE which synergizes exopolysaccharide biosynthesis with flagellar motility by acting as a clutch through interaction with the flagellar protein FliG to limit rotation and therefore motility [55] . This study presents a possible mechanism for how c-di-GMP production affects motility and biofilm formation through modulating MshA pilus production . Both pili and flagella contribute to near surface motility and initial attachment [10] . Besides the generation of near-surface motility modes conducive to surface attachment , it is known that van der Waals forces depend crucially on the extent of surface contact [56] . That V . cholerae has a comma-like helicoid shape with smaller surface contact areas implies that adhesive forces between the surface and the cell body will be decreased relative to more cylindrically symmetrical species such as P . aeruginosa for most cell orientations , so adhesive contributions from appendages like MSHA to ‘anchor’ the cell on surface will be comparatively more important . In fact , recent measurements of TFP adhesive forces show that they can be quite strong , in the hundreds of pico-Newton ( pN ) range , and are surface chemistry dependent , in agreement with our results [57] . That V . cholera select for surfaces that interact with MSHA strongly ( and thereby generate ‘orbiting’ motility ) implies that the existence of more functional MSHA induced by c-di-GMP can better anchor a cell mechanically and mitigate against flagellum driven motion . This work further strengthens the notion that there is a mechanistic link between c-di-GMP and initial attachment through modulation of flagellar motility and pilus activity .
The bacterial strains and plasmids used in this study are listed in S1 Table . All V . cholerae and Escherichia coli strains were grown aerobically , at 30°C and 37°C , respectively , unless otherwise noted . All cultures were grown in Luria-Bertani ( LB ) broth ( 1% Tryptone , 0 . 5% Yeast Extract , 1% NaCl ) , pH 7 . 5 . LB agar medium contains 1 . 5% ( wt/vol ) granulated agar ( BD , Sparks , MD ) . Concentrations of antibiotics and inducers used , where appropriate , were as follows: ampicillin , 100 μg/ml; rifampicin , 100 μg/ml; gentamicin , 50 μg/ml , kanamycin , 50 μg/ml , and arabinose , 0 . 2% ( w/v ) , 6 . 25–400μM IPTG . In-frame deletion and GFP-tagged strains were generated according to protocols previously published [13 , 14] . DNA manipulations were carried out by standard molecular techniques according to manufacturer’s instructions . Restriction and DNA modification enzymes were purchased from New England Biolabs ( Ipswitch , MA ) . Polymerase chain reactions ( PCR ) were carried out using primers purchased from Bioneer Corporation ( Alameda , CA ) and the Phusion High-Fidelity PCR kit ( New England Biolabs , Ipswitch , MA ) , unless otherwise noted . Sequences of the primers used in the present study are available upon request . Sequences of constructs were verified by DNA sequencing ( UC Berkeley DNA Sequencing Facility , Berkeley , CA ) . A region encompassing the PlacIq-lacI and the Ptac promoter elements was amplified from the pMAL-c5x plasmid ( New England Biolabs , Ipswitch , MA ) by PCR . The amplified product was joined by overlapping PCR to amplicons of ~500 bp that correspond to sequences upstream and downstream of the VCA0956 translational start site . The resulting amplicon was cloned into the suicide plasmid pGP704sacB28 and mobilized into Vibrio cholerae A1552 by biparental mating . The selection of double recombinants with the desired insertion of the PlacIq-lacI and Ptac promoter elements was performed as described in [14] . Sequences of constructs were verified by DNA sequencing ( UC Berkeley DNA Sequencing Facility , Berkeley , CA ) . To generate a library of transposon mutants , V . cholerae O1 El Tor strain A1552 ΔcdgJ was conjugated with the donor E . coli S-17-l λpir containing the Mariner transposon on the pSC189 backbone [58] . Transconjugants were selected on LB agar containing kanamycin 50μg/ml and rifampicin 100μg/ml . A total of 7054 mutants were isolated and screened for motility phenotypes on LB soft agar ( 0 . 3% ) motility plates . Motility plates consist of LB containing 0 . 3% agar supplemented with 100μM IPTG where appropriate . Plates were poured and allowed to dry at room temperature for 4 h prior to inoculation . Colonies from overnight LB agar plates grown at 30°C were transferred to motility plates and incubated for 16 h at 30°C . Motility diameter was measured and normalized to the average of WT on each plate . Experiments were performed with three biological replicates in triplicate and data were analyzed with a Oneway ANOVA followed by Dunnett’s multiple comparison test . Inoculation of flow cells was done by diluting overnight-grown cultures to an OD600 of 0 . 04 and injecting into a μ-Slide VI0 . 4 ( Ibidi , Martinsried , Germany ) . To inoculate the flow cell surface , bacteria were allowed to adhere at room temperature for 1 h . Flow of 2% v/v LB ( 0 . 02% tryptone , 0 . 01% yeast extract , 1% NaCl; pH 7 . 5 ) was initiated at a rate of 7 . 5 ml/h and continued for 24 h . Confocal images were obtained on a Zeiss LSM 5 PASCAL Laser Scanning Confocal microscope . Images were obtained with a 40X dry objective and were processed using Imaris ( Bitplane , Zurich , Switzerland ) . Quantitative analyses were performed using the COMSTAT software package [59] . Statistical significance was determined using Oneway ANOVA with Dunnett’s Multiple Comparison test . Three biological replicates were performed in triplicate . Images presented are from one representative experiment . Bacteria were cultured in full strength Luria–Bertani ( LB ) broth overnight under shaking at 30°C . Immediately prior to inoculation , cultures were diluted into 2% LB ( containing 171 mM NaCl ) to an OD600 0 . 01–0 . 03 . The V . cholerae cells were then injected into a sterile flow-cell containing the same media and imaged immediately . Imaging was performed with a Phantom V12 . 1 high-speed camera ( Vision Research ) collecting ~20 , 000 bright-field images at 5 ms resolution with a 100x oil objective on an IX71 Olympus microscope . All movies were recorded at the same frame rate , for the same duration , after the first , third and 15th minute post inoculation . For cell-tracking algorithms and analysis protocol , every frame of a movie was preprocessed in Matlab ( Mathworks ) by subtracting the background , scaling , smoothing and thresholding . Image processing this way causes the bacteria appear as bright regions . Tracking is done by locating all bright objects that overlap objects in the next frame by combining the two frames into a three-dimensional ( 3D ) matrix and then by locating 3D connected components . Results are stored in a tree-like data structure with multiple roots; every newly detected bacterium that appears is recorded as a ‘root’ of the tree . When bacteria interact , they are recorded as a ‘node’ of the tree; when they depart , they are recorded as a ‘leaf’ . Each root or node stores the sequence of pixel lists that comprise the bacterium in all frames until the next interaction or detachment event . We measure the instantaneous shape properties of the bacteria using the Matlab regionprops function [10] . Surface pili composed of MshA were quantified using an ELISA based on a previously published protocol [46] . Briefly , overnight culture was diluted 1:100 in fresh LB medium and grown to OD600 0 . 5 at 30°C . Cells ( 125μL ) were added to a 96-well plate ( Greiner Bio-One , Monroe , NC ) and incubated at 30°C for one hour . Cells were fixed with 100μL of methanol for 10 minutes at room temperature , then washed twice with PBS . Samples were blocked in 5% nonfat dry milk and immunoblotted with polyclonal rabbit anti-MshA ( 1:1000 dilution , gift of J . Zhu ) and horseradish peroxidase ( HRP ) -conjugated secondary antibody ( Santa Cruz Biotechnology , Santa Cruz , CA ) . After three washes in PBS , 100μL of TMB ( eBioscience , San Diego , CA ) was added and incubated for 30 minutes at room temperature followed by the addition of 100μL of 2N H2SO4 . Absorbance was recorded at 490nm and the samples were normalized to the change in WT . Three biological replicates were assayed in triplicate and statistical significance was determined with a Oneway ANOVA followed by a Dunnett’s Multiple Comparison test . Samples were grown to mid-exponential phase ( OD6000 . 5 ) in LB or LB with IPTG . Cells were collected via centrifugation and cell pellets were resuspended in 2% SDS and boiled for 5 minutes . Lysates were cleared via centrifugation and total protein was quantified via BCA assay ( Pierce , Rockford , IL ) . Two hundred μg of protein was separated on a 12% SDS PAGE gel and transferred to a PVDF membrane using a semi-dry transfer apparatus ( Bio-Rad , Hercules , CA ) . Blots were blocked in 5% nonfat dry milk and immunoblotted with polyclonal rabbit anti-MshA ( 1:2000 dilution , gift of J . Zhu ) and horseradish peroxidase ( HRP ) -conjugated secondary antibody ( Santa Cruz Biotechnology , Santa Cruz , CA ) . Chemiluminescence was detected with the SuperSignal West Pico reagents ( Pierce , Rockford , IL ) on the ChemiDoc MP Imager ( Bio-Rad Hercules , CA ) . Densitometry was performed using the Image Lab software v4 . 0 . 1 ( Bio-Rad , Hercules , CA ) using the band in the WT lane as a reference . Blots were performed in triplicate for densitometry analysis and a representative image is shown . Bacteria were prepared for electron microscopy by inoculating a single colony of each V . cholerae strain in LB broth grown overnight at 30°C with shaking at 200 rpm after which each culture was diluted 1:100 in LB broth and allowed to grow similarly to OD 0 . 4 . An aliquot of each culture was diluted to yield an optical density of 0 . 1–0 . 2 and then applied to a 300 mesh carbon-coated Formvar grid ( Electron Microscopy Sciences , Hatfield , PA ) . After 2 minutes , each grid was washed five times with deionized water , and negatively stained with 2% ( w/v ) aqueous uranyl acetate solution for 90 seconds . Imaging was performed with a JEOL JEM-1400 transmission microscope . E . coli BL21 harboring plasmids for gene expression were grown to an OD600 of 0 . 4 at 30°C in LB containing 100μg/mL ampicillin . Cultures were shifted to 18°C and IPTG was added to a final concentration of 100μM . 16h post induction , cells were harvested by centrifugation at 10 , 000 x g for 15 minutes and stored at -80°C . Cell pellets were resuspended in GST Lysis Buffer ( 50mM Tris ( pH 8 . 0 ) , 1M NaCl , 0 . 5% Tween-20 containing PI cocktail tablets ( Roche Life Science , Indianapolis , IN ) . Cells were lysed by sonication and cell lysate was cleared via centrifugation . Cleared lysate was loaded onto GST FPLC column as follows . GSTPrep FF16/10 column ( GE Healthcare , Piscataway , NJ ) was equilibrated in lysis buffer at 1mL/minute using a BioLogic DuoFlow FPLC system ( Bio-Rad , Hurcules , CA ) . Sample was loaded and washed with 1 column volume of GST Lysis Buffer ( 20mL ) . Subsequent washes were performed with 20mL of Wash Buffer 2 ( 50mM Tris ( pH 8 . 0 ) , 0 . 25M NaCl , 0 . 5% Tween-20 , 0 . 5mM DTT ) and 3 ( 50mM Tris ( pH 8 . 0 ) , 0 . 25M NaCl , 0 . 5mM DTT ) . Bound protein was eluted with 80mL of GST Elution Buffer 4 ( 50mM Tris ( pH 8 . 0 ) , 0 . 25M NaCl , Glutathione 1 . 5g/L ) and collected in 15mL fractions . These fractions were pooled and concentrated to approximately 10mL using an Amicon 10KDa cutoff spin fliter ( EMD Millipore , Darmstadt , Germany ) . Samples were dialyzed against ITC buffer ( 25mM Tris-HCl , 150mM NaCl , 250μM DTT , pH 7 . 5 ) overnight using 12 kDa cutoff dialysis tubing ( Fisherbrand , Pittsburgh , PA ) . An aliquot of dialyzed protein was diluted in 6M guanidinium HCl and concentration determined via A280 . MshE ( 18 . 9μM ) , PilT ( 20 . 2μM ) , PilU ( 21 . 1μM ) , VpsT ( 19 . 5μM ) and c-di-GMP ( 250μM ) were prepared in 25mM TrisHCl pH 7 . 5 , 150mM NaCl , and 200μM DTT and degassed prior to analysis . ITC was performed in with a VP-ITC ( MicroCal , Northampton , MA ) with the following parameters: 3 initial injections of 2μL followed by 40 injections of 10μL spaced at 180 seconds . The data were normalized to a run injecting c-di-GMP into buffer to account for the heat of dilution . Data were processed in Origin v7 . 0 software ( OriginLab , Northampton , MA ) and fit to a single site model . Thermal shift assays were performed as previously described with modifications [41 , 42] . Briefly , purified protein was added to the reaction to a final concentration of 3μM in the presence or absence of 2mM concentration of the indicated nucleotide in buffer ( 25Mm TrisHCl pH 7 . 5 , 100mM NaCl , 1:1000 dilution of SYPRO Orange Dye ( Invitrogen ) , and 0 . 2mM MgCl2 . A melt curve protocol was run on an Applied Biosystems ViiA7 qPCR instrument . The fluorescence was measured using the ROX reporter with a temperature gradient of 20–95°C in 0 . 5°C increments at 30 second intervals . Melt curve data were trimmed to three data points after maximum and the data were plotted with Boltzmann model to obtain the temperature midpoint of unfolding ( Tm ) of the protein in each condition using Prism 5 . 0 software ( GraphPad ) . The fluorescence baseline of each sample was normalized to the buffer control for visualization purposes . Three biological replicates were assayed in triplicate and statistical significance was determined with a Oneway ANOVA followed by a Dunnett’s Multiple Comparison test . ATPase activity of purified proteins was determined by measuring the production of inorganic phosphate from ATP using the Enzchek Phosphate Assay Kit ( Invitrogen ) . The standard reaction mixture was prepared with the addition of 2mM MgCl2 , 10mM KCl , and 1mM DTT . Purified protein in buffer ( 25mM TrisHCl pH 7 . 5 , 100mM NaCl ) was added to the standard reaction mixture to a final concentration of 5μM . After a 10 minute incubation at room temperature , ATP was added to a final concentration of 10mM and reactions were incubated at 22°C for 30 minutes . Production of inorganic phosphate was monitored by reading OD360 and compared to a standard curve of solutions of KH2PO4 . The data are reported as specific activity ( nmol Pi/min/mg of protein ) . BSA was included as a negative control . Three independent experiments were run in triplicate . c-di-GMP extraction was performed as described previously [36] . Briefly , 40 ml of culture grown to OD600 ~0 . 4 was centrifuged at 3220 x g for 30 min . Cell pellets were allowed to dry briefly then re-suspended in 1 ml extraction solution ( 40% acetonitrile , 40% methanol , 0 . 1% formic acid , 19 . 9% water ) , and incubated on ice for 5 min . Samples were then centrifuged at 16 , 100 g for 5 min and 800 μl of supernatant was dried under vacuum and lyophilized . Samples were re-suspended in 50 μl of 184 mM NaCl and analyzed by liquid chromatography-tandem mass spectrometry ( LC-MS/MS ) on a Thermo-Electron Finnigan LTQ mass spectrometer coupled to a surveyor HPLC ( Thermo , Waltham , MA ) . The Synergin Hydro 4u Fusion-RP 80A column ( 150 mm x 2 . 00 mm diameter; 4-μm particle size ) ( Phenomenex , Torrance , CA ) was used for reverse-phase liquid chromatography . Solvent A was 0 . 1% acetic acid in 10 mM ammonium acetate , solvent B was 0 . 1% formic acid in methanol . The gradient used was as follows: time ( t ) = 0–4 min , 98% solvent A , 2% solvent B; t = 10–15 minutes , 5% solvent A , 95% solvent B . The injection volume was 20 μl and the flow rate for chromatography was 200 μl/minutes . The amount of c-di-GMP in samples was calculated with a standard curve generated from pure c-di-GMP suspended in 184 mM NaCl ( Biolog Life Science Institute , Bremen , Germany ) . Concentrations used for standard curve generation were 50 nM , 100 nM , 500 nM , 2 μM , 3 . 5 μM , 5 μM , 7 . 5 μM , and 10 μM . The assay is linear from 50 nM to 10 μM with an R2 of 0 . 999 . c-di-GMP levels were normalized to total protein per ml of culture . To determine protein concentration , 4 ml from each culture was pelleted , the supernatant was removed , and cells were lysed in 1 ml of 2% sodium dodecyl sulfate . Total protein in the samples was estimated with BCA assay ( Pierce , Rockford , IL ) using bovine serum albumin ( BSA ) as standards . Each c-di-GMP quantification experiment was performed with four biological replicates . Levels of c-di-GMP were compared to WT with Oneway ANOVA followed by a Dunnett’s Multiple Comparison test .
|
The human pathogen Vibrio cholerae causes the debilitating disease cholera through ingestion of contaminated food and water . V . cholerae is a natural inhabitant of aquatic environments . Transmission of V . cholerae to the human host is dependent on survival of the pathogen in aquatic reservoirs where it is challenged with many stressors , including changes in the physiochemical parameters of environments and predation by protozoa and phages . One method utilized to endure these assaults is to form a multicellular community called a biofilm . The signaling molecule cyclic dimeric guanosine monophosphate ( c-di-GMP ) is utilized to induce biofilm formation in many bacteria , including V . cholerae . We demonstrate that c-di-GMP promotes the production of a cell surface structure called MshA pili by binding the molecular motor responsible for polymerizing pilus subunits . These pili are adhesive appendages that are essential for attachment to surfaces . This study identifies a novel mechanism for c-di-GMP regulation of pilus production through interactions with the molecular motor responsible for pilus assembly . Since many bacteria utilize pili for attachment to surfaces and c-di-GMP as a pro-biofilm signaling molecule , the mechanism for pilus regulation and biofilm formation described here may be widespread among many pathogens .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
C-di-GMP Regulates Motile to Sessile Transition by Modulating MshA Pili Biogenesis and Near-Surface Motility Behavior in Vibrio cholerae
|
Mobile group II introns consist of a catalytic intron RNA and an intron-encoded protein with reverse transcriptase activity , which act together in a ribonucleoprotein particle to promote DNA integration during intron mobility . Previously , we found that the Lactococcus lactis Ll . LtrB intron-encoded protein ( LtrA ) expressed alone or with the intron RNA to form ribonucleoprotein particles localizes to bacterial cellular poles , potentially accounting for the intron's preferential insertion in the oriC and ter regions of the Escherichia coli chromosome . Here , by using cell microarrays and automated fluorescence microscopy to screen a transposon-insertion library , we identified five E . coli genes ( gppA , uhpT , wcaK , ynbC , and zntR ) whose disruption results in both an increased proportion of cells with more diffuse LtrA localization and a more uniform genomic distribution of Ll . LtrB-insertion sites . Surprisingly , we find that a common factor affecting LtrA localization in these and other disruptants is the accumulation of intracellular polyphosphate , which appears to bind LtrA and other basic proteins and delocalize them away from the poles . Our findings show that the intracellular localization of a group II intron-encoded protein is a major determinant of insertion-site preference . More generally , our results suggest that polyphosphate accumulation may provide a means of localizing proteins to different sites of action during cellular stress or entry into stationary phase , with potentially wide physiological consequences .
Mobile group II introns , found in bacterial and organelle genomes , consist of a catalytic intron RNA and a multifunctional intron-encoded protein ( IEP ) , which interact to promote RNA splicing and intron mobility [1 , 2] . The IEP binds to the intron in unspliced precursor RNA , promotes its splicing by stabilizing the catalytically active RNA structure , and remains tightly bound to the excised intron lariat RNA in a ribonucleoprotein particle ( RNP ) that promotes intron mobility . Mobility occurs by a remarkable mechanism in which the intron RNA inserts ( reverse splices ) directly into a DNA strand and is reverse-transcribed by the IEP , with the primer being either the opposite DNA strand cleaved by the IEP or a nascent strand at a DNA replication fork ( reviewed in [1] ) . By using this reverse splicing mechanism , mobile group II introns insert at high frequency into specific DNA target sites ( “retrohoming” ) and at low frequency into ectopic sites that resemble the normal homing site ( “retrotransposition” ) [1] . The latter process has led to the wide dispersal of mobile group II introns among bacterial species and also may have been used to invade eukaryotic nuclear genomes , where mobile group II introns are thought to have evolved into spliceosomal introns and non-long-terminal-repeat retrotransposons [2] . Although their DNA integration mechanism has been elucidated , little is known about how mobile group II introns function in a cellular context or about how their mobility is influenced by other cellular processes . The Lactococcus lactis Ll . LtrB intron , which has been used as a model system , is highly mobile not only in L . lactis but also in a variety of other bacteria including Escherichia coli , where it has been studied extensively by using the facile genetic and biochemical methods available for that organism [1 , 3] . The broad host range of the Ll . LtrB intron reflects that RNPs containing only the IEP and intron RNA by themselves carry out the initial reverse splicing and reverse transcription reactions , while the subsequent cDNA integration steps use common host DNA repair functions [4 , 5] . Ll . LtrB and other mobile group II introns recognize their DNA target sequences by using both the IEP and the base pairing of the intron RNA , with the latter contributing most of the target specificity [6–8] . For Ll . LtrB , these base-pairing interactions involve intron RNA sequences denoted EBS2 , EBS1 , and δ and complementary DNA target sequences denoted IBS2 , IBS1 , and δ' , extending from position −12 to +2 from the intron-insertion site . ( EBS and IBS denote exon- and intron-binding site , respectively . ) Because the DNA target site is recognized largely by base pairing of the intron RNA , Ll . LtrB can be retargeted to insert ( retrohome ) into different chromosomal sites , enabling its development into a gene targeting vector ( “targetron” ) [6 , 8 , 9] . Further , an Ll . LtrB intron with randomized EBS2 , EBS1 , and δ sequences inserts at sites distributed throughout the E . coli genome , analogous to global transposon mutagenesis . Surprisingly , however , we found that although Ll . LtrB could be retargeted to insert efficiently into any region of the E . coli genome , an Ll . LtrB intron with randomized EBS and δ sequences shows a strong proclivity to insert at sites clustered around the bidirectional replication origin ( oriC ) , with 57% of the sites lying within 5% of the chromosome on either side of oriC [10] . Coros et al . [11] studying retrotransposition of the wild-type Ll . LtrB intron into ectopic sites in E . coli observed a similar clustering of insertion sites in both the oriC and the ter regions . Previously , we hypothesized that the preferential insertion of the Ll . LtrB intron into the oriC and ter regions of the E . coli chromosome might reflect the intracellular localization of Ll . LtrB RNPs [12] . In E . coli , the oriC and ter regions are localized near the cellular poles during much of the cell cycle [13 , 14] , and we found by using both LtrA fusions with green fluorescent protein ( GFP ) and immunofluorescence microscopy that LtrA expressed alone or with Ll . LtrB RNA to form RNPs localizes to the cellular poles in both E . coli and L . lactis [12] . Further analysis in E . coli showed that the bipolar localization of LtrA occurs over a wide range of cellular growth rates and LtrA expression levels , is independent of oriC function , and occurs in anucleate cells , suggesting that LtrA simply is not forced to the poles by nucleoid occlusion [12] . We also found that LtrA expression in E . coli interferes with the polar localization of coexpressed Shigella flexneri IcsA protein , suggesting competition for a common localization determinant [12] . Beauregard et al . [15] found that the polar localization of LtrA is maintained in E . coli mutants with defects in nucleoid condensation , chromosome partitioning , and DNA replication , and as expected from its continued pole localization , Ll . LtrB retrotransposition sites remained clustered in the oriC and ter regions in such mutants . While the above findings are consistent with the possibility that the bipolar localization of LtrA is responsible for the clustering of Ll . LtrB-insertion sites in the oriC and ter regions of the E . coli chromosome , to prove this connection , it is necessary to obtain mutations that change LtrA's intracellular localization and show that they correspondingly change the chromosomal distribution of Ll . LtrB-insertion sites . However , mutations that affect the polar localization of proteins in E . coli have been difficult to find . Nilsen et al . [16] manually screened ∼7 , 000 E . coli mutants for altered localization of the S . flexneri IcsA protein . The only mutant identified was in the mreB gene , which encodes a bacterial actin homologue required for maintenance of the cell's rod shape , and in this case , an IcsA/GFP fusion protein simply showed multiple foci in a portion of the spherical mutant cells instead of two foci at poles . Here , we used automated fluorescence microscopy of cell microarrays [17] to screen a transposon-insertion library for mutants with altered LtrA localization . We identified five E . coli genes ( gppA , uhpT , wcaK , ynbC , and zntR ) whose disruption leads to both a more diffuse intracellular distribution of LtrA and a more uniform genomic distribution of Ll . LtrB-insertion sites , indicating that group II intron protein localization is a major determinant of insertion-site preference . Surprisingly , we find that the common factor affecting LtrA localization in these disruptants is the accumulation of intracellular polyphosphate poly ( P ) . We confirmed this connection by analyzing ppx disruptants , which lack the exopolyphosphatase that degrades poly ( P ) , and found that disruptants that accumulate poly ( P ) also show delocalization of other pole-localized basic proteins ( Neurospora crassa CYT-18 and E . coli XapR ) . Our findings show that poly ( P ) , which accumulates in response to cell stress or entry into stationary phase , can localize proteins to different sites of action , with potentially wide physiological consequences .
To screen for E . coli mutants with altered LtrA localization , we constructed a library of mariner transposon insertions in E . coli HMS174 ( DE3 ) , a standard host strain that contains an integrated λDE3 prophage with an isopropyl β-d-1-thiogalactopyranoside ( IPTG ) -inducible T7 RNA polymerase for Ll . LtrB intron expression ( see Materials and Methods section ) . For screening , the library was transformed with intron-donor plasmid pACD2X-GFP/LtrA , which uses a T7lac promoter to express an Ll . LtrB-ΔORF intron with short flanking exons , plus a GFP/LtrA fusion protein downstream of the 3′ exon ( E2; Figure 1A ) . This configuration gives high intron mobility frequencies and permits independent manipulation of the intron RNA and IEP . We showed previously that GFP/LtrA expressed from this plasmid is active in promoting both RNA splicing and intron mobility [12] . The library in 96-well plate format was arrayed onto microscope slides and screened for mutants with altered GFP/LtrA localization by automated fluorescence microscopy ( Figure 1B; see Materials and Methods section ) . The images were stored and then examined individually to characterize GFP/LtrA localization patterns ( Figure 1C and 1D ) . A total of 9 , 600 disruptants were screened under two different Ll . LtrB induction conditions ( overnight with 500 μM IPTG at 30 °C or 100 μM IPTG at 37 °C ) . Of 277 initial candidates , 36 showed similarly altered GFP/LtrA localization patterns in duplicate arrays , and five that showed the most strongly altered GFP/LtrA localization patterns in liquid culture were studied further . The mariner transposon-insertion sites in the five disruptants were amplified and sequenced via thermal-asymmetric-interlaced ( TAIL ) PCR and found to be in the gppA , uhpT , wcaK , ynbC , and zntR genes . gppA encodes guanosine pentaphosphatase A , which removes the 5′ phosphate from pppGpp to produce the stringent response regulator , ppGpp ( “magic spot” ) [18]; uhpT encodes a hexose phosphate transport component [19]; wcaK is a predicted colanic acid biosynthesis pyruvyl transferase [20]; ynbC encodes a 585-amino-acid ORF of unknown function [21]; and zntR encodes a zinc-responsive transcriptional regulator [22] . In each disruptant , the transposon-insertion site was confirmed by additional PCRs to amplify and sequence both the 5′- and the 3′-transposon-integration junctions ( Figure S1 ) and by Southern hybridization , which also showed that each strain contains a single transposon insertion ( Figure S2 ) . Immunoblots showed that the GFP/LtrA expression levels in the disruptants are similar to or lower than those in the wild-type strain , with the expression level particularly low in the ynbC disruptant ( Figure S3 ) . These findings indicate that the disruptions affect GFP/LtrA localization over a wide range of protein expression levels . gppA and uhpT are in single gene transcription units , and zntR is the last gene of a two-gene operon , while ynbC and wcaK are upstream genes in operons whose disruption could exert polar effects on downstream genes [23] . To characterize their GFP/LtrA localization patterns , the wild-type and disruptant strains were grown in liquid culture , and GFP/LtrA expression was induced overnight with 500 μM IPTG at 30 °C , one of the induction conditions used for screening the transposon-insertion library . GFP/LtrA localization patterns in ≥200 randomly selected cells of each strain were then analyzed by fluorescence microscopy ( Figure 2 and Table 1 , top ) . As found previously , in the wild-type strain , a high proportion of the cells ( 81 . 2% ) showed polar localization of LtrA . Such cells contain one or two small , discrete GFP/LtrA foci at their poles , with elongated cells , presumably those ready for division , typically showing a third focus in the middle . Only 2 . 8% of the wild-type cells showed diffuse GFP/LtrA localization patterns . In wild type as well as the disruptants , ∼20% of cells did not show GFP/LtrA fluorescence detectable above background . In comparison to the wild-type strain , each of the disruptants showed a substantially increased proportion of cells with more diffuse GFP/LtrA localization patterns ( 26 . 4–46 . 1% ) , which were classified into two types: completely diffuse ( C ) , with GFP/LtrA uniformly distributed throughout the cell , or partially diffuse ( P ) , with GFP/LtrA spread out from the poles but leaving a clear area in the middle of the cell ( percentages of C and P are indicated in parentheses in Table 1 ) . The gppA disruptant was the most strongly affected , with only 20 . 5% of the cells still showing polar GFP/LtrA localization , 39 . 5% showing partially or completely diffuse fluorescence , and 16 . 9% filamentous cells with multiple foci and/or irregular diffuse patches of GFP/LtrA fluorescence ( Figure 2 and Table 1 , top ) . The wcaK disruptant also showed an increased proportion of filamentous cells ( 18 . 6% ) , while the remaining three disruptants showed few or no filamentous cells . The ratio of cells with completely or partially diffuse LtrA localization differed for each of the disruptants , being highest for the zntR disruptant and lowest for the uhpT disruptant . When Ll . LtrB intron expression was induced under the higher temperature conditions ( overnight , 100 μM IPTG , 37 °C ) , we observed similar GFP/LtrA localization patterns , but the proportions of cells showing completely or partially diffuse fluorescence shifted in some cases ( Table S1 ) . Although their cellular phenotypes are heterogeneous , PCR using primers flanking the target gene showed that cultures of four of the disruptants ( gppA , uhpT , wcaK , and ynbC ) were homogenous for the disrupted allele , while the fifth ( zntR ) contained predominantly the disrupted allele but also showed a light band ( star ) comigrating with that for the wild-type allele ( Figure S4 ) . This light band was found by sequencing to contain a 4-bp insertion at the transposon-insertion site in the middle of the ORF and is presumably a null allele resulting from transposon excision . Thus , in all cases , the heterogeneous cellular phenotype is not due to persistence of the wild-type allele . Next , we tested whether the disruptions affect intron mobility frequencies . For these experiments , we used an Ll . LtrB-ΔORF intron that had been retargeted by modification of its EBS and δ sequences to insert at a site within the E . coli lacZ gene ( LacZ-1063a ) so that the intron integration frequency could be scored simply by blue–white screening ( see Materials and Methods section and Figure S5 ) . Intron-donor plasmid pACD2X expressing the retargeted intron was transformed into the wild-type and disruptant strains and induced with IPTG at 30 or 37 °C prior to plating the cells on Luria–Bertani ( LB ) medium containing X-Gal . As summarized in Figure 3 , at both temperatures , Ll . LtrB-ΔORF intron mobility frequencies in the disruptants were somewhat higher than those in the wild-type strain ( 68–96% compared to 47–51% ) . Immunoblots of proteins isolated from the induced cells showed slightly reduced levels of LtrA protein in all cases ( 69–78% wild type; Figure 3B ) , indicating that the increased mobility frequencies simply are not due to higher protein expression levels . Possible reasons for the increased chromosomal insertion frequencies in the disruptants are discussed below ( see Discussion section ) . Having demonstrated that the Ll . LtrB-ΔORF intron remains mobile in the disruptants , we next examined whether the changes in LtrA localization patterns are correlated with changes in the genomic distribution of Ll . LtrB-insertion sites . As done previously to analyze insertion-site preference [10] , we used a donor plasmid library that expresses Ll . LtrB-ΔORF introns with randomized target site recognition ( EBS2 , EBS1 , and δ ) sequences plus a TpR retrotransposition-activated marker ( RAM ) to detect chromosome integrations . This marker consists of a small trimethoprim-resistance ( TpR ) gene inserted within the intron in the orientation opposite to intron transcription but interrupted by a small , self-splicing group I intron ( the phage T4 td intron ) in the forward orientation . During retrohoming via an RNA intermediate , the group I intron is spliced , reconstituting the marker and enabling the selection of cells containing integrated introns by trimethoprim resistance . The TpR colonies were isolated and analyzed by TAIL PCR and sequencing to identify the intron-insertion sites . The chromosomal distributions of Ll . LtrB-ΔORF intron-insertion sites in wild-type HMS174 ( DE3 ) and the disruptants are shown in Figure 4 , with the previous distribution for wild-type HMS174 ( DE3 ) from Zhong et al . [10] shown for comparison . The two sets of data for the wild-type strain are in close agreement , showing Ll . LtrB intron-insertion sites strongly clustered around oriC ( blue bars; Figure 4A and 4B ) . By combining the two data sets for the wild-type strain , we defined the region around oriC containing clustered Ll . LtrB-insertion sites as encompassing minutes 69–90 of the E . coli chromosome ( 21% of the genome ) . In the wild-type strain , the proportion of Ll . LtrB-insertion sites in the oriC region defined as above was 75% in the previous work ( Figure 4A ) and 78% in the present work ( Figure 4B ) , strikingly higher than 21% expected for random integration . By contrast , in the disruptants only 35–53% of the Ll . LtrB-insertions sites were located in the oriC region , with the lowest proportions being found in the zntR and ynbC disruptants ( 35% and 38% , respectively; Figure 4C–G ) . The differences in distribution patterns were statistically significant at p < 0 . 001 , calculated by χ-square test . These findings show that the more diffuse intracellular localization of GFP/LtrA in the disruptants is in fact correlated with a more uniform distribution of Ll . LtrB-insertion sites . Among the genes whose disruption results in altered GFP/LtrA localization , the best characterized is gppA , which encodes guanosine pentaphosphatase A ( GPPA ) . This enzyme has two functions: it dephosphorylates pppGpp to produce ppGpp ( “magic spot” ) , and it processively hydrolyzes intracellular poly ( P ) , liberating orthophosphate [18] . To test whether the altered LtrA localization pattern in the gppA disruptant is due to impaired synthesis of magic spot , we examined LtrA localization in two other mutants that affect the synthesis of magic spot in different ways: a relA disruptant derived from wild-type HMS174 ( DE3 ) by targetron mutagenesis ( Figure S5 ) and a previously isolated relA/spoT* double mutant obtained from GenoBase [24] . ( Note that strains obtained from GenoBase were derivatized with λDE3 for intron expression and are denoted with an asterisk to indicate different genetic background . ) Both the relA and the relA/spoT* mutants still showed predominantly polar localization of GFP/LtrA ( 72 . 2% and 69 . 2% of cells , respectively ) , with only small proportions of cells ( 4 . 2% and 6 . 3% , respectively ) showing diffuse GFP/LtrA localization ( Figure 5A and Table 1 , bottom ) . Thus , impaired ability to produce ppGpp does not strongly affect LtrA localization . The mobility frequency of the Ll . LtrB-ΔORF intron assayed by targeted integration into the lacZ gene as above was increased in the relA disruptant and decreased in the relA/spoT* mutant compared to the wild-type strain ( Figure 5B ) . The lower intron mobility frequency in the relA/spoT* mutant could be due to the relA/spoT mutations , the different genetic background , or a combination of the two . The alternate possibility was that GFP/LtrA localization in the gppA disruptant might be due to the accumulation of intracellular poly ( P ) . To test this possibility , we used targetron mutagenesis to disrupt the E . coli HMS174 ( DE3 ) ppk and ppx genes , which encode polyphosphate kinase and exopolyphosphatase , respectively [25 , 26] ( Figure S5 ) . The ppk and ppx genes are expressed from the same operon , with ppk upstream of ppx . Thus , the disruption of ppk is expected to affect the expression of both genes and lead to decreased levels of intracellular poly ( P ) , while the disruption of ppx should affect only the expression of that gene and lead to accumulation of poly ( P ) [26 , 27] . The ppx disruptant , which we confirmed below accumulates poly ( P ) , did in fact show a substantially increased proportion of cells with more diffuse GFP/LtrA localization ( 27 . 4% ) , while the ppk disruptant , expected to have decreased levels of poly ( P ) , showed predominantly the normal bipolar GFP/LtrA localization pattern ( Figure 5A and Table 1 , bottom ) . Similar results were obtained with independently constructed ppx and ppk deletions from the Keio collection obtained from GenoBase ( i . e . , more diffuse GFP/LtrA localization in the ppx* deletion ( 18 . 7% ) and predominantly bipolar GFP/LtrA localization in the ppk* deletion ( Table 1 , bottom ) ) . In both the ppk and the ppx disruptants , the mobility frequency of the Ll . LtrB-ΔORF intron assayed by targeted integration into the lacZ gene was increased relative to the wild-type strain , with the increases more pronounced if normalized to the LtrA expression level ( Figure 5B and 5C ) . We conclude from these findings that the common feature correlated with the more diffuse LtrA localization in the gppA and ppx disruptants is the accumulation of intracellular poly ( P ) . To investigate further the relationship between poly ( P ) accumulation and GFP/LtrA localization , we used fluorescence microcopy to examine the intracellular localization of poly ( P ) detected by 4′ , 6-diamidino-2-phenylindole ( DAPI ) staining in the wild-type strain and the gppA and ppx disruptants expressing GFP/LtrA . Under fluorescence microscopy with excitation at 360 nm , DAPI bound to poly ( P ) emits at 550 nm and appears yellow or orange , while DAPI bound to DNA emits at 490 nm and appears blue [28] . In wild-type HMS174 ( DE3 ) under standard GFP/LtrA induction conditions at 30 °C , 16 . 5% of the cells showed detectable poly ( P ) fluorescence localized in discrete foci , mainly at the cell poles ( P ) , another 8 . 9% showed more diffuse poly ( P ) fluorescence ( D ) , and the remainder ( 74 . 6% ) showed no detectable poly ( P ) fluorescence ( N ) , likely reflecting at least in part the sensitivity of the detection method ( Figure 6A ) . The localization of poly ( P ) in discrete foci ( “volutin granules” ) at the cellular poles has been found previously in other bacteria [29–31] . By contrast , under the same conditions , both the gppA and the ppx disruptions , which lack enzymes involved in poly ( P ) degradation , showed substantially increased proportions of cells with detectable poly ( P ) fluorescence ( 55 . 9% and 61 . 1% , respectively; Figure 6A ) . Further , in a high proportion of gppA and ppx disruptant cells , poly ( P ) was no longer present in discrete foci but was instead dispersed throughout the cell ( 44 . 4% and 25 . 4% , respectively , excluding filamentous gppA cells ) . Conversely , the ppk disruptant , which lacks the major poly ( P ) biosynthetic enzyme , shows a reduced proportion of cells with detectable poly ( P ) fluorescence , which was localized again mainly at or near the cell poles ( 6 . 6%; Figure 6A ) . The residual poly ( P ) in the ppk disruptant may be synthesized by an alternate pathway [32] . Similar differences in poly ( P ) localization between the wild-type and the disruptant strains were observed after IPTG induction in strains carrying the empty vector , indicating that poly ( P ) accumulation in the disruptants is not due to expression of GFP/LtrA ( unpublished data ) . Importantly , although the gppA and ppx disruptants have heterogeneous cellular phenotypes , poly ( P ) accumulation and dispersed GFP/LtrA localization were correlated strongly in individual cells . Thus , for the gppA disruptant , 90% of the doubly fluorescent cells with diffuse poly ( P ) localization showed diffuse GFP/LtrA localization , and 96% with diffuse GFP/LtrA localization showed diffuse poly ( P ) fluorescence . Similarly , in the ppx disruptant , 89% of the doubly fluorescent cells with diffuse poly ( P ) localization showed diffuse GFP/LtrA localization , and 88% with diffuse GFP/LtrA localization showed diffuse poly ( P ) localization . Further , in a high proportion of the doubly fluorescent cells , the poly ( P ) and GFP/LtrA fluorescence were either overlapping ( 94 . 7% and 87 . 9% in the gppA and ppx disruptants , respectively ) or completely colocalized ( 78 . 4% and 45 . 5% in the gppA and ppx disruptants , respectively; Figure 6C ) . This degree of colocalization supports the hypothesis that the altered GFP/LtrA localization is due to the binding of the basic LtrA protein to the negatively charged poly ( P ) . Those cells in which both poly ( P ) and GFP/LtrA are dispersed but not completely colocalized could reflect that poly ( P ) associated with LtrA dissociates or is degraded after the protein is dispersed , that some poly ( P ) bound to LtrA is less than the concentration limit required for fluorescence detection with DAPI , or that the GFP tag is clipped , rendering some proportion of LtrA nonfluorescent . After the above findings , we also re-examined the wild-type strain , where 2 . 8% of the cells showed diffuse GFP/LtrA localization ( Table 1 , top ) . Strikingly , even in wild type , where only a small proportion of cells is affected , diffuse GFP/LtrA localization was correlated again with the accumulation and dispersal of poly ( P ) ( 88% of wild-type cells with diffuse GFP/LtrA showed dispersed poly ( P ) , and 83% with dispersed poly ( P ) showed diffuse GFP/LtrA ) . The accumulation of poly ( P ) in a small proportion of wild-type cells is likely due to cell stress , which triggers poly ( P ) synthesis [33] . Collectively , these findings suggest that the more diffuse intracellular localization of GFP/LtrA results from the accumulation of poly ( P ) , which binds to LtrA and delocalizes it away from the cell poles . The remaining four disruptants ( uhpT , wcaK , ynbC , and zntR ) with more diffuse GFP/LtrA localization do not involve genes that are known to function in poly ( P ) metabolism . However , the above findings raised the possibility that they might also accumulate poly ( P ) as a result of cell stress caused by the disruptions . Fluorescence microscopy of DAPI-stained cells revealed that this is in fact the case , with all four of the above disruptants in the HMS174 ( DE3 ) genetic background showing an increased proportion of cells with higher and more dispersed poly ( P ) fluorescence ( Figure 6B ) . By contrast , strains having deletions of the same four genes ( uhpT* , wcaK* , ynbC* , and zntR* ) in the Keio genetic background did not accumulate poly ( P ) and correspondingly did not exhibit altered GFP/LtrA localization patterns ( unpublished data ) . The likely explanation is that their different genetic background makes the Keio strains less prone to cell stress caused by these mutations and/or less prone to accumulate poly ( P ) in response to cell stress than in the HMS174 ( DE3 ) background . In comparison to the ppx disruption in the HMS174 ( DE3 ) background ( Figure 6 ) , the ppx* deletion in the Keio genetic background also showed less detectable poly ( P ) accumulation ( 1 . 0% P , 6 . 0% D , 93 . 0% N ) and correspondingly had a more muted effect on GFP/LtrA localization ( Table 1 , bottom ) . We conclude that poly ( P ) accumulation is the common factor correlated with altered GFP/LtrA localization in all of the strains analyzed here . To test directly whether poly ( P ) binds to LtrA , we examined its effect on the electrophoretic mobility of the LtrA protein in a nondenaturing polyacrylamide gel ( Figure 7A ) . In the absence of poly ( P ) , the LtrA protein , which is highly positively charged ( calculated pI = 9 . 6 ) , could not migrate toward the positive pole and failed to enter the gel . By contrast , with increasing concentrations of poly ( P ) , an increasing proportion of LtrA entered the gel and migrated toward the positive pole , indicating complex formation . Control lanes ( right ) show that the highest concentration of poly ( P ) tested had no effect on the electrophoretic mobility of an acidic protein , bovine serum albumin ( pI = 4 . 6 ) , run in the same gel . Figure 7B extends these findings by showing that equimolar poly ( P ) completely inhibited the reverse transcriptase ( RT ) activity of LtrA . Poly ( P ) also inhibited the RT activity of Moloney murine leukemia virus and Superscript RTs , although these enzymes appeared somewhat less sensitive to poly ( P ) inhibition than LtrA ( Figure 7C and 7D ) . Together , the above findings show that poly ( P ) binds LtrA and inhibits its RT activity . The ability of poly ( P ) to bind LtrA and carry it toward the opposite pole in a nondenaturing gel supports its hypothesized mechanism of action in the cell . Finally , we used the gppA and ppx disruptants to test whether poly ( P ) accumulation might similarly affect the polar localization of proteins other than LtrA . First , we tested GFP fusions of two basic proteins , which were shown previously to be pole-localized in E . coli: the N . crassa CYT-18 protein ( calculated pI = 9 . 29 ) [12] and the E . coli transcriptional regulator XapR ( calculated pI = 8 . 91 ) [34] . Strikingly , we found that both proteins are largely pole-localized in wild type and the ppk disruptant but showed more dispersed localization patterns in the gppA and ppx disruptants , which accumulate poly ( P ) ( Figure 8A and 8B ) . For GFP/CYT-18 , the gppA and ppx disruptants showed high proportions of cells with more diffuse protein localization , while for GFP/XapR these disruptants showed high proportions of cells with multiple large foci scattered throughout the cell . For GFP/XapR , we confirmed by dual-fluorescence microscopy that the scattered localization in individual cells is again correlated with the accumulation and dispersal of poly ( P ) ( 86% of the gppA and 83% of the ppx cells showing scattered GFP/XapR localization also showed dispersed poly ( P ) localization ) . Immunoblots showed that GFP/XapR is expressed at similar levels in the wild-type and gppA and ppx disruptant strains ( Figure S6 ) . By contrast to these basic proteins , a GFP fusion with a S . flexneri IcsA protein subsegment ( IcsA507–620 pI = 7 . 15 ) [35] remained pole-localized in the gppA and ppx disruptants , although for both mutants an increased proportion of cells with pole-localized IcsA507–620/GFP showed additional tiny foci distributed throughout the cell ( P*; Figure 8C ) . Thus , as expected , poly ( P ) accumulation most strongly affects the localization of positively charged proteins to which it can bind directly .
Here , by using high-throughput methods to screen a transposon-insertion library , we identified five E . coli genes ( gppA , uhpT , wcaK , ynbC , and zntR ) whose disruption affects the polar localization of LtrA , a group II intron-encoded reverse transcriptase that functions with the intron RNA in RNPs to promote intron mobility . All of the above disruptants show both an increased proportion of cells with more diffuse GFP/LtrA localization and a more uniform genomic distribution of Ll . LtrB-insertion sites , which in the wild-type E . coli strain are strongly clustered in the oriC region . These findings indicate that the intracellular localization of LtrA is a major determinant of Ll . LtrB-insertion-site preference , to our knowledge the first such demonstration for any mobile genetic element . Further analysis showed that the common factor leading to more diffuse GFP/LtrA localization in these and other disruptants is the accumulation of intracellular poly ( P ) and that disruptants that accumulate poly ( P ) also show altered intracellular distributions of other basic proteins that are normally pole-localized ( CYT-18 and XapR ) . The latter findings suggest that poly ( P ) accumulation may be part of a cellular mechanism that leads to relocalization of basic proteins in response to cell stress or entry into stationary phase . From a technical standpoint , our results demonstrate the feasibility of using bacterial cellular arrays for high-throughput screens to identify mutations affecting protein localization or morphology , and they suggest a method for obtaining more uniform group II intron-gene disruption libraries by using mutants that accumulate poly ( P ) . The polar localization of LtrA in wild-type cells may facilitate group II intron mobility by increasing access of group II intron RNPs to exposed DNA segments in the oriC and ter regions of the E . coli chromosome . It also may provide favorable sites for interaction with DNA replication components and a means of coordinating group II intron mobility with DNA replication and/or cell division [12] . Other mobile elements use a variety of mechanisms for coordinating transposition with DNA replication ( [36 , 37] and references therein ) . Such coordination may be particularly important for group II introns , which can use nascent strands at DNA replication forks to prime reverse transcription [11 , 38 , 39] . Indeed , Coros et al . [11] found that in E . coli the frequency of Ll . LtrB retrotransposition events that use this priming mechanism is highest in the oriC domain and decreases in a gradient toward the ter domain , while retrotransposition events that utilize the DNA cleavage activity of the IEP to generate the primer for reverse transcription do not show such a gradient . The selection for polar localization may have originated with ancestral group II introns whose IEPs lacked DNA cleavage activity and relied entirely on nascent DNA strands as primers . In addition to changing the intracellular localization of LtrA , poly ( P ) also potentially could affect the genomic distribution of Ll . LtrB-insertion sites by binding basic nucleoid-associated proteins , leading to nucleoid decondensation , as observed in a Pseudomonas aeruginosa ppk1 mutant [40] . Indeed , it has been speculated that such a mechanism contributes to global changes in gene expression accompanying poly ( P ) accumulation in response to cell stress or entry into stationary phase [41] . However , Beauregard et al . [15] found that neither LtrA localization nor the distribution of Ll . LtrB-insertion sites is affected significantly by mutations in the nucleotide-associated proteins H-NS , StpA , or MukB , which lead to nucleoid decondensation in different ways . These findings indicate that nucleoid decondensation does not by itself lead to a more uniform distribution of Ll . LtrB-insertion sites as long as LtrA remains pole-localized . It is possible , however , that nucleotide decondensation contributes to the more uniform distribution of insertion sites that we observe when LtrA has been delocalized by poly ( P ) . Our results suggest that negatively charged poly ( P ) delocalizes LtrA and other basic proteins by binding to them directly . A direct binding mechanism is supported by: ( i ) the colocalization of GFP/LtrA and poly ( P ) in high proportions of gppA and ppx disruptant cells ( Figure 6C ) ; ( ii ) the finding that poly ( P ) accumulation affects the polar localization of three basic proteins ( LtrA , CYT-18 , and XapR ) but not the polar localization of the acidic protein IcsA507–620/GFP ( Figure 8 ) ; and ( iii ) biochemical experiments showing that poly ( P ) can bind LtrA to inhibit its RT activity and cause it to migrate toward the opposite pole in a nondenaturing polyacrylamide gel ( Figure 7 ) . The binding of poly ( P ) to LtrA and other basic proteins is presumably nonspecific , but we cannot at this stage exclude a more specific binding component in some cases . In the cell , poly ( P ) may bind to LtrA and other basic proteins at the poles and relocalize them as it moves through the cell , and/or it may bind to the nascent proteins in other regions , preventing them from becoming pole-localized . The poly ( P ) that colocalizes with LtrA could be present in specific complexes , akin to volutin granules , which may be passively or actively localized or delocalized . In addition to affecting the intracellular localization of LtrA , the binding of poly ( P ) may directly affect LtrA's biochemical activities in RNA splicing and intron mobility . Precedents for such effects include the findings that poly ( P ) activates Lon protease degradation of ribosomal proteins [42] , may play a role in activating mammalian TOR kinase [43] , and is required for the lytic growth of phages P1 and fd [44] . Our finding that poly ( P ) inhibits LtrA's RT activity ( Figure 7B ) may explain why the ppk mutant , which has decreased levels of poly ( P ) , shows increased Ll . LtrB mobility frequencies in a chromosomal lacZ gene integration assay ( Figure 5B ) . However , we also find that Ll . LtrB mobility frequencies are increased moderately in gppA , ppx , and other mutants that accumulate poly ( P ) ( Figures 3A and 5B ) . The latter findings indicate that inhibition of LtrA's RT activity by poly ( P ) in vivo must be either mild or transient , perhaps due to dissociation or degradation of the bound poly ( P ) after the protein has been delocalized . A factor that may contribute to the moderately increased Ll . LtrB mobility frequencies in mutants that accumulate poly ( P ) is that their more uniform intracellular distribution of LtrA makes it easier to target the lacZ gene , which is located outside the oriC or ter chromosomal regions . The disruptants that we identified with altered GFP/LtrA localization patterns accumulate poly ( P ) for different reasons . The gppA and ppx disruptions inhibit poly ( P ) degradation , while the uhpT , wcaK , ynbC , and zntR disruptants presumably accumulate poly ( P ) as a result of cellular stress caused by the disruptions . All of disruptants , including the gppA and ppx disruptants , have heterogeneous cellular phenotypes , with an increased proportion of cells showing both poly ( P ) accumulation and altered GFP/LtrA localization , and the remainder having wild-type localization patterns . This heterogeneity may reflect that only some cells in the population are under sufficient stress to trigger poly ( P ) accumulation . The wild-type strain shows similar heterogeneity but with only a small proportion of cells showing poly ( P ) accumulation and altered GFP/LtrA localization . Thus , the disruptions appear to increase the normal propensity for stress-induced poly ( P ) accumulation , either by increasing the degree of stress or by decreasing the degradation rate of poly ( P ) , making it easier to achieve elevated poly ( P ) concentrations . We note that the gppA disruption causes a more extreme accumulation and dispersal of poly ( P ) than does the ppx disruptant ( Figure 6 ) , possibly reflecting that the gppA disruptant not only lacks GPPA , which contributes to poly ( P ) degradation , but also accumulates pppGpp , which inhibits the remaining exopolyphosphatase PPX [42] . Like a previous screen for E . coli localization factors [16] , our cell array screen did not identify a pole-localized receptor protein whose absence leads to diffuse LtrA localization . It is possible that a specific polar receptor for LtrA does not exist and that LtrA is localized to the poles by default because it is not actively localized elsewhere . Alternatively , the receptor may require essential proteins or be redundant or nonspecific ( e . g . , acidic phospholipids [45] ) . Additionally , the intriguing finding that in wild-type E . coli poly ( P ) detected by DAPI staining is found frequently in discrete foci at the poles raises the possibility that poly ( P ) itself contributes as a receptor for the polar localization of basic proteins under nonstress conditions . The finding that GFP/LtrA and other basic proteins remain pole-localized in ppk mutants , which have ∼10-fold decreased poly ( P ) levels [27] , does not exclude this possibility because poly ( P ) may be present in excess in wild-type strains and the residual poly ( P ) in the mutant may be sufficient for protein localization . Finally , the finding that disruptants that accumulate poly ( P ) show altered distributions not only of LtrA but also of other pole-localized basic proteins suggests that poly ( P ) accumulation may be part of a mechanism that relocalizes proteins to different sites of action in response to cellular stress . One can imagine that reservoirs of certain enzymes , such as transcription factors or DNA repair enzymes , accumulate at the cellular poles and are mobilized to new sites of action by poly ( P ) during entry into stationary phase or under stress conditions , as shown here for LtrA . Such relocalization may be particularly advantageous when the cell's biosynthetic capacity is impaired as it is faster and more economical than synthesizing a specific receptor and transport system for each protein . Protein relocalization by binding to poly ( P ) may have wide physiological consequences , not only in prokaryotes but also in eukaryotes , where poly ( P ) also exists but has remained enigmatic .
E . coli HMS174 ( DE3 ) ( F– recA hsdR rifR ) ( Novagen ) was used for LtrA localization and Ll . LtrB intron-integration assays; DH5α was used for cloning; and S17–1λpir [46] ( obtained from Ram Narayanaswamy and Andy Ellington , University of Texas at Austin ) was used for mariner transposon mutagenesis . Derivatives of HMS174 ( DE3 ) with disruptions of the ppk , ppx , and relA genes were constructed by targetron mutagenesis , as described ( [8]; Figure S5 ) . Keio deletion strains obtained from GenoBase ( http://ecoli . naist . jp/GB6/search . jsp ) were ppk , ppx , uhpT , wcaK , ynbC , and zntR . Other mutants obtained from GenoBase were gppA and relA/spoT . GenoBase strains were lysogenized with λ ( DE3 ) carrying an IPTG-inducible T7 RNA polymerase gene by using a kit ( Novagen ) and are indicated with an asterisk in the text to denote their different genetic background . Strains were grown in LB medium at 30 or 37 °C , with antibiotics used at the following concentrations: ampicillin , 100 μg/ml; chloramphenicol , 25 μg/ml; kanamycin , 40 μg/ml; rifampicin , 100 μg/ml; trimethoprim , 10 μg/ml . pACD2X-GFP/LtrA expresses the Ll . LtrB-ΔORF intron and a GFP/LtrA fusion protein ( Figure 1A ) [12] . The pACD3-TpR-RAM library expresses Ll . LtrB-ΔORF introns with randomized EBS2 , EBS1 , and δ sequences plus a TpR-RAM for detecting chromosome integrations [10] . Plasmids expressing other GFP fusion proteins were: pAC-GFP/CYT-18 , N . crassa CYT-18 protein with an N-terminal GFP fusion [12]; pBAD24-icsA507–620::gfp , pole-localized segment of the S . flexneri IcsA protein with a C-terminal GFP fusion [35]; and pAC-GFP/XapR , E . coli XapR protein with an N-terminal GFP fusion . The latter plasmid was derived from pAC-GFP/LtrA [12] by replacing the LtrA ORF with the xapR ORF ( codons 1–294 ) amplified from E . coli HMS174 ( DE3 ) by PCR . pSC189 expresses a mariner transposon with a kanR marker and a separately encoded hyperactive C9 transposase [47] . The protein expression plasmid pImp-1P contains the LtrA ORF cloned behind a tac promoter in pCYB2 ( New England Biolabs ) [4] . E . coli strains HMS174 ( DE3 ) and S17–1λpir carrying pSC189 ( see above ) were grown separately to OD600 = 0 . 5 at 37 °C in 10 ml of LB without and with ampicillin , respectively , then mixed , and incubated at room temperature overnight without shaking for conjugation . The conjugated cells were washed twice with LB medium by centrifugation , resuspended , and grown overnight at 37 °C in fresh LB medium containing rifampicin and kanamycin to kill the donor strain and select for recipient HMS174 ( DE3 ) cells containing mariner transposon insertions . Loss of pSC189 , which carries an ampR marker and does not replicate in HMS174 ( DE3 ) , was confirmed by plating on LB agar with and without ampicillin ( <1% AmpR colonies ) . The HMS174 ( DE3 ) isolate used for library construction inadvertently carried a TetR broad-host-range plasmid pBBR1MCS-3 [48] . For the GFP/LtrA localization screen , cells grown in LB with kanamycin were electroporated with the intron-donor plasmid pACD2X-GFP/LtrA and plated on LB agar containing chloramphenicol and kanamycin . Approximately 9 , 600 colonies were picked into one hundred 96-well plates and stored at −80 °C . TAIL PCR of isolates showed that >92% contained different mariner transposon insertions , both before and after introduction of pACD2X-GFP/LtrA . The E . coli HMS174 ( DE3 ) mariner transposon-insertion library carrying pACD2X-GFP/LtrA was inoculated into new 96-well plates containing LB medium with chloramphenicol plus kanamycin and incubated overnight at 37 °C . The cultures then were inoculated 1:10 into fresh medium plus 17% glycerol in a new 96-well plate and grown for 5 h at 37 °C . In one screen , cells in fifty-one 96-well plates were induced overnight with 100 μM IPTG at 37 °C , and in another screen , cells in forty-nine 96-well plates were induced overnight with 500 μM IPTG at 30 °C . Culture transfer and media additions were done by a Biomek FX Laboratory Automation Workstation ( Beckman Coulter ) . Cell microarrays ( cell chips ) were constructed as described [17] . Briefly , ∼5 , 000 knockouts plus a wild-type HMS174 ( DE3 ) control were printed onto poly-l-lysine-coated microscope slides using a custom-built DNA microarray-printing robot . In each experiment , ∼30 cell chips were made , of which two were used for imaging and the remainder were stored at −80 °C . Before being imaged , the cell chips were washed briefly with double-distilled water to remove glycerol and debris and then mounted with VECTASHIELD hard-set mounting medium containing 1 . 5 μg/ml DAPI . Cell images were collected by automated microscopy , using a Nikon E800 fluorescence microscope with computer-controlled X-Y stage and piezoelectric-positioned objective . The automated microscope scanned the position of each spot , focused , and captured the image with a Coolsnap CCD camera ( Photometrics ) . Images were stored in a custom cell microarray image database ( Cellma , http://cellma . icmb . utexas . edu/ ) and examined individually to identify strains with altered GFP/LtrA localization patterns . Fluorescence microscopy was done as described [12] . Cells carrying pACD2X-GFP/LtrA were grown in LB medium with appropriate antibiotics at 37 °C to OD600 = 0 . 3 and then induced overnight with 500 μM IPTG at 30 °C or 100 μM IPTG at 37 °C . DAPI ( 25 μg/ml ) was added to cultures 30 min before the end of induction . Cells were examined by fluorescence microscopy using a Leica DMIRBE microscope ( Leica ) with a 100× oil lens ( PL APO 1 . 4–0 . 7 NA ) and a GFP filter for GFP fluorescence or a wide DAPI filter for poly ( P ) fluorescence . Photographs were taken with a Leica DFC350 FX fluorescence camera ( GFP fluorescence ) or a Leica DFC320 FX color camera ( poly ( P ) fluorescence ) . Ll . LtrB-ΔORF intron mobility frequencies were determined by using a retargeted intron that inserts at a site within the E . coli lacZ gene ( targetron LacZ-1063a; Figure S5 ) . Cells containing the retargeted intron cloned in pACD2X were grown in LB medium with chloramphenicol at 37 °C to OD600 = 0 . 2–0 . 3 and induced with 500 μM IPTG for 0 . 5 h at 30 °C or 100 μM IPTG for 1 h at 37 °C . The cells then were washed with fresh LB medium and plated on LB containing X-Gal ( Fisher Scientific ) . After overnight incubation at 37 °C , the lacZ integration frequency was calculated as the percentage of white colonies . Cells transformed with a pACD3-TpR-RAM library of Ll . LtrB introns with randomized EBS2 , EBS1 , and δ sequences and a TpR-RAM inserted in intron domain IV [10] were grown overnight at 37 °C in LB medium containing chloramphenicol ( wild-type strains ) or chloramphenicol plus kanamycin ( disruptants ) , then inoculated 1:100 into fresh medium , grown to OD600 = 0 . 3 , and induced with 500 μM IPTG overnight at 30 °C . Cells containing chromosomally integrated Ll . LtrB introns carrying the activated TpR-RAM were selected by plating on Mueller–Hinton medium with trimethoprim and thymine [10] . TAIL PCR [49] was done on colonies resuspended in PCR premix . Integration junctions were amplified by two successive PCRs , using two nested specific primers in combination with a single degenerate primer . For mariner transposon insertions , the first PCR used the specific primer TailP1 ( 5′-GTTCTTCTGAGCGGGACTCTGGGG-3′ ) and the degenerate primer AD2 ( 5′-NGTCGASWGANAWGA-3′ , where N = A , C , G , or T; S = C or G; and W = A or T ) , and the second PCR used the specific primer TailP2 ( 5′-CGGCCGCGAAGTTCCTATTCCG-3′ ) and AD2 . The final PCR product was gel-purified and sequenced using the TailP2 primer . For Ll . LtrB-intron insertions , the specific primers used in the first and second PCRs were Ell1 ( 5′-CTGATTAACATTGCGACTCAGTCGTACCC-3′ ) and Ell2 ( 5′-CAACCGTGCTCTGTTCCCGTATCAG-3′ ) , respectively , and the degenerate primer was again AD2 . The final PCR products were sequenced by using primer Ell3 ( 5′-GGTTGGCTGTTTTCTGTGTTATCTTACAGAG-3′ ) . Proteins isolated from equal amounts of cells ( OD600= 0 . 02 or 0 . 04 ) were run in either a 7 . 5% polyacrylamide/0 . 1% SDS gel ( GFP/LtrA , Figure S3 ) or a 10% polyacrylamide/0 . 1% SDS gel ( LtrA , Figures 3 and 5; GFP/XapR , Figure S6 ) , then transferred to a nitrocellulose membrane ( Bio-Rad ) by using a semidry transfer apparatus ( Hoefer Semiphor TE 70; Amersham Biosciences ) . The blots were probed with a 1:5 , 000 dilution of anti-GFP antibody JL-8 ( BD Biosciences ) , followed by a 1:10 , 000 dilution of horseradish peroxidase goat anti-mouse secondary antibody ( Bio-Rad ) , or with a 1:1 , 000 dilution of an anti-LtrA antibody preparation ( obtained from Gary Dunny , University of Minnesota ) , followed by a 1:100 , 000 dilution of horseradish peroxidase goat anti-rabbit secondary antibody ( Pierce ) . Blots were developed with Amersham ECL western blotting detection reagents ( GE Healthcare ) . Equal loading was confirmed by Coomassie blue staining of a parallel gel . DNA was isolated from wild-type and mutants strains by using a genomic DNA isolation kit ( Qiagen ) , digested with restriction enzymes , and run in a 0 . 8% agarose gel . The gel was blotted to a nylon transfer membrane ( Magna , 0 . 45 μm; GE Osmonics Labstore ) and hybridized with a 32P-labeled probe corresponding to mariner transposon positions 1385–1868 . The probe was generated by PCR of pSC189 with primers Mar3200 ( 5′-GGGTGGAGAGGCTATTCGGCTATGACTGGGC-3′ ) and Mar-3650 ( 5′-CCTTGAGCCTGGCGAACAGTTCGGCTGG-3′ ) , followed by labeling with [α-32P]dTTP ( 3 , 000 Ci/mmol; PerkinElmer ) , using a High Prime DNA labeling kit ( Roche ) . The blots were scanned with a Typhoon Trio fluorescence scanner ( Amersham Biosciences ) . The LtrA protein was expressed in E . coli Rosetta ( DE3 ) ( Novagen ) from the intein-based expression plasmid pImp-1P ( see above ) and purified , essentially as described [4] . The LtrA protein used for RT activity assays was purified through an additional heparin-Sepharose chromatography step . For nondenaturing PAGE , poly ( P ) ( Type 65; Sigma-Aldrich ) was incubated with purified LtrA protein or bovine serum albumin ( Fraction V ( Sigma-Aldrich ) , with 20% glycerol added to match LtrA ) in 20 μl of 100 mM KCl , 5 mM MgCl2 , 20 mM Tris-HCl , pH . 7 . 5 , for 30 min at 30 °C . The samples were then run in a nondenaturing 5–15% polyacrylamide gradient gel in Tris-acetate , pH 6 . 5 , buffer . The gel was silver-stained , and a parallel gel was used for immunoblotting with anti-LtrA antibody ( see above ) . RT activity was assayed as described [50] by polymerization of [32P]-dTTP ( 3 , 000 Ci/mmol; PerkinElmer ) with poly ( rA ) /oligo ( dT ) 18 as template in reaction medium containing 10 mM KCl , 10 mM MgCl2 , and 50 mM Tris-HCl , pH 7 . 5 , for 10 min at 37 °C . Samples were spotted onto DE-81 paper ( Whatman ) , washed three times with 2× SSC ( SSC is 0 . 15 M NaCl and 0 . 15 M citric acid , pH 7 . 0 . ) , and counted in a scintillation counter .
Accession numbers for Keio deletion strains from GenoBase ( http://ecoli . naist . jp/GB6/search . jsp ) are ppk ( JW2486 ) , ppx ( JW2487 ) , uhpT ( JW3641 ) , wcaK ( JW2030 ) , ynbC ( JW1407 ) , zntR ( JW2354 ) , gppA ( JD24693 ) , and relA/spoT ( AQ4319 ) .
|
Group II introns are bacterial mobile elements thought to be ancestors of introns—genetic material that is discarded from messenger RNA transcripts—and retroelements—genetic elements and viruses that replicate via reverse transcription—in higher organisms . They propagate by forming a complex consisting of the catalytically active intron RNA and an intron-encoded reverse transcriptase ( which converts the RNA to DNA , which can then be reinserted in the host genome ) . The Ll . LtrB group II intron-encoded protein ( LtrA ) was found previously to localize to bacterial cellular poles , potentially accounting for the preferential insertion of Ll . LtrB in the replication origin ( oriC ) and terminus ( ter ) regions of the Escherichia coli chromosome , which are located near the poles during much of the cell cycle . Here , we identify E . coli genes whose disruption leads both to more diffuse LtrA localization and a more uniform chromosomal distribution of Ll . LtrB-insertion sites , proving that the location of the LtrA protein contributes to insertion-site preference . Surprisingly , we find that LtrA localization in the disruptants is affected by the accumulation of intracellular polyphosphate , which appears to bind basic proteins and delocalize them away from the cellular poles . Thus , polyphosphate , a ubiquitous but enigmatic molecule in prokaryotes and eukaryotes , can localize proteins to different sites of action , with potentially wide physiological consequences .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"biochemistry",
"cell",
"biology",
"microbiology",
"evolutionary",
"biology",
"molecular",
"biology",
"genetics",
"and",
"genomics"
] |
2008
|
Group II Intron Protein Localization and Insertion Sites Are Affected by Polyphosphate
|
Salmonella serovars Typhi ( S . Typhi ) and Paratyphi A ( S . Paratyphi A ) , the causative agents of enteric fever , have been routinely isolated organisms from the blood of febrile patients in the Kathmandu Valley since the early 1990s . Susceptibility against commonly used antimicrobials for treating enteric fever has gradually changed throughout South Asia since this time , posing serious treatment challenges . Here , we aimed to longitudinally describe trends in the isolation of Salmonella enterica and assess changes in their antimicrobial susceptibility in Kathmandu over a 23-year period . We conducted a retrospective analysis of standardised microbiological data from April 1992 to December 2014 at a single healthcare facility in Kathmandu , examining time trends of Salmonella-associated bacteraemia and the corresponding antimicrobial susceptibility profiles of the isolated organisms . Over 23 years there were 30 , 353 positive blood cultures . Salmonella enterica accounted for 65 . 4% ( 19 , 857/30 , 353 ) of all the bacteria positive blood cultures . S . Typhi and S . Paratyphi A were the dominant serovars , constituting 68 . 5% ( 13 , 592/19 , 857 ) and 30 . 5% ( 6 , 057/19 , 857 ) of all isolated Salmonellae . We observed ( i ) a peak in the number of Salmonella-positive cultures in 2002 , a year of heavy rainfall and flooding in the Kathmandu Valley , followed by a decline toward pre-flood baseline by 2014 , ( ii ) an increase in the proportion of S . Paratyphi in all Salmonella-positive cultures between 1992 and 2014 , ( iii ) a decrease in the prevalence of MDR for both S . Typhi and S . Paratyphi , and ( iv ) a recent increase in fluoroquinolone non-susceptibility in both S . Typhi and S . Paratyphi isolates . Our work describes significant changes in the epidemiology of Salmonella enterica in the Kathmandu Valley during the last quarter of a century . We highlight the need to examine current treatment protocols for enteric fever and suggest a change from fluoroquinolone monotherapy to combination therapies of macrolides or cephalosporins along with older first-line antimicrobials that have regained their efficacy .
Enteric ( typhoid ) fever is a systemic disease caused by Salmonella serovars Typhi ( S . Typhi ) and Paratyphi A ( S . Paratyphi A ) . Global estimates suggest that 12 to 27 million new cases of enteric fever occur annually resulting in 130 , 000 to 220 , 000 deaths , which predominantly occur in low-middle income countries ( LMICs ) [1–3] . Enteric fever has been a public health concern in Nepal for some time , with S . Typhi and S . Paratyphi A consistently being regularly isolated from the blood of febrile patients in the Kathmandu Valley since the early 1990s [4–6] . The sustained high prevalence of S . Typhi and S . Paratyphi A in the Kathmandu Valley has been attributed to a contaminated water supply and poor sanitation [6] . Enteric fever is an infection that requires antimicrobial therapy , but antimicrobial resistance ( AMR ) hinders effective treatment , prolonging the duration of fever and leaving patients at risk of further complications . As a consequence of horizontal acquisition of resistance genes and point mutations , the antimicrobial susceptibility profiles of S . Typhi and S . Paratyphi A have changed substantially in Asia over the past 30 years . Organisms resistant to ampicillin , chloramphenicol , and co-trimoxazole ( multidrug-resistant , MDR ) first appeared in the 1980s and 1990s and were associated with large focal outbreaks [7] . The emergence and persistence of resistance against these antimicrobials diverted clinical practice towards more frequent use of fluoroquinolones to treat enteric fever [8] . This shift towards the use of fluoroquinolones was predictably followed by a general decline in susceptibility against these antimicrobials through sequential mutations in the genes encoding the enzymes targeted by fluoroquinolones [8] . Understanding longitudinal changes in isolation patterns of Salmonella serovars and fluctuations in their corresponding antimicrobial susceptibility profiles in locations where enteric fever is endemic is important for disease surveillance . These data are critical for improving clinical care , updating treatment guidelines , and guiding public health interventions . Aiming to delineate trends in the isolation of invasive Salmonella and their corresponding antimicrobial susceptibility profiles , we conducted a retrospective analysis of 23 years of Salmonella isolated from blood cultures taken from patients attending a major healthcare facility in the Kathmandu Valley in Nepal .
Data for this study consisted of anonymised laboratory results devoid of individual patient information or identifiers . This study was therefore part of the routine surveillance measures within this healthcare facility and ethical approval and individual informed consent were not necessary . However , a written permission , for access and analysis of the data , was sought and obtained from the Patan Hospital management . This study was a retrospective analysis of all blood cultures performed between April 1992 and December 2014 at Patan Hospital . Patan Hospital is located in Lalitpur Sub-metropolitan City ( LSMC ) within the Kathmandu Valley in central Nepal . It is one of three general hospitals in the greater metropolitan area of Kathmandu . The hospital had a capacity of 138 beds in 1992 , 350 beds in 2014 , a current capacity of 592 beds , and provides emergency and elective services to outpatients ( approximately 200 , 000 outpatient visits per year ) and inpatients , 90% of which live in LSMC . For each patient , the date of blood draw , the ward in which the blood sample was taken , the result of the culture ( whether positive or negative ) and the susceptibility of the isolate to commonly used antimicrobials were recorded by a member of the hospital laboratory personnel . For the purposes of this investigation , all handwritten data recorded between the specified dates were transcribed into bespoke software designed by Bonfire Technologies , Nepal . All data were double entered; 10% of all entries were subjected to random inspection by a team of dedicated data entry personnel from the Nepal Family Development Foundation ( NFDF ) . These were the source data for this study . Conventional manual blood cultures were performed for the majority of the data used in these analyses with the exception of the paediatric population ( <14 years of age ) . Conventional blood cultures were performed by inoculating 3–5 ml of blood from paediatric patients and 5–8 ml of blood from adult patients into 30–50 ml of media containing tryptone soya broth containing 0 . 05% sodium polyanetholesulfonate . BACTEC Peds plus bottles ( Becton Dickinson , Sparks , MD , USA ) were used for the paediatric samples after an automated system was introduced . Conventional bottles were incubated at 37°C in a standard microbiological incubator and examined daily for growth for a period of seven days . An automated BACTEC ( Becton Dickinson , MD , USA ) culture system was used for paediatric samples from 2005 onwards . Organisms isolated from blood were identified using Gram staining , standard biochemical tests , and specific antisera . Antimicrobial susceptibilities were tested at the time of isolation by the modified Kirby-Bauer disc diffusion method . Zone size interpretations were initially recorded according to the guidelines provided in the antimicrobial packaging; from 2003 zone sizes were interpreted according to the annual CLSI guidelines . Dependant on the period of isolation and the organism a range of antimicrobials were tested , including: Amikacin ( AMK ) , amoxicillin ( AMX ) , ampicillin ( AMP ) , cefixime ( CFM ) , cefotaxime ( CTX ) , ceftriaxone ( CRO ) , cephalexin ( LEX ) , chloramphenicol ( CHL ) , ciprofloxacin ( CIP ) , trimethoprim-sulphamethoxazole ( SXT ) , gatifloxacin ( GAT ) , gentamicin ( GEN ) , nalidixic acid ( NAL ) , norfloxacin ( NOR ) , ofloxacin ( OFX ) , tetracycline ( TET ) , tigecycline ( TGC ) , tobramycin ( TOB ) . Following antimicrobials were occasionally tested for: ampicillin-sulbactam ( SAM ) , cefepime-tazobactam ( FEP . TZB ) , cefoperazone ( CFP ) , cefoperazone-sulbactam ( CFP . SUL ) , cefoxitin ( FOX ) , ceftriaxone-sulbactam ( CRO . SUL ) , cloxacillin ( CLO ) , colistin ( CST ) , erythromycin ( ERY ) , imipenem ( IPM ) , meropenem ( MEM ) , nitrofurantoin ( NIT ) , oxacillin ( OXA ) , penicillin ( PEN ) , piperacillin ( PIP ) , piperacillin-tazobactam ( TZP ) , and teicoplanin ( TEC ) . Resistant and intermediate isolates were grouped as “non-susceptible” and MDR was defined as non-susceptibility to ampicillin/amoxicillin , chloramphenicol , and co-trimoxazole . All statistical analysis was performed using statistical software R , version 3 . 3 . 2[9] . Increasing or decreasing trends over time were identified using the Cox and Stuart test for trends ( R-package “randtrends” ) . Davies’ test ( R-package “segmented” ) was used to detect a non-constant regression parameter ( i . e . change in the slope ) in linear regressions .
From April 1992 to December 2014 , a total of 224 , 741 individual patient blood samples were drawn for microbiological culture at this healthcare facility . Of these , 173 , 892 ( 77 . 4% ) were culture-negative , 10 , 496 ( 4 . 7% ) were positive for non-Salmonella bacteria , and 20 , 496 ( 9 . 1% ) were contaminated or contained fungi ( Table 1 ) . The remaining 8 . 8% of samples were positive for Salmonella enterica sub-species I , accounting for 65 . 4% ( 19 , 857/30 , 353 ) of all the bacteria positive blood cultures . S . Typhi and S . Paratyphi A were the dominant serovars , constituting 68 . 5% ( 13 , 592/19 , 857 ) and 30 . 5% ( 6 , 057/19 , 857 ) of all isolated Salmonellae , respectively ( Table 2 ) . S . Paratyphi A represented 99 . 44% of all S . Paratyphi ( 6 , 057/6 , 091 ) . The annual distributions of Salmonella-positive , positive for non-Salmonella bacteria , and culture-negative/contaminated blood cultures are shown in S1 Fig . The annual distributions of Salmonella-positive blood cultures ( in all bacteriologically positive blood cultures ) are shown in Fig 1 . The proportion of Salmonella positive blood cultures did not exhibit a uniformly increasing or decreasing trend throughout the study period . However , we observed a peak of 2 , 590 positive cultures in 2002 , a year of unusually high rainfall and flooding in the Kathmandu Valley . Subsequently , the total number of Salmonella-positive cultures decreased towards pre-2002 baseline numbers and the proportion of Salmonella-positive blood cultures ( in all bacteria-positive cultures ) exhibited a significant decreasing trend from 2002 to 2014 ( p = 0 . 016 ) . During this 13-year period ( 2002–2014 ) , the proportion of Salmonella in all bacteria positive blood cultures decreased by a mean of 2 . 4% ( 95% CI: 1 . 1–3 . 7 ) per year ( S2 Fig ) . Additionally , there was a significant decreasing trend in the absolute number of blood cultures from which Salmonella were isolated during this corresponding period ( p = 0 . 016 ) . S . Typhi was markedly the more prevalent of the two Salmonella serovars over the period of investigation ( Fig 1 ) . However , the proportion of S . Typhi ( with respect to all isolated Salmonella ) exhibited a significant decreasing trend from 1992 to 2014 ( p = 0 . 033 ) , decreasing on average by 0 . 81% a year ( 95% CI: 0 . 3–1 . 3 ) . Conversely , the proportion of S . Paratyphi A ( with respect to all isolated Salmonella ) exhibited a significant increasing trend over the same time period ( p = 0 . 033 ) . To better appreciate the population presenting at this healthcare facility with enteric fever , we investigated from which department within the hospital the blood samples that were positive for Salmonella had originated ( Fig 2 ) . Almost half of the S . Typhi and S . Paratyphi positive cultures ( 6 , 713/13 , 592; 49 . 4% and 3 , 298/6 , 091; 54 . 1% , respectively ) originated from the outpatient department . Notably , the proportion of S . Typhi and S . Paratyphi positive blood cultures arising from the emergency department increased from the year 2000 onwards; the emergency department continued to contribute a large proportion of Salmonella-positive blood samples after this year . Fig 3 shows the monthly time series for the Salmonella-positive cultures . June , July , August , and September were the months when the majority of Salmonella were isolated from blood cultures . The number of Salmonella-positive blood cultures ( all S . enterica , S . Typhi , and S . Paratyphi A ) exhibited maximal autocorrelation 12 months apart , therefore confirming seasonality of Salmonella infections ( S3 Fig ) . The time trends , which represent longer-term fluctuations in the number of positive blood cultures ( after filtering for seasonality and random variation ) , exhibited major peaks for S . Typhi and S . Paratyphi in 2002 , as well as smaller confined peaks for S . Typhi in 2005 , 2010 , and 2013 and S . Paratyphi in 2010 . Trends of the antimicrobial susceptibility profiles of the Salmonella isolated from blood between 1992 and 2014 are shown in Fig 4 . We observed a significant change in the susceptibility patterns against various antimicrobials over the study period , including increasing non-susceptibility against fluoroquinolones and increasing susceptibility against ampicillin and tetracycline for both S . Typhi and S . Paratyphi . We next investigated changes in susceptibility over the entire study period for the current first-line antimicrobials , ciprofloxacin , nalidixic acid ( as a marker of reduced susceptibility against fluoroquinolones ) , and cefotaxime . Segmented linear regression was exploited to detect changes in regression slopes between the proportion of non-susceptible organisms and time ( Fig 5 ) . A change in slope ( or “breakpoint” ) was interpreted as a shift in the rate of resistance acquisition at a particular time point . Non-susceptibility against nalidixic acid was consistently high for S . Typhi and S . Paratyphi from 2002 to 2014 , although a change in slope was observed for ciprofloxacin non-susceptibility between 2010 and 2011 for S . Typhi , and between 2009 and 2010 for S . Paratyphi . These breakpoints demarcated period when ciprofloxacin non-susceptibility remained low and constant and a sustained increase in ciprofloxacin non-susceptibility . For cefotaxime , a decrease in the proportion of non-susceptible S . Typhi was observed from 1993 to 1994 , this was followed by a period of low and constant non-susceptibility from 1996 to 2014 . For S . Paratyphi , non-susceptibility to cefotaxime was low and remained constant from 1995 to 2014 We lastly examined the proportion of MDR organisms through time ( Fig 5 ) . We observed a significant decrease in the proportion of MDR S . Typhi and S . Paratyphi from 1992 to 1997 and from 1992 to 1995 , respectively . This period was followed by a period of low MDR prevalence , which was maintained throughout the study period . Notably , in 2002 , a year with an atypically high number of Salmonella isolations , the proportion of MDR S . Typhi was irregularly high ( 19 . 7% ) , while the proportion of MDR S . Paratyphi isolates remained consistently low ( 1 . 0% ) .
Our study describes changes in the epidemiology and antimicrobial susceptibility patterns of Salmonella isolated in a Kathmandu healthcare facility from 1992 to 2014 . We document ( i ) a significant decline in the number of Salmonella-positive cultures from 2002 ( a year of heavy rainfall and flooding in the Kathmandu Valley ) to 2014 , ( ii ) an increase in the proportion of S . Paratyphi in all Salmonella-positive cultures between 1992 and 2014 , ( iii ) a decline in MDR S . Typhi and S . Paratyphi , and ( iv ) the recent increase in fluoroquinolone non-susceptibility in both S . Typhi and S . Paratyphi . In this comprehensive retrospective analysis of nearly 20 , 000 Salmonella-positive blood cultures we report that overall prevalence of Salmonella isolation from blood cultures was 8 . 8% . These data are consistent with data ( 7 . 2% and 13 . 4% ) recently reported in other studies from Nepal ( 10 , 11 ) . Over the 23-year study period , S . Typhi and S . Paratyphi A represented almost all Salmonella-positive cultures . However , their proportions varied over time: the proportion of S . Typhi ranged from >80% in 1992 to <60% in 2014 . These data are consistent with studies citing a prevalence of 65 . 1% for S . Typhi and 34 . 9% for S . Paratyphi A in Alka Hospital , a neighbouring Kathmandu hospital [10] . A prevalence of 57 . 8% for S . Typhi and 42 . 3% for S . Paratyphi A was reported in a tertiary hospital in Kathmandu in 2012–2013 [11] . From 1992 to 2014 , we observed a decreasing trend in the proportion of S . Typhi in all Salmonella-positive cultures; this decrease was concurrent with an increasing in the proportion of S . Paratyphi isolated . Similar increase have been reported in China [12] and Cambodia [13] and has been suggested to be a major shift in the aetiology of enteric fever in Asia [8 , 14] . The precise mechanism for this potential serovar replacement is unknown , but may be dependent on various environmental , ecological , or epidemiological factors [14] . Additionally , S . Paratyphi A does not harbour the Vi- polysaccharide capsule antigen , which is the key antigen used in all available S . Typhi vaccines in Nepal . We have only limited data on the uptake of Vi vaccine in this location , but a large Vi vaccine study was conducted in 2012 in all school-age children[15] . This program likely had a major impact on S . Typhi incidence and the apparent proportional increase in S . Paratyphi A cases . However , the vaccination program cannot entirely explain the proportional increase of S . Paratyphi A , which seems to have started before 2012 . The study period was not characterized by a uniform decline in Salmonella , but rather by a major peak in Salmonella-positive cultures in 2002 . From 2002 to 2014 , the absolute number of both S . Typhi and S . Paratyphi positive cultures declined . It is unclear whether the observed decline in S . enterica isolation from 2002 to 2014 was a return to baseline after a year of exceptionally high incidence in 2002 , or if it reflected a true decrease in Salmonella incidence . High numbers of S . enterica-positive blood cultures were also reported from nearby hospitals in 2002 . From May to July 2002 , a large single-point source outbreak of MDR S . Typhi associated with the water supply system was reported in Bharatpur , a city ~100 km southwest of Kathmandu , with almost 6 , 000 confirmed or suspected cases [16] . This spike in Salmonella infections is likely associated with exceptional rainfall in Nepal that year , resulting in severe flooding in the Kathmandu Valley in July . We have previously shown an association between rainfall and an ability to detect S . Typhi and S . Paratyphi A DNA in the local water supplies [17] . The major Salmonella outbreak in the middle of the study period is a reminder that the Kathmandu Valley retains an epidemic potential for enteric fever . While we describe a decrease in Salmonella incidence since 2002 , the residual enteric fever burden may flare up and cause large outbreaks under conducive conditions . Our data additionally documented a decrease in the percentage of MDR Salmonella isolates; this was observed for both S . Typhi ( 1992 to 1997 ) and S . Paratyphi ( 1992 to 1995 ) . Other researchers have suggested a decline in MDR-Salmonella in Nepal by showing that the majority of Salmonella isolated between 2011 and 2013 were susceptible to chloramphenicol , ampicillin , and co-trimoxazole [10 , 11] . Our longitudinal study confirms and expands these findings by presenting detailed kinetics in MDR Salmonella trends over the last 23 years: a decrease in the 1990’s followed by a long period of exceptionally low MDR prevalence . Nalidixic acid resistance was found to be high for both S . Typhi and S . Paratyphi from 2003 , the same year when antimicrobial susceptibility testing began to be performed routinely . Resistance to nalidixic acid is associated with a decline in fluoroquinolone susceptibility and the imminent emergence of fluoroquinolone resistance[18] . Indeed , we report a dramatic increase in ciprofloxacin non-susceptibility in both S . Typhi and S . Paratyphi from 2009 , approaching 100% non-susceptibility by 2014 . A gradual increase in ciprofloxacin non-susceptibility has been implied by various studies reporting ciprofloxacin-resistance ( or non-susceptibility ) for Salmonella in Nepal over the relevant years , ranging from 0% ( S . Typhi ) and 0 . 4% ( S . Paratyphi A ) non-susceptibility in 1993–1998[5] , 5% ( S . Typhi ) and 13% ( S . Paratyphi A ) non-susceptibility in 1999–2003 [5] , 2 . 8% ( S . Typhi ) and 10% ( S . Paratyphi A ) intermediate resistance in 2006–2007 [19] , 8 . 4% resistance in 2010 [20] , 0% ( S . Typhi ) and 3 . 3% ( S . Paratyphi A ) resistance in 2011–2012 [10] , and 83 . 1% non-susceptibility ( resistant or intermediate ) in 2012–2013 [11] . Furthermore , an increase in ciprofloxacin minimal inhibitory concentrations ( MIC ) was reported for S . Typhi between 2005 and 2014 , with a sharp increase in 2009 [21] . Therefore , our data confirms this trend of increasing ciprofloxacin non-susceptibility in S . Typhi and S . Paratyphi A , and details the kinetics of this transition over time . Notably , resistance to the third generation cephalosporins was low throughout the whole study period for both S . Typhi and S . Paratyphi A . Ceftriaxone remains the empirical antimicrobial of choice in this setting , which we have shown to be efficacious in those with culture confirmed enteric fever [22] . Though azithromycin is increasingly being used for the treatment of uncomplicated enteric fever , the drug was not used within the Patan Hospital treatment regimens until 2006 . Within Patan Hospital it was rarely used for the treatment of enteric fever until 2014 , when it was used as a “rescue drug” when the other antimicrobials failed . Due to the unavailability of break points for azithromycin against Salmonella , the microbiology laboratory did not perform any susceptibility tests until 2015 . We are therefore unable to make any conclusions on how azithromycin resistance has evolved for Salmonella within our population . However , as of 2015 , after the introduction of breakpoints for azithromycin on Salmonella by the CLSI , the hospital performs susceptibility testing and an ongoing surveillance monitors the MIC values . The fluctuations in AMR profiles described in this study largely reflect changes in clinical practices . MDR decreased as empirical treatment shifted from the fist-line antimicrobials ampicillin , chloramphenicol , or co-trimoxazole to ciprofloxacin in the 1990s . As the MDR phenotype is generally acquired via IncH1 resistance plasmids in S . Typhi [23 , 24] , the reduction of selective pressure after cessation of chloramphenicol , ampicillin and/or co-trimoxazole seems to induce the loss of these resistance plasmids and a reversion to a non-MDR phenotype . Subsequently , the routine use of ciprofloxacin to treat febrile diseases of presumed bacterial origin was likely associated with the dramatic increase in ciprofloxacin-non-susceptibility observed since 2009–2010 . Recently , co-trimoxazole has been successfully used to treat a fluoroquinolone-resistant S . Typhi infection in Nepal , opening the debate about the re-deployment of first-line antimicrobials [25] . Whilst currently efficacious against fluoroquinolone resistant organisms , including the H58 variant , there is an obvious risk of re-emergence of resistance with the reintroduction of the older antimicrobials , therefore methods such as antimicrobial cycling could be considered [21] . Our findings from almost 20 , 000 Salmonella enterica isolates highlight the need to review the current treatment protocols of enteric fever from fluoroquinolones as recommended by the WHO and local health ministries . There is sufficient evidence to use azithromycin or cephalosporins as the first antimicrobial therapy instead of fluoroquinolones [22 , 26] . With the re-emergence of efficacy in older antimicrobials used to treat enteric fever , they could be considered for use , but perhaps in combination with other antimicrobials to avoid an early re-emergence of resistance to these antimicrobials . However , treatment guidelines should be meticulously guided by microbiology data . Our analysis though comprehensive has a few limitations , being a passive surveillance focused on one particular hospital . One of our primary limitations is the use of different susceptibility guidelines used for resistance interpretation over time as the CLSI guidelines were unavailable locally prior to 2003 . We were not able to provide a consistent interpretation through the retrospective use of CLSI guidelines as the isolates were not stored and the disc diameters were not recorded . Another limitation is the population that is seen in the hospital , though the catchment area of the hospital has remained the same throughout the study period , many other health care facilities have opened up over time within this area where a certain subset population would go to seek medical care . However , Patan Hospital , being the biggest general hospital in the area still does see a majority of the patients within the catchment area . Hospital records showing individuals seeking care in Patan Hospital has increased consistently over the years , in accordance to the increase of the local population .
|
Aiming to understand the epidemiology and changing patterns of the Salmonella enterica serovars Typhi and Paratyphi A within a single healthcare facility in Kathmandu , we retrospectively analysed 23 years of microbiological blood culture data . From 224 , 741 blood cultures performed , 30 , 353 were confirmed to be positive for pathogenic bacteria , of which Salmonella enterica accounted for 65 . 4% ( 19 , 857/30 , 353 ) of all the bacteria positive blood cultures . S . Typhi and S . Paratyphi A were the dominant serovars , constituting 68 . 5% ( 13 , 592/19 , 857 ) and 30 . 5% ( 6 , 057/19 , 857 ) of all isolated Salmonellae . We observed that S . Typhi and S . Paratyphi A remained the leading cause of bacterial febrile illness since the 1990s . Additionally , antimicrobial resistance is a major public health challenge , and resistance against the antimicrobials most commonly used for treating enteric fever has developed over the last two decades . In this analysis we were able to document a decrease in the number of Salmonella-positive cultures from 2002 to 2014 and an increase in the proportion of S . Paratyphi A in all Salmonella-positive cultures between 1992 and 2014 . Concurrently , we also observed a decrease in MDR for S . Typhi and S . Paratyphi and a recent increase in fluoroquinolone non-susceptibility in both serovars .
|
[
"Abstract",
"Introduction",
"Material",
"and",
"methods",
"Results",
"Discussion"
] |
[
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"drugs",
"microbiology",
"salmonella",
"typhi",
"health",
"care",
"bacterial",
"diseases",
"signs",
"and",
"symptoms",
"enterobacteriaceae",
"pharmacology",
"bacteria",
"bacterial",
"pathogens",
"infectious",
"diseases",
"antimicrobial",
"resistance",
"medical",
"microbiology",
"microbial",
"pathogens",
"salmonella",
"enterica",
"salmonella",
"diagnostic",
"medicine",
"blood",
"anatomy",
"health",
"care",
"facilities",
"fevers",
"physiology",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2017
|
A 23-year retrospective investigation of Salmonella Typhi and Salmonella Paratyphi isolated in a tertiary Kathmandu hospital
|
Our ability to respond appropriately to infectious diseases is enhanced by identifying differences in the potential for transmitting infection between individuals . Here , we identify epidemiological traits of self-limited infections ( i . e . infections with an effective reproduction number satisfying ) that correlate with transmissibility . Our analysis is based on a branching process model that permits statistical comparison of both the strength and heterogeneity of transmission for two distinct types of cases . Our approach provides insight into a variety of scenarios , including the transmission of Middle East Respiratory Syndrome Coronavirus ( MERS-CoV ) in the Arabian peninsula , measles in North America , pre-eradication smallpox in Europe , and human monkeypox in the Democratic Republic of the Congo . When applied to chain size data for MERS-CoV transmission before 2014 , our method indicates that despite an apparent trend towards improved control , there is not enough statistical evidence to indicate that has declined with time . Meanwhile , chain size data for measles in the United States and Canada reveal statistically significant geographic variation in , suggesting that the timing and coverage of national vaccination programs , as well as contact tracing procedures , may shape the size distribution of observed infection clusters . Infection source data for smallpox suggests that primary cases transmitted more than secondary cases , and provides a quantitative assessment of the effectiveness of control interventions . Human monkeypox , on the other hand , does not show evidence of differential transmission between animals in contact with humans , primary cases , or secondary cases , which assuages the concern that social mixing can amplify transmission by secondary cases . Lastly , we evaluate surveillance requirements for detecting a change in the human-to-human transmission of monkeypox since the cessation of cross-protective smallpox vaccination . Our studies lay the foundation for future investigations regarding how infection source , vaccination status or other putative transmissibility traits may affect self-limited transmission .
Many infections only occur as isolated cases , short chains of transmission , or as small infection clusters ( i . e . intertwined transmission chains ) . Examples include zoonotic infections with relatively weak human-to-human transmission as well as vaccine-preventable infections in settings of high vaccination coverage [1]–[7] . Even though transmission is limited , these diseases are an important public health concern . For example , zoonotic infections can adapt for increased human-to-human transmission and then cause greater or even pandemic spread [8]–[10] . In addition , decreased voluntary vaccination , difficulty with vaccine delivery or changes in vaccine efficacy can allow growth of the number of individuals susceptible to preventable diseases and thus cause larger outbreaks [3] , [11] . Self-limited ( or subcritical ) transmission also characterizes diseases that are on the brink of elimination such as smallpox during its worldwide eradication campaign or polio today [12]–[14] . Despite a need to monitor disease burden , manage the risk of disease emergence or enhance disease elimination , the surveillance and control of subcritical infections can be challenging . Resource-poor countries , which are home to many zoonoses , have many logistical hurdles that impact the quality of surveillance and control interventions . Meanwhile , even in developed countries , reactive control strategies such as isolation protocols for vaccine-preventable diseases have significant sociological impact beyond the immediate financial costs . Because of these challenges , the overarching goal is to optimize control interventions for the least amount of effort and expense . It is therefore important to gain as much quantitative information about disease transmission as possible from existing surveillance data . This includes monitoring how transmission varies with time , location and other epidemiological characteristics of individual cases . By improving the understanding of mechanisms of disease transmission , finer tuning within the spectrum of intervention strategies becomes possible [15] , [16] . Such mechanistic understanding can guide the response to a diverse range of threats that include emerging infections ( e . g . , Middle East respiratory syndrome coronavirus ) , vaccine-preventable infections ( e . g . , measles ) and antibiotic resistance [17] , [18] . For ethical and logistical reasons , population-level studies of infectious disease transmission in humans typically involve retrospective statistical analysis rather than controlled prospective experimentation . Given this constraint , one approach for evaluating mechanisms underlying transmission patterns is to compare the transmissibility of two distinct , but related populations . In this manuscript , we demonstrate how the strength and heterogeneity of transmission can be compared for two different populations or types of infection sources . We then show how our framework provides insight into the transmission patterns of a variety of subcritical diseases . This analysis builds upon earlier studies that were limited to estimating transmission parameters from chain size distributions and addressing issues of surveillance bias [19] , [20] . Mathematically , the transmissibility of a group of infected individuals can be quantified by determining the group's effective reproduction number , . This number represents the mean number of secondary cases caused by an infected case . However , because of the stochastic nature of disease transmission , the realized numbers of secondary infections caused by a given infected individual will vary . is a more general parameter than the oft cited basic reproduction number , which more specifically represents the mean number of secondary cases caused by the first infected case in a completely susceptible population [21] . When , transmission cannot reach epidemic proportions , whereas if there is a potential for epidemic spread . Thus , our focus on subcritical diseases implies that , overall , will be less than one and transmission will be characterized by self-limited clusters of infection . However , our method still permits the possibility that cases can be divided into two groups in which one group has a , and the other group has a . Our study builds upon the prior success of inferring from the size distribution of observed transmission chains [1] , [2] , [22] . The same distributions can also be used to infer the degree of transmission heterogeneity , represented by the dispersion parameter , [19] , [20] , [23] . A high degree of heterogeneity represents a scenario where some individuals are predisposed to spreading infection to a larger number of people ( i . e . , ‘superspreaders’ ) . When models of chain size distributions incorporate both and , excellent agreement can often be found between observed data and model predictions [19] , [20] , [23] . Our goal is to evaluate specific hypotheses regarding disease transmission by testing whether and differ between two groups of cases . Our analyses differ from more traditional epidemiological approaches based on case-control studies ( and many other study designs ) in that we focus on transmissibility instead of individual-level risk factors for disease susceptibility . We demonstrate our methodology by considering four subcritical infections ( MERS-CoV , measles , monkeypox and smallpox ) and three types of data ( size distribution of infection clusters , transmission chain data and infection source classification ) to answer four different questions based on published data . For MERS-CoV , we use chain size distributions to determine whether an apparent decrease in during the latter half of 2013 was statistically significant . Assessing temporal trends of has important implications for evaluating the risk of endemic MERS-CoV transmission and the impact of control interventions . For measles , we use chain size distributions to compare two locations ( United States and Canada ) and test whether there is a significant difference in , which would suggest important differences in vaccine distribution , social connectedness , and/or demographics . For smallpox and monkeypox , we use case series resolved by infection generation to determine whether there are significant differences between the first and subsequent generations of spread [24] , [25] . This analysis allows us to assess whether variation in the number of contacts or the timing of control interventions can be linked to changes in . It also allows us to test the validity of a specific ‘random network’ model that relates the contact patterns of primary and secondary cases . We then test whether there is a significant difference between inferred transmission parameters for animal-to-human and human-to-human transmission of monkeypox , which provides insight into the mechanisms of zoonotic spillover . Our analysis of chain size distributions also provides perspective on the surveillance required to detect a change in , such as the expected increase in human monkeypox transmission following the eradication of smallpox . Each of the scenarios considered represents a unique example of how quantitative characterization of transmissibility can provide insight into the effectiveness of control interventions and risk assessment for future spread .
The stochastic nature of infectious disease transmission is particularly important when , as it can result in substantial variation in the size distribution of transmission chains . In this case it is helpful to model transmission as a branching process [26] . In this formulation , the offspring distribution specifies the probability that an infected individual will cause new infections . We specify the corresponding offspring probabilities to be , with . To facilitate likelihood calculations ( as seen below ) , the offspring distribution can be represented as a generating function , , in which the polynomial coefficients are the offspring probabilities [26]–[28] . In line with research demonstrating how the strength and variability of transmission can be modeled [23] , we assume the qi's follow a negative binomial offspring distribution with a mean of and a dispersion parameter of . The dispersion parameter represents the degree of transmission heterogeneity , with lower values of corresponding to higher variance . The supplementary methods ( Text S1 ) explains how our simple model of disease transmission can be used to calculate the likelihood for various types of observed data . These likelihood calculations permit inference of the strength and variability of transmission for individual cases , in terms of and . All calculations were conducted with either Matlab or R . Code for all analyses is available at: https://github . com/sbfnk/nbbpchainsizes . By calculating the likelihood of an observed set of transmission events , we can probe whether there is statistical support for differences in transmission between two pre-specified populations , and . In our general model , the two types of individuals have distinct negative binomial characterizations and thus there are four parameters in total . We label these four parameters , , and with the subscripts corresponding to the type of individual . Five simpler models that are nested within the 4-parameter model can be constructed by assuming , and/or ( Figure 1 ) . The specific test case of is chosen for the nested models because this corresponds to a geometric offspring distribution which is the expectation for a traditional SIR or SEIR model . These models assume homogenous mixing with constant infectivity over an exponentially distributed infectious period [29] . For each model , we determine the parameter values ( MLE ) that maximize the log-likelihood . The 95% confidence intervals and confidence regions shown in the figures were found by profiling on and/or and employing the likelihood ratio test [30] . Model comparison is accomplished via the Akaike Information criterion ( AIC ) [31] . To identify whether there is statistical support for a difference in for two data sets , the AIC scores were computed for all six aforementioned models . A difference in was deemed statistically significant according to the rule that the model with the best AIC score cannot be within two AIC units of a model that supports identical values of for the two sets of simulations . This rule is in approximate alignment with the commonly used likelihood ratio test for establishing statistical support for the use of an extra parameter with 95% confidence , but we could not employ the likelihood ratio test explicitly because some pairs of models we consider are not nested . We verified the internal consistency of our modeling framework by applying this method to simulated data ( Supplementary material , Text S1 ) . We used parametric bootstrapping to evaluate the type I error and the power for detecting a change in for our analyses . Specifically , for every analysis we simulated 20 , 000 new data sets . Each simulated data set replicated the two populations involved in the analyses ( e . g . MERS-CoV chains before and after June 1 , 2013 ) . Two models were simulated . Half of the simulations used two distinct values of and that matched the inferred values of our unrestricted four-parameter model . The other half of the simulations used a single value of and that matched the inferred values of our two-parameter model , which requires both and to be the same for all cases seen in the observed data . Our inferential algorithm for ascertaining a statistically significant difference in the inferred value of was then applied to all simulations . The type I error of an analysis ( i . e . the probability that the analysis would falsely claim that is different for the two types of cases considered ) was estimated as the proportion of simulations based on the two-parameter model that were found to have a statistically significant difference in for the two types of cases . The parametric bootstrap probability ( or power ) of detecting a change in was estimated as the proportion of simulations based on the four-parameter model that were found to have significant difference in for the two types of cases .
Since 2011 , there have been over 500 confirmed cases of MERS-CoV , and over 140 associated deaths , suggesting a case fatality rate of 28% [32] . The persistent occurrence of small outbreaks is due to zoonotic spillover [33]–[35] . MERS-CoV may be a new virus , as the most recent common ancestor of viral samples from infected patients was estimated to have occurred after September 2010 [34] . The novelty of this virus and its high case fatality rate underscore the significance of monitoring the transmission of MERS-CoV . Although human-to-human transmission has been relatively limited so far , with likely less than one , there is concern that future adaptation that could lead to spread similar to sudden acute respiratory syndrome ( SARS ) in 2003 . Health authorities have prudently instituted a variety of infection control policies and procedures and a trend towards decreasing has been reported [34] . Since verification of the effectiveness of control has important implications , we reconsidered the evidence for a trend towards decreasing . To avoid artifacts of assembling multiple data sources , we restricted our analysis to the previously reported chain size distribution for all MERS-CoV cases in the Arabian Peninsula occurring before August 8 , 2013 [34] . Previous analysis of these data shows that is 0 . 74 ( 95% CI 0 . 53–1 . 03 ) before June 1 , 2013 and 0 . 32 ( 95% CI 0 . 14–0 . 65 ) after June 1 , 2013 . Our results replicate the finding that independent evaluation of cases before and after June 1 , 2013 results in an estimate of 0 . 7 and 0 . 3 for respectively ( Figure 2 and Table 1 ) . When our six models are compared , we do not find statistical support for models with different values of before and after June 1 , 2013 . This is again consistent with the results of prior studies that determined a p-value of 0 . 07 for change in , but our analysis allows the possibility of a high degree of transmission heterogeneity . Local elimination of measles is dependent on vaccination programs , and the potential for re-emergence necessitates continued surveillance and re-assessment of vaccination strategy [1] , [3] , [36]–[38] . Even where elimination has been achieved , there can be sporadic clusters of infection due to a combination of geographic importation and pockets of susceptibility [39]–[41] . Geographical differences in transmission may arise due to differences in cultural practices , public health guidelines , population density and other factors . Methods that delineate whether differences in are statistically significant for two different regions can therefore help to identify key differences in transmission potential and thus pinpoint opportunities for improved control . Measles data in the United States ( 1997–1999 ) and Canada ( 1998–2001 ) are reported according to the size of infection clusters [39] , [40] . Most infection clusters have a single primary infection , but even when multiple primary infections exist ( as in the case of a cluster with six cases in the United States ) , the likelihood calculation needed for assessing differences in is straightforward ( Supplementary Material , Text S1 ) . When the two data sets are compared , the results indicate that for the United States and Canada are significantly different ( Figure 3 and Table 2 ) . Meanwhile , the results also confirm previous studies that infer a high degree of transmission heterogeneity in measles transmission [19] , [23] . This can be seen from Table 2 since the MLE estimates for and are less than one and the value of the model with is large . On the other hand , there is negligible statistical support for distinct values of in the two countries . The type I error for this situation was estimated to be 4 . 9% by parametric bootstrapping . Smallpox is the only human disease to have been eradicated and thus represents a tremendously successful use of control [12] . During the endgame of smallpox eradication in the middle of the 20th century , smallpox cases in Europe resulted in rapid implementation of quarantine and control procedures . Transmission data for smallpox infections in Europe that occurred during this period provide an opportunity to investigate how control interventions impacted the transmissibility of primary cases caused by geographic importation relative to secondary cases resulting from local transmission [12] . Smallpox clusters were tabulated according to the number of cases in each generation of spread [12] . The inference results indicate that secondary cases transmitted significantly less than primary cases ( seen by the lack of overlap of contours with the grey line in Figure 4 and by the statistical selection of the non-restricted model in Table 3 ) . In fact , the effectiveness of control procedures can be quantified by looking at the ratio of reproduction numbers for primary and secondary transmission ( Figure 4 inset ) . The ratio of the maximum likelihood values for to suggests that control reduced by 75% . Meanwhile , for both primary and secondary transmission , a high degree of transmission heterogeneity is evident ( since the MLE estimates of and are substantially less than one and the value of the model is large ) . Based on selection of the unrestricted model , and the associated estimates of , there appears to be significantly more heterogeneity of disease transmission for secondary cases than for primary cases . The type I error for this analysis was estimated to be 5 . 1% by parametric bootstrapping . Following the eradication of smallpox in 1979 , the World Health Organization was concerned that subsequent cessation of smallpox vaccination would allow other diseases to flourish [42] . Monkeypox was of particular concern because exposure to smallpox or smallpox vaccination provided protection against monkeypox . Estimates of , extrapolated from contact tracing data gathered during rigorous surveillance in the Democratic Republic of Congo ( formerly Zaire ) during 1981–1984 , provided re-assurance that endemic transmission would not be sustainable even when population immunity to monkeypox waned [43] . The initial analysis of monkeypox transmission did not quantitatively compare the transmission of primary cases ( i . e . those caused by animal-to-human transmission ) to the transmission of secondary cases ( i . e . those caused by human-to-human transmission ) . Since the characteristics of these cases differ ( i . e . only primary cases required exposure to infected animals ) , differences in transmission are possible . Increased transmission of secondary cases could also arise from population structure [25] , or evolutionary adaptation [8] , [10] . For example , network models have proposed that social structure impacts the effective reproduction number of individual cases [44]–[48] . In particular , the random network model that we have considered ( Supplementary material , Text S1 ) predicts that secondary cases transmit more than primary cases since highly-connected individuals are most likely to both acquire and spread infection . If this aspect of the random network model is accurate , the risk of endemic spread as population immunity wanes may be higher than previously expected . This is because for secondary transmission would be expected to increase more than for primary transmission . It is thus important to ascertain whether there is a difference between primary and secondary transmission that is consistent with the random network hypothesis . As part of the monkeypox surveillance efforts , transmission was tabulated according to the number of cases in each generation of spread [43] , [49] . These data can be used to ascertain whether there is a statistically significant difference in primary versus secondary transmission ( Figure 5 and Table 4 ) . The results indicate a lack of evidence for a difference between the of primary and secondary cases ( seen by noting the overlap of contours with the grey line in Figure 5 and because the preferred model in Table 4 has ) . The low values for the maximum likelihood estimates of are consistent with previous studies that infer a high degree of transmission heterogeneity in monkeypox transmission [20] , [23] . Animal-to-human transmission of monkeypox is an important contributor to overall disease burden . Determining the factors that allow continual introduction of monkeypox into human populations requires knowledge of how monkeypox maintains itself in reservoir hosts and the mechanisms that allow its transmission to humans [6] , [50] . In this section we assess whether an infected animal in contact with humans has a distinct set of inferred transmission parameters than infected humans . The relationship between infection source and transmissibility is an active area of research for many multi-host diseases systems [51]–[55] , particularly for zoonotic infections . Since the infection cluster data for monkeypox contains information on how many primary infections are in each cluster , it can be used to infer the amount of animal-to-human transmission that occurs when infected animals make contact with humans . To accomplish this , we assume that the negative binomial offspring distribution that has been shown to be a good description of human-to-human transmission [23] is also an effective model of animal-to-human transmission . We let represent the average number of primary cases caused by an infected animal that has contact with humans . Our results indicate that the for human-to-human transmission is similar to ( Figure 6 and Table 5 ) . There is also evidence that animal-to-human transmission is relatively homogeneous ( since the for the preferred model ) . If one takes the MLEs of and for the preferred model at face value , then we estimate that at least one infection occurs 25% of the time that a infected animal has contact with humans . Recently , a 20-fold increase in the incidence of monkeypox has been reported in the Democratic Republic of Congo [56] , and there is concern that for monkeypox may have increased . The lack of cross-protective immunity to monkeypox from either smallpox vaccination or natural exposure to smallpox provides a mechanism for why would increase [57] . However , land-use changes that impact the potential for animal-human transmission have also been suggested as a cause of an increase in monkeypox incidence [58] , [59] , and could do so without changing . There are no active interventions in place for monkeypox , so it is important to determine if has changed in order to understand the source of increased incidence . Due to logistical barriers and the rare nature of the disease , acquiring data on monkeypox is a challenge [42] , [56] . In the wake of smallpox eradication , the infrastructure for monkeypox surveillance in 1980–1984 was strong and well funded [42] . The detailed transmission data from this surveillance effort provide an estimate of 0 . 30 for ( 95% CI: 0 . 21–0 . 42 ) and 0 . 33 for ( 95% CI: 0 . 17–0 . 75 ) [20] . For the 2005–2007 surveillance effort , specific data on cluster sizes and individual-level transmission are unavailable , so an assessment of cannot be made . However , we can quantify the amount of data that would be needed in order to detect a change in relative to 1980–1984 [42] , [43] , [49] . Simulations show that 200 clusters would provide 70% power to detect an increase in from 0 . 3 to 0 . 5 ( Figure 7A ) . As the number of observations increase , smaller changes are more readily noticeable . Consideration of the relationship between , the number of chains and the number of cases provides perspective on the power of the recent surveillance efforts ( 2005–2007 ) to detect a change in [56] . It appears that there is 95% power to detect an increase in from 0 . 3 to 0 . 55 with analysis of the 760 observed cases ( Figure 7B ) .
To reduce the burden of MERS-CoV and reduce the risk of global spread , effective control procedures are of obvious importance . Given the large amount of resources and effort that have already been directed towards the control of MERS-CoV , it would be reassuring to see a statistically significant decrease in . When analyzing data on MERS-CoV cases that presented before Aug 8 , 2013 , the unrestricted model had the best score . This unrestricted model suggested that because decreased from 0 . 7 to 0 . 3 , control is over 50% effective . However , there is not enough data to show statistical significance for this result . Meanwhile , our analysis is likely biased by the large outbreak that initiated the observational period for the data , so further studies are needed to more accurately evaluate the impact of control interventions [60] . Unfortunately , the number of recent confirmed MERS-CoV cases remains significant and the overall incidence may be increasing [32] . An increase in the number of cases can be caused by an increased , an increased rate of primary cases , or a combination of these effects [61] . Based on our observation that is more likely to be decreasing after June 2013 than increasing , the paradigm of emergence that is most consistent with the previously published data we have analyzed is that MERS-CoV incidence may be increasing in its non-human reservoir , but that human-to-human transmission remains stable . In fact , sequence data support the possibility of an expanding epidemic in animal hosts of MERS-CoV that could lead to an increased incidence of primary cases [34] . However , other factors , such as seasonal drivers of transmission could also impact the temporal trend of . An increased case load could also be observed if transmission patterns have not changed much , but greater interest in and knowledge of MERS-CoV has led to improved surveillance . This could paradoxically lead to both an increase in the number of observed cases and a decrease in the observed value of because of a greater chance of seeing a larger proportion of smaller outbreaks [19] , [62] . Given the relative paucity of cases and uncertainties regarding case observation probability , it would be inappropriate to make a definitive statement concerning the cause of the apparent increase in MERS-CoV incidence at this time . However , as more data on MERS-CoV are reported , the types of analyses presented in this manuscript can be rapidly applied to address hypothesis-driven questions concerning the temporal trends of incidence and the impact of control intervention . In particular there may be concerns that certain subgroups of MERS-CoV cases may have increased transmission , such as those occurring in health care settings where nosocomial transmission is higher or in geographic regions where control interventions are harder to implement . Alternatively , as we have shown with smallpox , there may be a difference in the transmissibility of primary cases versus secondary cases . With more data , our method can help to quantify differences in transmission , and evaluate whether certain population subgroups may have an that exceeds the critical value of one . While it is not necessary for future data to be resolved to the level of individual transmission events , the types of analyses we have presented do require knowledge of chain size distributions rather than aggregate epidemic curve data . Meanwhile , an important gap in the currently available data is a quantitative assessment of the case reporting probability for MERS-CoV cases and whether this is increasing with time . Improved knowledge of the reporting probability would permit adjustments to the likelihood calculations and reduce the bias of imperfect case ascertainment [19] . Our comparison of measles transmission in the United States and Canada provides a framework for elucidating geographic differences in transmission ( Figure 3 ) . Interestingly , while our analysis supported a difference in between the two countries , a difference in the degree of transmission heterogeneity ( as quantified by the dispersion parameter ) was not identified . This apparent disassociation between the strength of transmissibility and the mechanisms of transmission heterogeneity may occur if the heterogeneity is due to intrinsic biological processes such as variability in viral shedding . However , the relationship between the value of dispersion parameter and various mechanisms of transmission heterogeneity is not straightforward so the interpretation of similar values of dispersion is unclear . There are many reasons why the value of may differ between the United States and Canada . One consideration is a potential difference in the timing of the introduction of two-dose vaccination . The Advisory Committee on Immunization Practices and the American Academy of Pediatrics recommended two-dose coverage in 1989 [63] . Although the coverage in 2004 appeared similar between the United States and Canada [38] , it is unclear whether this level of coverage was achieved at the same time in both countries . To assess whether a difference in vaccine coverage explains the difference in observed here , it would be helpful to run a similar analysis on more recent data . Other factors that could contribute to the difference in include a greater tendency in the United States to conduct contact tracing for susceptible cases and vaccinate close contacts , a greater sensitivity in Canada for reporting milder cases of measles , or greater difficulty of detecting isolated cases via passive surveillance in Canada [37] , [38] . More detailed information of the impact of contact investigation , stratification of cases based on disease severity , and quantitative comparison of case ascertainment in passive versus active surveillance would provide additional insight . Smallpox control is already known to have been very effective; however , our analysis of smallpox transmission in Europe around the time of eradication quantifies the impact of interventions for control ( Figure 4 ) showing that there was a reduction of for secondary cases by 75% compared to primary cases . This effect of control may be an underestimate because it does not account for the possibility of late arrival of imported cases during the course of infection . Since the infectious period of imported primary cases may have occurred outside of the country of residence , the actual for primary cases might be higher than seen in the data and thus the effect of control may be even greater than our estimates indicate . Here we have shown how for each generation can be quantitatively compared , using published transmission data . Our analysis of differences in the transmissibility of cases as an outbreak develops is not unique ( see for example [64] ) . However , previously published methods rely on symptom-onset data to determine at various stages of an outbreak and thus these approaches could not be performed on the smallpox data set . Aside from the change in , the marked increase in degree of transmission heterogeneity for secondary cases ( as evidenced by a decreased in the observed value of ) suggests that control tended to be individual-specific rather than population-wide . Here , individual-specific control refers to an intervention that is completely effective for 75% of cases but not effective at all for the remaining cases , whereas population-wide control refers to an intervention that reduces the transmissibility of each case by 75% [23] . For individual-specific control , a large number of cases become dead ends for infection so the observed degree of heterogeneity increases [19] , [20] . In contrast , the observed degree of transmission ( as quantified by the dispersion parameter ) would not change for population-wide intervention . The support for individual-specific control is highly consistent with the quarantine and ring vaccination methods employed during smallpox elimination efforts [12] . These observations show how understanding the variation in both the strength and heterogeneity of transmission can provide insight into disease dynamics . Our analysis of monkeypox in the Democratic Republic of Congo demonstrates how our method can be used to inform surveillance planning . In particular , by determining the number of chains that needed to be observed in order to detect various degrees of change in , we provide perspective regarding the extent to which the 760 monkeypox cases observed between 2005 and 2007 [56] can provide enough information to detect increased transmissibility ( Figure 7 ) . Based on our power analysis , it appears that a change in due to declining population immunity should be detectable , since is expected to approach [43] . However , this result needs to be interpreted in context because our model assumes that the probability of case observation is high and that distinct infection clusters can be determined . Given the logistical challenges of recent surveillance efforts [56] , these assumptions are unlikely to have been met , so the realized power for detecting a change in is probably lower . Nevertheless , this simulation analysis provides perspective concerning the trade-offs of thoroughness in detecting and characterizing cases versus observing cases within a greater catchment area for any future surveillance efforts for which measurement of is of interest . When we focused on more detailed generation-level data for monkeypox transmission from 1980–1984 , we found no support for enhancement of by highly-connected individuals in secondary generations ( Figure 5 ) . This suggests that the high degree of transmission heterogeneity may be caused by biological factors , rather than variability in social contact . However , a key assumption of the network model we tested is that primary cases are infected at random relative to their degree ( as might reasonably be expected for a zoonotic infection ) . It may be that high-connected individuals are also more likely to get a primary infection . If this were the case , then highly connected individuals would contribute to heterogeneity of both primary and secondary transmission . Meanwhile , the lack of increased for secondary transmission provides assurance that significant viral adaptation is not occurring , although local depletion of susceptible individuals within small sub-networks such as households could obscure signals of viral adaptation . We found that humans and animals in contact with humans produce similar numbers of human cases ( Figure 6 ) . Moreover , we estimated that 25% of human exposure to an infected animal lead to at least one detected human case . While the truncated negative binomial distribution produces unbiased estimates of transmission parameters , the confidence intervals can be quite large [19] . Furthermore , the a priori specification that the offspring distribution will be characterized by negative binomial distribution is a strong assumption . Thus the inferred proportion of animal-to-human exposures leading to infection deserves cautious interpretation . Nevertheless , this type of analysis could be useful for informing surveillance and detection efforts in wildlife species . In particular , since the overall incidence of monkeypox is quite low ( 14 . 42 per 10 , 000 per year [56] ) , the observation that there may be only 4-fold more infected animals in contact with humans than the number of observed infection clusters provides perspective on the fact that monkeypox virus has only been isolated from one wild animal ( as of 2011 ) [58] . If contacts with infected animals account for a small proportion of overall human contact with reservoir species , the use of targeted-surveillance strategies that can exploit spatial-temporal data to identify likely hotspots of incidence [58] , [59] , [65] may be essential to improve detection efforts in wildlife hosts . As with any model selection or measurement scheme , a small portion of the data , or even a single data point , can have a particularly large influence . For example , the largest transmission chain in the Canadian measles data consists of 155 cases while the second largest chain has just 30 cases . Moreover , the chain with 155 cases was associated with a religious community that resisted immunization , thus it could be argued that this chain is not representative of the population as a whole . If the 155-case chain were excluded from the analysis , our method would no longer find statistical support for a difference in between the United States and Canada ( Supplementary material , Text S1 ) . However , rather than excluding a possible outlier , our preference is to treat the data at face value . From a modeling perspective , it is often unclear whether the mechanism responsible for a purported outlier is absent in the rest of the data . For example , in the case of Canadian measles data set , the second largest chain of 30 cases was also associated with a religious community . In addition , a particularly large chain does not represent a single large transmission event , but rather an entire group of individuals who collectively had relatively high transmission . Mathematically , a high degree of transmission heterogeneity ( represented by low values of ) is expected to have a big tail for the distribution of the number of cases that each case causes [23]; thus , a large transmission event or chain in a set of data will increase the estimated value of , but will also decrease the estimated value of . A lower will be associated with a wider confidence interval for and this would make it harder for our analysis to find a statistically significant difference in [19] , [20] . Thus our modeling framework has a built-in mechanism that compensates for large transmission events and chains that are consequences of intrinsic population-level or individual-level mechanisms of heterogeneity . A key caveat of our analyses is that we have assumed perfect observation of cases . Some surveillance programs , such as measles in the United States , have documented evidence of high case observation [36] . However , this level of case ascertainment cannot be expected of all diseases , particularly those such as MERS-CoV that are quite new . Meanwhile , even meticulously collected data are prone to multiple sources of observation bias due to limited surveillance resources , subclinical infections , laboratory error , or other factors . When the limitations of observation can be quantified , likelihood calculation for observed transmission events can be adjusted appropriately [19] , [20] . The challenge is that the limitations of surveillance systems and case ascertainment are often difficult to quantify . An alternative to explicit correction of observation bias is to simply consider what level of observation bias would impact key results . For example , in our analysis of the difference between animal-to-human and human-to-human transmission of monkeypox , it is quite possible that a number of animal-to-human infections are unobserved — particularly if the resulting primary infection is mild and has no further transmission . When we treated observation of an infection cluster as an all-or-none process with an independent probability , , that each case would activate surveillance ( thus implying many isolated cases would be unobserved ) , our preferred model of transmission remained stable even for a of 0 . 1 ( Supplementary material , Text S1 ) . This provides re-assurance that our methodology is not necessarily sensitive to imperfect observation . However , different data sets or a different type of observation bias could yield less stable results . In our analyses , we have allowed for at most two values of and in a data set rather than permitting additional stratification or a continuous distribution of values . These simplifications are not always valid assumptions . However , modifications to the likelihood calculation can often be made in order to accommodate more complicated data sets so that our framework for detecting a difference in can be utilized . For example , the offspring generating function used for the likelihood calculation can be written in terms of a continuous variable that provides a smooth transition between the extreme limits of classification . In fact this approach has been used to investigate whether there is a temporal trend of measles transmissibility in the United States [61] . Although we have mainly focused on differences in between two populations , our method can also be used to identify whether these populations differ in the observed degree of heterogeneity . Clustering of individuals with higher transmissibility may favor models with two distinct values for whenever two distinct values of are observed . Meanwhile , situations that would favor a model with two distinct values of and one value of could arise if different mechanisms of control were used to maintain below a given threshold , as seen in the smallpox example . Regardless of which model is the preferred model for a given data set , the estimated or assigned value of can be useful to assess the overall degree of transmission heterogeneity and the likely presence of super-spreaders [20] , [23] . On the other hand , the specific mechanism of heterogeneity ( e . g . differences in transmission potential among cases versus clustering of susceptible individuals ) cannot be ascertained from estimation of alone . Our analysis is focused on determining whether there is statistical support for a difference in for individuals having a specific trait . Also , as exemplified by our direct comparison to the random network model ( Figure 5 and Table 4 ) , we can evaluate specific models of transmission . However , in the absence of a mechanistically derived model , our analysis cannot identify the cause of differences in . For-example , population-level factors favoring transmission ( e . g . increased human density ) cannot be directly distinguished from biological factors ( e . g . evolutionary adaptation ) . Furthermore , the decrease in secondary transmission due to local depletion of susceptibles cannot be directly distinguished from decreases due to control mechanisms . Instead , our method needs to be considered as a tool that can identify differences in transmission ( e . g . temporal trends for MERS-CoV , and geographic distinctions in measles ) or quantify changes in transmission that are expected to occur ( e . g . decreased transmission due to quarantine of smallpox cases or ring vaccination ) . By addressing diverse questions within varied data sets , we have demonstrated that a set of inter-related models within a branching process framework allows rigorous statistical assessment of whether particular characteristics of infectious cases impact transmission potential . We have focused on subcritical diseases , in large part because the type of surveillance data gathered for these diseases is most compatible with our computational approach . For MERS-CoV , we evaluated the possibility of a temporal trend towards decreasing that may indicate stronger control , but did not find enough statistical evidence to confirm this finding . For measles , we found evidence of geographic variability that provides potential insight into the effectiveness of surveillance and public health interventions . For smallpox , we identified signatures of effective control by comparing primary and secondary transmission . For monkeypox , we found that the most parsimonious models are ones that incorporate a high degree of transmission heterogeneity , but do not differentiate between animal-to-human transmission , transmission of primary cases , and transmission of secondary cases . In general , the statistical support we observed for models that allow flexible inference of both and reinforces the importance of quantifying both the strength and variability of disease transmissibility . By providing a diverse array of applications and analyses , the method we have demonstrated can increase the value of existing surveillance data and improve strategies for future data collection . Through identifying specific risk factors for transmissibility and by assessing different sources of transmission heterogeneity , we hope that disease monitoring and control interventions can become more targeted and thus more effective .
|
The goal of this paper is to identify epidemiological factors that correlate with either an increased or decreased risk of transmitting a particular disease . We are particularly interested in identifying such factors for diseases that are self-limited ( meaning that infections tend to occur in isolated clusters ) , because targeted control of these diseases can facilitate public health goals for minimizing the risk of disease emergence or promoting disease elimination . For example , we show that there is a significant difference in the transmission of measles between the United States and Canada . In contrast , we find that an observed decrease in the transmission of Middle East respiratory syndrome coronavirus during the latter half of 2013 cannot be ascertained with sufficient confidence . We then quantify the degree to which control was effective in eradicating smallpox in Europe . We also consider how the transmission of monkeypox in humans depends on whether the infection source is an animal or a human . Finally , we demonstrate how our approach can be used by surveillance programs to detect changes in transmission that may occur over time .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"infectious",
"disease",
"epidemiology",
"population",
"dynamics",
"immunology",
"epidemiological",
"methods",
"and",
"statistics",
"preventive",
"medicine",
"plant",
"science",
"mathematics",
"statistics",
"(mathematics)",
"population",
"modeling",
"global",
"health",
"plant",
"pathology",
"population",
"biology",
"biostatistics",
"vaccination",
"and",
"immunization",
"veterinary",
"science",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"veterinary",
"diseases",
"zoonoses",
"disease",
"ecology",
"theoretical",
"biology",
"smallpox",
"epidemiology",
"travel-associated",
"diseases",
"infectious",
"disease",
"modeling",
"disease",
"dynamics",
"measles",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"computational",
"biology",
"physical",
"sciences"
] |
2014
|
Detecting Differential Transmissibilities That Affect the Size of Self-Limited Outbreaks
|
Parasitic helminths release molecular effectors into their hosts and these effectors can directly damage host tissue and modulate host immunity . Excreted/secreted proteins ( ESPs ) are one category of parasite molecular effectors that are critical to their success within the host . However , most studies of nematode ESPs rely on in vitro stimulation or culture conditions to collect the ESPs , operating under the assumption that in vitro conditions mimic actual in vivo infection . This assumption is rarely if ever validated . Entomopathogenic nematodes ( EPNs ) are lethal parasites of insects that produce and release toxins into their insect hosts and are a powerful model parasite system . We compared transcriptional profiles of individual Steinernema feltiae nematodes at different time points of activation under in vitro and in vivo conditions and found that some but not all time points during in vitro parasite activation have similar transcriptional profiles with nematodes from in vivo infections . These findings highlight the importance of experimental validation of ESP collection conditions . Additionally , we found that a suite of genes in the neuropeptide pathway were downregulated as nematodes activated and infection progressed in vivo , suggesting that these genes are involved in host-seeking behavior and are less important during active infection . We then characterized the ESPs of activated S . feltiae infective juveniles ( IJs ) using mass spectrometry and identified 266 proteins that are released by these nematodes . In comparing these ESPs with those previously identified in activated S . carpocapsae IJs , we identified a core set of 52 proteins that are conserved and present in the ESPs of activated IJs of both species . These core venom proteins include both tissue-damaging and immune-modulating proteins , suggesting that the ESPs of these parasites include both a core set of effectors as well as a specialized set , more adapted to the particular hosts they infect .
Parasitic nematodes continue to be a major source of mortality and morbidity worldwide , infecting nearly 25% of the global population [1 , 2] . The molecules that are released by these parasites , including the excreted/secreted proteins ( ESPs ) , represent the major interface between hosts and parasites , and directly influence the survival and health of the parasites as well as the pathology they cause to the hosts [3 , 4] . Despite an abundance of studies addressing mechanistic aspects of host immune response to nematode parasites , there is a distinct paucity of molecular information about most parasitic nematodes , where few secreted molecules have been studied in detail . Further , the role of the parasite ESP composition in determining host specificity is unknown . What is known relies largely on ESP studies where release of the ESPs is stimulated and collected in vitro . An underlying assumption is that the ESPs collected under these conditions are relevant and similar to the ESPs released in in vivo infections , though this assumption has not been experimentally validated for most if not all such studies [3] . Obtaining enough ESPs from nematodes that are actively involved in a host infection for subsequent analysis is difficult if not impossible . However , sequencing the transcriptomes of individual nematodes [5 , 6] , provides a way of comparing transcriptional profiles of parasites undergoing in vitro activation and in vivo infection . Entomopathogenic nematodes ( EPNs ) are parasites of insects that rapidly kill their hosts . When EPNs deplete host nutrients the developing generation emerges from the cadaver as infective juveniles ( IJs ) , an alternative third-stage larval form ( L3 ) that is developmentally arrested , similar to the dauer juvenile stage in C . elegans [7] . The IJs are the only free-living stage of these nematodes , and they actively seek hosts to infect [8 , 9] . Upon entering a new host , the IJs undergo the process of activation , or recovery from dauer , which entails resumption of growth and development , along with changes in morphology and gene expression that facilitate transition from a free-living form to an actively parasitic form [5 , 10–12] . EPNs are being used as models for host-parasite interactions including ecology [13 , 14] , host-seeking behavior [9 , 15] , neurobiology [8] , parasite activation [5 , 16 , 17] , and the role of secreted products in parasitism [5 , 18 , 19] . There are more than 70 described species of EPNs in the genus Steinernema , and these vary in their host range and specificity [20 , 21] , making these nematodes a potential model for understanding the evolution of ESPs and their role in niche partitioning among parasites . For example , S . carpocapsae is a generalist parasite capable of infecting more than 250 different species of insects from at least 13 orders [22 , 23] , while other species such as S . scapterisci and S . scarabaei are specialist parasites infecting a much narrower range of species [24 , 25] . A recent study of the S . carpocapsae secretome found that this generalist parasite releases more than 450 different proteins when initiating active parasitism . Many of these proteins were hypothesized to be involved in tissue damage and immunosuppression of the host [5] . S . feltiae is another generalist EPN parasite but with a more limited host range than S . carpocapsae and in a different clade within Steinernema [26 , 27] . Several studies have shown that S . feltiae IJs use their cuticle to suppress and evade host immunity [28–30] . It has even been postulated that unlike S . carpocapsae , S . feltiae does not use secretion processes or secreted proteins to induce host immunosuppression [31] . Here we utilized RNA-seq from individual S . feltiae nematodes throughout a time course of in vitro and in vivo activation to compare the induction of ESPs under these different conditions . We reported the secretome of S . feltiae and tested its activity in vivo . We showed that activated S . feltiae IJs release a variety of proteins likely involved in tissue damage as well as immune modulation . By analyzing the in vivo time course of activation , we identified putative neuropeptide pathway genes likely to be involved in host-seeking behavior as the expression of these genes decreased as the nematodes is activated . Further , using comparative analysis we identified a core suite of 52 ESPs released by both S . feltiae and S . carpocapsae during active parasitism , indicating that despite differences in host range and specificity , some proteins may be broadly useful in parasitizing insect hosts . Most of these core proteins are conserved in nematode parasites of mammals , suggesting that they have an important and conserved role in parasitism .
We utilized an in vitro activation method previously used for S . carpocapsae and S . scapterisci [5 , 17] to determine how S . feltiae IJs activate . We exposed S . feltiae IJs to insect homogenate and found that they activated in a manner similar to what has been described for S . carpocapsae and S . scapterisci ( Fig 1 ) . Expansion of the pharyngeal bulb was found to be a reliable indicator of IJ activation [5 , 16 , 17] and this feature was used to quantify activation . In naïve IJs ( IJs not exposed to host tissue ) the pharyngeal bulb is often difficult to observe at 400x magnification ( Fig 1A ) . At 1000x magnification ( Fig 1B ) the pharyngeal bulb can be seen , however the bulb is typically more compressed , seemingly deflated , when compared to activated nematodes . As IJs are exposed to host tissue over time they begin exhibiting partially-activated morphology characterized by partial expansion of the pharyngeal bulb ( Fig 1D ) which , in contrast to naïve IJs , is more expanded and can be readily observed at 400x ( Fig 1C ) . These differences allow us to quickly and efficiently differentiate between non-activated and activated IJs under 400x magnification . After 6 hours of exposure to insect tissue , approximately 25% of IJs exhibit fully activated morphology with full expansion of the pharyngeal bulb , which is wider and appears rounder than the oval shape of partially activated nematodes ( Fig 1F and 1D ) . Similar to what was observed for S . carpocapsae , S . feltiae exhibits high levels of activation ( combined partial and full activation ) after only 6 hours of exposure to host tissues ( Fig 1G ) . However , S . feltiae IJs exhibited a higher percentage of fully activated morphology ( approx . 25% ) compared to S . carpocapsae ( approx . 15% ) at 6 hours . And while both species displayed time-dependent increase in activation rates , S . feltiae activation rates were often higher than S . carpocapsae with significantly higher full activation rates after 6 , 24 , and 30 hours of exposure ( Fig 1G , S1 Table ) . After determining the activation dynamics of S . feltiae IJs , we collected the ESPs of activated S . feltiae IJs to determine their effect in insects . S . feltiae IJs were activated in insect homogenate for 0 , 6 , 12 , 18 , 24 , or 30 hours , washed to remove the insect homogenate , and incubated in PBS for 3 hours where they continued releasing ESPs . The PBS ( with accumulated ESPs ) was then filtered through a 0 . 22 μm filter to remove the IJs and concentrated for further experiments . The relative age of all the ESPs were the same; at most , they were 3 hours old . We found that the profile of S . feltiae ESPs changed over time with proteins between 25 and 37 kDa being consistently present from 6–30 hours while proteins between 37–75 kDa peaked at 12 hours and diminished in abundance thereafter ( Fig 2A ) . There was an overall time-dependent decrease in proteins released by S . feltiae ( S1 Fig ) . Comparing the protein band profiles of S . feltiae and S . carpocapsae ESPs side-by-side shows that the majority of S . feltiae ESPs are between 25 and 75 kDa while S . carpocapsae ESPs are more concentrated in a narrower size range , between 25 and 50 kDa ( Fig 2B ) . Naïve S . feltiae IJs produced a relatively large amount of ESPs , with most of these proteins below 37 kDa ( Fig 2A and 2B ) while naïve S . carpocapsae IJs produced undetectable levels of ESPs ( Fig 2B ) . Next , we tested the activity of S . feltiae ESPs in insect hosts . We injected 20 ng of S . feltiae ESPs into Drosophila melanogaster adults and monitored their survival . We found that the ESPs from naïve ( 0 hour ) IJs were not toxic ( Fig 2C ) . ESPs collected from the early activation time points ( 6 and 12 hours ) exhibited the highest toxicity while ESPs from later activation time points ( 18 , 24 , and 30 hours ) decreased in toxicity ( Fig 2C ) . This activation-dependent toxicity is in stark contrast with S . carpocapsae ESPs , which maintained consistently high toxicity levels , even for ESPs collected after 30 hours of activation ( Fig 2D ) . Late stage L4 and early adults were present at the later time points ( 24 and 30 hours ) and since the more developed nematodes are fragile it was possible that some of these nematodes were damaged and unable to continue producing ESPs or were producing different ESPs . To address this possibility , we quantified the number of damaged nematodes throughout activation using a vital stain ( 0 . 2% trypan blue ) . Since it was the later time points ( 18 , 24 , and 30 hours ) that exhibited notable decreases in ESP amount and toxicity we compared the number of damaged nematodes in these groups to that found among the 6-hour activated nematodes . The number of damaged nematodes did increase at the later time points ( as expected ) but the only group that exhibited a significantly higher percentage of damaged nematodes was the 30-hour time point , which accounted for less than 5% of the population ( S2 Fig and S2 Table ) . Further , to simulate harsh experimental handling of the nematodes we repeated the activation time course but applied manual crushing/pressing of the activation sponge before washing the nematodes out for staining and observation . We found that manual crushing/pressing of the sponge caused significant increases in the percentages of damaged nematodes , with the highest average just below 12% at the 30-hour time point ( S2 Fig and S2 Table ) . We also evaluated whether the toxicity we observed was primarily from nematode-derived ESPs or contamination from its symbiotic bacteria , Xenorhabdus bovenii . We compared ESPs from axenic S . feltiae IJs activated for 6 hours and found that the profile of ESPs and the toxicity ( S3 Fig ) were similar to those of symbiotic IJs ( Fig 2A–2C ) , leading us to conclude that the toxicity in these experiments is a result of nematode-derived ESPs . We performed single-nematode RNA-seq analysis [6] in order to identify the similarities and differences between the activation of S . feltiae in vivo and in vitro . We collected RNA from 3 individual nematodes activated in vitro for 3 , 6 , and 9 hours and from nematodes dissected out of infected waxworms ( in vivo ) at 3 , 6 , 9 , 12 , and 15 hours . We performed differential expression ( DE ) analysis using edgeR [32] and found 5670 genes to be differentially expressed between 6 hours in vitro activated IJs and naïve IJs ( Fig 3A ) . Among these genes , 3 general gene expression patterns were observed: Increasing expression over time , increasing first and then decreasing over time , and high levels of expression in naïve IJs with expression decreasing over time ( Fig 3A ) . With the 5670 differentially expressed genes between 6-hour in vitro activated IJs and naïve IJs , we then used MaSigPro to identify genes with significant expression differences and similarities between in vitro and in vivo time courses and identified 3 major clusters ( Fig 3B ) , similar to the result from edgeR analysis ( Fig 3A ) [33] . Cluster 1 consists of 366 genes that demonstrate a distinct profile between in vitro ( red ) and in vivo ( green ) conditions ( Fig 3B ) . While the 6-hour in vitro and 6-hour in vivo samples had similar gene expression levels , many of these genes showed increasing expression up to 15 hours in vivo , whereas they showed decreasing expression by 9 hours in vitro . GO terms for defense response ( p-value 3 . 24e-7 ) , proteolysis ( p-value 1 . 92e-5 ) as well as enzymatic activities such as peptidase ( p-value 3 . 43e-11 ) and hydrolase ( p-value 3 . 36e-15 ) are enriched in cluster 1 ( Fig 3C ) . Enzymatic activity is also a feature of cluster 2 ( 815 genes ) with enzymes such as oxidoreductase ( p-value 1 . 73e-22 ) and serine-type peptidase ( p-value 4 . 44e-4 ) reaching a peak of expression at 6 hours in vitro and in vivo . Lastly , cluster 3 consists of 2578 genes that decrease within 3 hours of activation . GO analysis of cluster 3 genes found enrichments in terms involved with response to hydrolase activity ( p-value 4 . 56e-5 ) , response to chemical ( p-value 3 . 98e-4 ) and enzyme activity such as phosphoric ester hydrolase activity ( p-value 6 . 90e-6 ) and peptidase regulator activity ( p-value 5 . 97e-4 ) ( Fig 3C ) . An analysis of changes in gene expression over the time course ( 3 , 6 , 9 , 12 , and 15 hours post infection ) of in vivo activation also identified 3 major patterns of expression or clusters ( S4 Fig ) . Cluster 1 has 286 genes and GO terms for defense response ( p-value 1 . 44e-5 ) and enzymatic activity such as hydrolase ( p-value 4 . 01e-9 ) and peptidase ( p-value 6 . 49e-9 ) ( S4 Fig ) . Cluster 3 consists of 1 , 153 genes and GO analysis found enrichments in terms involved in enzymatic regulation such as negative regulation of catalytic activity ( p-value 4 . 71e-4 ) , regulation of serine kinase activity ( p-value 2 . 21e-4 ) and regulation of protein phosphorylation ( p-value 2 . 18e-5 ) . Cluster 2 has 1 , 353 genes which have a high expression in IJs and a sharp decrease in gene expression by 3 hours with a minor peak at 6 hours ( S4 Fig ) . GO analysis reveals enzymatic activity is also a feature of cluster 2 with enzymes such as kinase ( p-value 7 . 89e-5 ) and phosphoprotein phosphatase ( p-value 5 . 8e-4 ) . Interestingly , the GO analysis is also enriched for neuropeptide signaling pathway ( p-value 4 . 18e-10 ) ( S4 Fig , Cluster 2 ) . We investigated further into the neuropeptide pathway genes and found that L889_g32029 ( Sf-flp-21 ) , which is orthologous to C . elegans flp-21 and is a neuropeptide important for host-seeking behavior [34] , decreases 8-fold in expression ( S4 Fig ) . Similarly , L889_g7374 ( Sf-flp-11 ) , which is an ortholog of C . elegans flp-11 , demonstrates strong expression at the IJ stage but has the sharpest decrease by 15 hours ( S4 Fig ) . Other neuropeptides such as L889_g30047 ( orthologous to C . elegans flp-3 ) , L889_g15885 ( orthologous to C . elegans flp-18 ) , L889_g27993 ( orthologous to C . elegans flp-14 ) and L889_g32992 ( orthologous to C . elegans flp-7 ) are highly expressed at the IJ stage and progressively decrease by 15 hours post infection ( S4 Fig ) . Overall , both in vivo and in vitro time courses showed significant downregulation of a set of naïve IJ genes within 3 hours as well as equivalent activation of another set of genes by 6 hours and differentially express similar sets of genes associated with proteolytic enzymes ( peptidases ) . The in vivo-only analysis is similar to the in vivo and in vitro DE analyses for both clusters 1 and 3 but have a different profile for cluster 2 . In cluster 2 of the in vivo-only time course there is a decrease in the expression of neuropeptides ( including ones thought to function in host-seeking behavior ) at the later time points , which is likely correlated with reduction of host-seeking sensory functions after successful infection of a host . Because of the high toxicity of the ESPs collected at the 6-hour time point and the similarity in gene expression between 6-hour in vitro and in vivo activated IJs , we chose to primarily focus on the 6-hour ESPs along with further analysis of ESPs from naïve IJs . Using mass spectrometry , we identified 266 proteins ( False Discovery Rate , FDR < 5% , S3 Table ) . To determine the level of correlation between gene expression and relative protein abundance , an mRNA abundance ( TPM , transcripts per million ) to protein abundance ( emPAI , exponentially modified protein abundance index ) correlation analysis of the 266 proteins was performed . We found a weak positive correlation between mRNA and protein abundance with Pearson’s correlation value of 0 . 452 and Spearman’s rank value of 0 . 438 ( S5 Fig ) . We then analyzed the protein sequences for protein domains using Pfam , an online database of protein families [35] . Fig 4A lists the 12 most abundant Pfam domains in S . feltiae ESPs with peptidase domains being the highest in abundance followed by glycosyl hydrolases , lectins , Ig-related ( Immunoglobulin like ) , and peptidase inhibitors . VW ( Von Willebrand ) domains and FAR domains were also found in relatively higher abundance ( Fig 4A ) . A Merops ( peptidase and peptidase inhibitor database ) analysis detected 92 peptidases and 17 peptidase inhibitors with metallo and serine peptidase being the highest in abundance ( Fig 4C ) . In analyzing the ESPs of naïve IJs we identified 682 proteins ( FDR < 5% , S3 Table Sheet 2 ) . Peptidase domains were also the highest in abundance in the ESPs from naïve IJs , followed closely by ribosomal , Ca-related ( calcium interacting/regulating proteins ) and ATPases ( Fig 4B ) . A Merops analysis detected 79 peptidases and 28 inhibitors with both metallo and serine peptidases in high abundance; with the number of metallo peptidases more than double of serine peptidases ( Fig 4D ) . We confirmed that the mRNA of the 266 S . feltiae ESPs were detected at the 6-hour in vitro time point , and that these are expressed similarly at 6 , 9 , 12 , and 15 hours in vivo ( Fig 5A ) . We compared the gene expression of these 266 proteins between 6 hours in vitro and naïve IJs and found that 54 genes are downregulated and 96 genes are upregulated upon activation ( Fig 5B ) . Gene ontology terms ( GO ) for the 96 upregulated genes show strong enrichment for enzymes such as hydrolases ( p-value 2 . 63e-25 ) and peptidases ( p-value 4 . 96e-18 ) and endopeptidase ( p-value 3 . 71e-8 ) , indicating that the activated nematodes increase the synthesis and release of enzymes to degrade host components , including proteins , at early stages of infection . In contrast , the 54 downregulated genes are related to muscle cell development ( p-value 3 . 09e-5 ) , protein complex assembly ( p-value 4 . 13e-5 ) and morphogenesis ( p-value 2 . 96e-5 ) . These data suggest that at 6 hours in vitro the nematodes are at peak production of venom proteins . We then conducted a comparative gene expression analysis of ESPs from S . feltiae and S . carpocapsae to understand the similarities and differences of genes involved in killing hosts . Our orthology analysis between 266 ESPs in S . feltiae and 472 S . carpocapsae found 52 genes in common ( Fig 5C , S4 Table ) . This is a lower number than expected , given that 112 of the 266 S . feltiae ESPs have homologs in S . carpocapsae ( S5 Table ) and 183 of 472 ESPs found in S . carpocapsae have homologs in S . feltiae ( Fig 5D , S6 Table ) . However , most of these homologs are not detected in the ESPs of the other species even when they are expressed ( Fig 5E ) suggesting that these enzymes might have been coopted over time to become part of the venom of either species . Interestingly , both S . feltiae and S . carpocapsae have a high expression of the shared 52 genes . GO terms analysis of the 52 genes shows enrichment of peptidases ( p-value 1 . 25e-7 ) , hydrolases ( p-value 6 . 71e-10 ) and alpha-glucosidase activity ( p-value 2 . 36e-5 ) ( S7 Table ) . These results correlate with Pfam domains found in common between S . feltiae and S . carpocapsae ( Fig 4A and [5] ) . We conclude that this small set of proteins form part of a core of venom proteins within Steinernema . Next , we wanted to determine whether these 52 ESPs from insect-parasitic nematodes were conserved in nematode parasites of vertebrates . We ran blastp on the 52 proteins ( E-value < 1e-3 ) and compiled the best non-Steinernema hits for each protein . More than half ( 31 out of 52 ) of these genes have orthologs in mammalian-parasitic nematodes ( S6 Fig ) that include Strongyloides ratti , Toxocara canis , and Ancylostoma duodenale ( S4 Table ) . The prevalence of these proteins in both insect- and vertebrate-parasitic nematode species leads us to speculate that these proteins may play critical roles during host infection and survival within the host for parasites in general .
Many nematodes have an alternate L3 stage of development , known as the dauer juvenile in free-living and necromenic species , or the infective juvenile for parasitic species [10 , 36 , 37] . The transition that parasitic IJs make when they enter a host and become actively parasitic and resume development is known as dauer recovery or activation . For parasitic nematodes , successful activation is critical to establishing a successful infection in their hosts [5 , 11 , 38 , 39] . Similar to other EPNs , S . feltiae activation rates increased in a time dependent manner after exposure to insect tissue in vitro [5 , 17] . After 30 hours of exposure to host tissue , essentially all the nematodes displayed some level of activation with non-activation rates being an average of 0 . 3% ( S1 Table ) . Although S . feltiae and S . carpocapsae are in the same genus , they are members of different clades within the genus [27 , 40 , 41] . The fact that these EPNs display similar behavior and morphology during activation when exposed to insect tissue demonstrates that the in vitro model of activation we used is a consistent and robust model of activation . We found that when activated in vitro , the S . feltiae population does not exhibit synchronous activation . Some individuals are fully activated , some are partially activated , and a small number are not activated at all . We found this resolution of activation quantification to be reliable and consistent however we do note that these 3 categories of activation are broad; encompassing different degrees of pharyngeal bulb expansion , and that the resolution could have been increased by including other factors such as active pumping of the pharyngeal bulb or expansion of the anterior gut . Along with this phased activation , the full activation rates seem to taper off when the nematodes are activated for a long time ( Fig 1G ) . Similar observations have been made for S . carpocapsae and S . scapterisci activation [5 , 17] . The phenomenon of non-synchronous activation is similar to the phased infectivity reported in in vivo infections , wherein a certain percentage of an IJ population is unable to infect insect hosts or displays reduced infectivity , but over time more individuals become infectious [42 , 43] . This characteristic is believed to be inherent to the IJ itself and does not seem to be significantly affected by factors such as IJ population or host population density . Studies have shown that phased infectivity correlates well with Heterorhabditis EPNs but not as well with Steinernema EPNs [44 , 45] . In contrast to H . bacteriophora , where the infectious percentage of the population seems to start out low , previous research suggests that a large percentage of a Steinernema IJ population is typically infectious [44] . It has been suggested that the phased infectivity hypothesis is incomplete , and many other factors , such as genetic/physical damage , attraction to infected vs non-infected hosts , and survival of the IJ within the host , could affect population infectivity [46] . The age of the IJs could also be a contributing factor and was previously shown to affect activation rates in steinernematids [16] . In our in vitro model , the IJs do not actually infect a host , but rather are exposed to host tissue as if they had already infected the host . In this context , all the IJs are exposed to host tissue at the same time and though the majority of the population activate to some degree some individuals seem to respond faster and become fully activated early on while another portion of the population activates slower . We did not test whether population density was a factor , nor did we strictly control for age ( IJs were between 2 weeks and 2 months post collection ) but our findings are consistent with previous studies of phased infectivity . Thus in vitro activation may be a useful tool in further exploring the potential relationship between infectivity and activation . It is widely recognized that helminths modulate host immune system and cause pathology mainly through the release of proteins and small molecules that interact with host cells and tissues , and that these molecules are key factors in disease pathology and parasite fitness [47 , 48] . However , nearly all previous and current helminth secretome and ESP studies have been done in vitro , due to the difficulty of detecting ESPs from helminth parasites in their hosts . Additionally , there has been little if any experimental validation that the in vitro induction of ESPs from various parasitic helminths accurately mimics in vivo conditions . Here , we utilized single-nematode RNA-seq to compare the transcriptomes of nematodes dissected out of waxworms after infection for 3 , 6 , 9 , 12 , and 15 hours and those of nematodes activated in vitro for 3 , 6 , and 9 hours . We found that the transcriptional profiles of nematodes activated in vitro were generally similar to those of nematodes from in vivo infections at each time point ( Fig 3A ) however some time points were more similar than others . We identified three major clusters of genes among the 5670 differentially expressed genes between activated and naïve IJs and within these three clusters the transcriptome profiles of the 6 h in vitro and 6 h in vivo activated nematodes exhibited the most consistent correlation ( Fig 3B ) . In contrast , the gene expression profiles of nematodes activated in vitro and in vivo at 3h and 9h had significantly different profiles and did not correlate consistently ( Fig 3B ) . Therefore , 3h and 9h in vitro are not representative of their in vivo counterparts . These data suggest that ( 1 ) activation of IJs in vitro can mimic in vivo infection and yield physiologically relevant results; ( 2 ) the fidelity of the in vitro results needs to be experimentally validated rather than simply assumed; and ( 3 ) selection of the timing of ESP collection should be based on the experimental evidence of when the in vitro system best mimics the natural process . It is important to determine the similarity of expression profiles for other parasites such as mammalian-parasitic nematodes freshly dissected from hosts compared to those stimulated under in vitro ESP collection conditions [49–51] . RNA-seq of individual nematodes , as we have done in this study , can be used to determine the similarity in the nematodes’ response to in vitro and in vivo conditions in order to optimize experimental in vitro conditions . This method is especially beneficial in parasitic studies where low parasite yield is a limiting factor . In addition , gene expression similarity should be optimized when using non-natural hosts , which are often used due to the difficulty of obtaining or maintaining natural hosts or lack of tools and techniques in non-model hosts compared to a model hosts such as a mouse . In EPN research , the nematode has been traditionally assumed to act primarily as a vector for the pathogenic bacterial symbiont . Once the bacterial pathogen is inside the host , it will kill the host while multiplying and providing nourishment ( the bacteria itself and the insect tissue ) for the nematode [10 , 20 , 52] . However , there is a growing body of research establishing the nematode as an active contributor to pathogenesis , and in some cases such as with S . scapterisci , the nematode may be the main driver of virulence [53] . It is clear that aside from serving as a vector for the bacteria they carry , EPNs contribute to pathogenesis in two ways: They directly damage host tissue and they dampen host immunity , acquiring more time for themselves and the bacteria they carry to overcome and kill the host . Past studies have shown that axenic S . carpocapsae IJs can kill and reproduce in insect hosts [54–56] and individual effector molecules from steinernematids have been characterized and shown to function in host immune suppression and tissue damage [18 , 19 , 57–61] . More recently the secretome of S . carpocapsae was shown to be a complex mixture containing many proteins and that collectively , this venom is toxic to insects . ESPs collected from axenic S . carpocapsae IJs had similar protein profiles as those from IJs associated with their bacterial symbiont , and the ESPs from both populations were similarly toxic [5] . We have shown these findings to also be true for S . feltiae , where S . feltiae IJs exposed to insect tissue become activated and produce ESPs ( Fig 2A ) that are toxic to insects ( Fig 2C ) . ESPs collected from axenic S . feltiae IJs also displayed similar protein profiles and toxicity ( Fig 2A and 2C; S3 Fig ) compared to their symbiotic counterparts . For EPNs in the genus Steinernema , the nematodes seem to play a much more active role in contributing to pathogenicity during infection than previously thought . We found that there are notable differences in ESP production and content among steinernematids . Whereas the protein profiles of S . carpocapsae ESPs were previously shown to be fairly constant after 6 to 30 hours of exposure to insect tissue [5] we found that the protein profiles and protein amount of S . feltiae ESPs change from 6 hours to 30 hours of exposure to host tissue ( Fig 2A; S1 Fig ) . Comparing the profiles of ESPs from S . feltiae and S . carpocapsae side by side ( Fig 2B ) , both have bands that are similar in size however the majority of intense S . carpocapsae bands are concentrated between 25–50 kDa while the majority of intense S . feltiae bands are not as concentrated and distinctly more spread out between 25–75 kDa . We found that there is a core suite of proteins found in the ESPs of both species and the differences in the protein profiles could be a result of adaptation to different bacterial symbionts or perhaps a result of host specialization . Another striking difference in ESP production between the two species is that when measuring ESPs from naïve IJs , S . carpocapsae was shown to produce few if any ESPs ( not detectable by Bradford assay nor any notable bands by silver-staining ( Fig 2B ) ) while naïve S . feltiae IJs produce a relatively large quantity of ESPs ( Fig 2B ) . ESPs from naïve S . feltiae IJs shared some similarities with those from 6-hour activated IJs; namely that they were produced in relatively large quantities and included peptidases , peptidase inhibitors , and glycosyl hydrolases ( Fig 4B ) . However , the protein profiles are different from each other ( Fig 2A ) where ESPs from naïve IJs contain a more diverse array of proteins ( S3 Table sheet 2 ) and there were generally more peptides detected for each protein domain ( Fig 4B ) . Further , the ESPs of naïve IJs were not toxic unlike their activated counterparts ( Fig 2C ) . The release of ESPs from naïve S . feltiae IJs without any stimulation from host cues seems metabolically wasteful . We evaluated the possibility that the ESPs from naïve IJs we collected were a result of damage from experimental handling rather than active release by the nematodes . We concluded that the contribution of ESPs from damaged nematodes is likely minimal for the following reasons: ( 1 ) S . feltiae IJs were treated exactly as S . carpocapsae IJs in a previous report [5] , yet naïve S . carpocapsae IJs did not release detectable amounts of protein . ( 2 ) The nematodes in these experiments , if exposed to host tissue , began producing ESPs with a considerably different composition than naïve IJs ( Figs 2B , 4A and 4B ) . ( 3 ) If allowed , the nematodes continued to develop into healthy , reproductive adults . Instead , our data reveals that naïve S . feltiae IJs are capable of producing a different set of ESPs , which could be involved with survival strategies including stress tolerance , lubrication and avoidance of desiccation , or maintaining the cuticle and other bodily structures . These strategies may be more pertinent to S . feltiae as it is categorized as more of a cruiser where it actively migrates in the soil seeking new hosts , while ambushers like S . carpocapsae tend to wait in epigeal habitats [13 , 15] . Another possibility for the role of naïve S . feltiae IJ ESPs is preparation of the IJ cuticle for host infection since the cuticle of S . feltiae IJs has suppressive effects against host immune responses [28 , 30 , 31] . Peptidases , peptidase inhibitors , and immunoglobin-like proteins are detected in high abundance in the ESPs and they can be produced to potentially coat/adhere to the cuticle . The production of ESPs from naïve S . feltiae IJs is an interesting find that differentiates S . feltiae from S . carpocapsae and merits further study to understand the biology of this parasite . The toxicity of activated S . feltiae ESPs was highest at the earliest time points tested ( 6 and 12 hours of exposure ) and toxic activity decreased in a time-dependent manner with those collected after 24 and 30 hours of exposure being significantly less toxic ( Fig 2C ) . The change in protein profiles ( Fig 2A ) and the reduced protein levels ( S1 Fig ) in S . feltiae ESPs over time seem to be correlated with the time-dependent toxicity decrease . However , it is unlikely that the reduction of toxicity is due to the decreasing abundance of total ESPs since the flies were exposed to the same amount of ESPs ( 20 ng per fly ) ; instead , it is more likely that some low abundance toxin ( s ) in the mixture decrease ( s ) over time , resulting in lower toxic activity . The correlation between protein profiles/abundance and toxicity was not observed for S . carpocapsae ESPs: Later time points ( 42 and 54 hours of exposure ) had similar protein profiles and protein abundance compared to earlier time points ( 6–30 hours of exposure ) , but were significantly reduced in toxicity or were not toxic at all [5] . This suggests that the toxic activity is due to low abundance proteins . Therefore , the toxins of both species are likely low abundance proteins and not the most abundant ones ( Fig 2A and 2B ) . Other proteins found in the ESPs likely have non-toxic functions during infection such as immunosuppression or immune evasion . We considered the possibility that damaged nematodes could be an explanation for the time-dependent decrease in ESP amount or toxicity and upon evaluation found a time correlated increase in the number of damaged nematodes . However , the highest level of damage we observed was less than 5% of the total population ( S2 Fig ) . Even manually crushing the activation arena to simulate excessive force averaged less than 12% of the nematodes being damaged . We believe that the percentage of damaged nematodes from our experimental handling alone is insufficient to explain the dramatic changes we see in S . feltiae IJ ESP production and activity . It could also be argued that instead of ( or in conjunction with ) the nematodes being significantly damaged , they become unhealthy at the later time points due to various factors such as depletion of resources . We acknowledge this possibility however , it is unlikely the limiting factor as this was not observed in S . carpocapsae [5] . Instead , the time-dependent decrease in toxicity and amount of S . feltiae ESPs compared to the much slower decrease in toxicity and amount of S . carpocapsae ESPs suggests that these nematodes utilize different strategies in establishing themselves as parasites . S . feltiae may have a stronger reliance on its bacterial symbiont , X . bovienii , in order to overcome and kill the host . Soon after activation and release of bacterial symbionts , the IJs may switch their priority from killing the host to survival , feeding , and development . Axenic S . feltiae IJs have been shown to be capable of killing insect hosts , however the studies are limited compared to studies of S . carpocapsae and they generally report reduced efficiencies [62 , 63] . We found no difference in activity between ESPs from axenic compared with symbiotic S . feltiae IJs , however we tested the activity of the ESPs alone and did not examine the larger context of an actual insect infection . It is possible that differences in ESP profiles between S . carpocapsae and S . feltiae are involved in niche partitioning and differences in host range and specificity . We found 266 proteins in S . feltiae ESPs which is significantly fewer than the 472 proteins that were detected in S . carpocapsae ESPs [5] . However , this difference may be due to the more fragmented nature of the available S . feltiae genome , which has an N50 of 47 . 5kb compared to the 300kb N50 of the S . carpocapsae genome [40] that was used in the previous study ( N50 is the length of the shortest contig that together with all the longer contigs cover 50% of the genome assembly ) . Although it is likely that the ESPs from EPNs are complex mixtures containing many different classes of molecules , we focused on analyzing the proteins . The most abundant group of proteins in activated S . feltiae venom are peptidases with a high proportion of serine and metallopeptidases ( Fig 4A and 4C ) . This is similar to what was previously reported in S . carpocapsae ESPs [5] . However , S . carpocapsae ESPs contained fewer metallopeptidases and significantly more serine peptidases . The high abundance of peptidases and peptidase inhibitors in the ESPs of both species illustrate the importance of these enzymes for EPNs as well as other parasites . Many studies have implicated their potential use in vaccine development and treatment [64–67] . Peptidases and peptidase inhibitors have been shown to have multiple functions in parasite pathogenesis including suppressing/evading host immune systems , host tissue damage , and parasite development [68] . Serine peptidases in particular have been suggested to be used by many parasites including Trichinella spiralis , Ascaris suum , and Brugia malayi , among others [69–71] . Some specific characterizations of nematode serine peptidase functions include collagen degradation , suppression of melanization , inhibition of blood clotting , and parasite sperm activation [72–74] . We analyzed the protein domains in the ESPs to determine the potential molecular functions of the proteins . For S . feltiae , the second most abundant protein domain after peptidases were domains associated with hydrolysis of glycosydic bonds . These enzymes are hypothesized to have many potential functions , including cleavage of glycosolated proteins and breakdown of structural components that contain glycosidic bonds , with many similarities to peptidases . Some of the other protein domains detected in higher abundance in both S . feltiae and S . carpocapsae ESPs are Ig ( immunoglobulin ) or Ig-like , Von Willebrand , and FAR ( fatty acid/retinol binding protein ) . The fact that both EPN species have high representation of these domains in their ESPs suggests their importance for EPN success . It is likely that some of these proteins are involved in immunomodulation . For example , it has been hypothesized that FAR proteins affect immune signaling [75] , and while this has been experimentally demonstrated in plants [76–78] , it has yet to be shown in an animal system . S . feltiae has been shown to modulate insect immunity using its cuticle but the use of specific excreted/secreted proteins in immune modulation by S . feltiae would be a novel finding [28 , 31] . Additionally , we evaluated the correlation between mRNA abundance and protein abundance for these ESPs . The correlation was weak but positive with a Pearson’s correlation of 0 . 452 and Spearman’s rank correlation of 0 . 438 ( S5 Fig ) . mRNA-protein abundance correlations have consistently been weak in various studies including those involving nematodes [79 , 80] and our data support this trend . The discrepancies between mRNA and protein abundance is likely due to post-transcriptional regulating systems that can include small non-coding RNAs and microRNAs which has been postulated before [79] . It has been pointed out that most studies of mRNA-protein abundance correlation have been focused on transcriptome-wide data and a study specifically focused on upregulated transcripts resulted in a higher distribution of strong correlations , but we did not evaluate this in the present study [81] . In examining the 266 ESPs released by S . feltiae and the 472 ESPs released by S . carpocapsae , we found 52 proteins conserved in the ESPs of both species . This was unexpectedly low since 112 of the 266 S . feltiae ESPs have homologs in S . carpocapsae and 184 of the 472 S . carpocapsae ESPs have homologs in S . feltiae ( Fig 5D ) . Both S . feltiae and S . carpocapsae have a high expression of the shared 52 venom genes , representing a core of effector proteins shared by these EPNs . Within this core set of ESPs there are peptidases , glycosyl hydrolases , lectins as well as proteins likely to be involved in immune modulation such as FAR proteins , immunoglobulins , and immunoglobulin-like proteins . The specific functions of these core venom proteins are yet unknown , but their conservation between S . carpocapsae and S . feltiae , which are in different clades within the genus , suggests that they are important effectors of parasitism and function in a variety of insect hosts . The genus Steinernema is the oldest known lineage of EPNs , potentially coevolving with their insect hosts for ~350 million years [26] . Determining the functions of the proteins in this core suite of ESPs may elucidate important steps in the evolution of EPNs and even more broadly parasitic nematodes in general .
Galleria mellonella ( waxworms ) were purchased from CritterGrub ( www . crittergrub . com ) . Oregon-R Drosophila melanogaster flies were reared in round bottom 8 oz bottles with food medium ( 129 . 4 g/L dextrose , 7 . 4 g/L agar , 61 . 2 g/L corn meal , 32 . 4 g/L yeast , and 2 . 7 g/L tegosept ) . The bottles were kept at 25°C with 60% relative humidity on a 12 hr light/dark cycle . S . feltiae IJs were cultured and propagated in vivo using waxworms as previously described [5] . Briefly , 15 wax worms were placed into a 10 cm petri dish with filter paper pressed to the bottom and 1 ml of tap water containing 750 S . feltiae IJs ( 50 IJs/worm ) was dispersed onto the filter paper . The infection plates were incubated at 25°C with 60% humidity in the dark for 10 days . Then , the waxworm cadavers were transferred to White traps [82] . After 2–3 days ( depending on IJ density ) the IJs were collected and washed using a glass vacuum filter holder ( Fisher Scientific , 09-753-1C ) with an 11 μm nylon mesh filter ( Millipore , NY1104700 ) . The IJs were stored at 15°C at a density of 7–10 IJs/μl . Insect homogenate was prepared as previously described [5] . Briefly , 25 g of waxworms were frozen and grounded in liquid nitrogen with a mortar and pestle into a fine powder . The waxworm powder was then transferred quickly into a glass beaker and resuspended in 100 mL of Phosphate Buffered Saline ( PBS , 137 mM NaCl , 2 . 7 mM KCl , 10 mM Na2HPO4 , 1 . 8 mM KH2PO4 , pH 7 . 4 ) . The mixture was then microwaved to a boil 7–8 times with stirring in between . The homogenate was then aliquoted into 50 mL conical tubes and centrifuged for 5 minutes at 3200 rcf to pellet the solid debris of the waxworm . The supernatant , including the top oil layer were transferred into a new container . PBS was then added to the 50 mL conical tubes containing the waxworm pellets , mixed , centrifuged , and the supernatant was collected . This was repeated until the desired volume and percent extract was reached . In this case , 25g of waxworm was used to make 100 ml of 25% waxworm homogenate . The waxworm homogenate extract was used immediately or aliquoted and stored at -20°C . IJ activation was done as previously described [5] . 100 mL of 25% waxworm homogenate was thawed and supplemented with 1x triple antibiotic Pen/Strep/Neo ( P4083 , Sigma-Aldrich ) . The homogenate was soaked into 8 . 2 g of autoclaved cut sponge pieces ( approximately 3x3x10 mm ) . 2 . 5 million S . feltiae IJs were washed 4 times with autoclaved 0 . 8% NaCl solution and excess liquid was removed from the washed IJs before gentle Pasteur pipette transferring/mixing into the homogenate-soaked sponge . The container was covered with aluminum foil and incubated in the dark at 25°C with 60% relative humidity for a specified amount of time . For most of the contents of this study , the IJs were incubated in waxworm homogenate for 6 hours . The IJs were then washed out of the sponge with 6–8 rounds of autoclaved 0 . 8% NaCl solution and once separated from the sponge , further washed with 4–5 rounds of 0 . 8% NaCl solution . Activations were replicated at least 3 times for each experiment . IJ activation quantification was done as described [5 , 16 , 17] . Briefly , activated IJs were observed under 400x magnification on a compound light microscope and scored for the activation phenotype based on expansions of the pharyngeal bulb . Fully activated phenotypes ( see Fig 1E ) , partially activated phenotypes ( Fig 1 image C ) , and Non-activated phenotypes ( see Fig 1A ) were scored . The difference between non-activated IJs from those that have been partially or fully activated is easily visualized as the absence of a visible pharyngeal bulb at 400x magnification . Differentiating between partially and fully activated IJs relies on the relative size and shape of the pharyngeal bulb; fully activated IJs have a wider , round-shaped bulb whereas partially activated IJs have a narrower , oval-shaped bulb . To minimize bias and double scoring the same nematode , scoring started with viewing IJs at one corner of the coverslip . All IJs with anterior/head region in view were scored before shifting the slide to view the next adjacent region . This was repeated until all regions of the coverslip was viewed without viewing the same region twice . Activations were done in 3 replicates for each time point ( naïve/0 hr , 6 hr , 12 hr , 18 hr , 24 hr , and 30 hr ) and each replicate was scored 3 times to obtain averages . Significant differences between S . feltiae and S . carpocapsae IJ activations were determined using the Prism 8 by paired two-way ANOVA with ( Prism recommended ) Sidak’s multiple comparisons between related groups ( i . e . rates of partially activated S . feltiae IJs at 30 hrs of exposure compared to rates of partially activated S . carpocapsae IJs at 30 hrs of exposure ) . ESP collection from the EPN was done as previously described [5] . After IJs were activated and thoroughly washed , they were transferred into a 1 L Erlenmeyer flask containing 100 mL of autoclaved PBS supplemented with 1x triple antibiotic Pen/Strep/Neo . The flask was shaken at 220 rpm in the dark for 3 hours and the nematodes were then centrifuged ( 700–800 rcf for 1 minute ) in 15 mL conical tubes to preliminarily separate the majority of the nematodes from the PBS . The PBS supernatant was then collected and filtered through a 0 . 22 um syringe filter ( Fisher Scientific , 9719001 ) and concentrated to approximately 300 μL using a 3 kD cut-off centrifuge column ( Millipore , UFC900308 ) . The protein concentration of the venom was quantified using a Bradford assay ( Bio-Rad , 500–0006 ) . S . feltiae ESPs were prepared for gel electrophoresis by boiling for 5–10 minutes in 1x Laemmli sample buffer supplemented with 50mM Dithiothreitol ( DTT ) ( Bio-Rad , 1610747 ) . The denatured proteins were loaded into a Mini-PROTEAN TGX precast gels ( Bio-Rad , 4561086 ) and electrophoresed at 100 V for 60–90 minutes . Silver staining was done following the manufacturer’s protocol ( Pierce , # 24600 ) . S . feltiae ESPs toxicity was tested in vivo on Drosophila melanogaster flies as previously described [5 , 83] . Adult male flies 5–6 days old were anesthetized with CO2 and injected with 20 ng of ESPs in a volume of 50 nl using pulled glass needles and a highspeed pneumatic microinjector ( Tritech Research , MINJ-FLY ) . PBS was injected as a negative control . After injection the flies were transferred to new vials containing food and stored at 25°C with 60% relative humidity on a 12hr light/dark cycle . Survival of the flies was recorded over 40 days or until all the flies had died . ESP collection and toxicity testing were done in 3 biological replicates for each time point ( PBS , 0 hr , 6 hr , 12 hr , 18 hr , 24 hr , 34 hr ) with 3 technical replicates of each biological replicate . At least 60 flies were used for each technical replicate totaling at least 180 flies for each biological replicate . Nematodes were activated in vitro as described in the “Activation of IJs” section of the methods however scaled down to fit a 9 cm petri dish ( 0 . 082 g of sponge , 1 mL of 25% insect homogenate , and approximately 25 , 000 S . feltiae IJs ) . The sponge pieces were each pressed down 5 times before the nematodes were washed out and rinsed with 4 rounds of autoclaved PBS . The nematodes were then stained by mixing an equal volume of nematodes with an equal volume of 0 . 4% trypan blue ( Sigma-Aldrich ) to give a final dye concentration of 0 . 2% . The mixture was allowed to sit for 5 minutes before transferring the nematodes to a microscope slide for viewing and counting . This was replicated 3 times for each time point ( 6 , 12 , 18 , 24 , 30 hours of activation ) with approximately 5000 counts each replicate ( 15 , 000 total counts for each time point ) Representative images are in S2 Fig and raw counts in S2 Table . Axenic nematode production and assaying was done as previously described [5] with some slight modifications . Axenic S . feltiae IJs were produced in vitro by growing bleach sterilized S . feltiae eggs on the colonizing defective mutant bacterial strain of Xenorhabdus nematophila , HGB315 [84] . HGB315 is unable to colonize the nematodes however can still be a food source . Phase I of the HGB315 bacteria colonies ( blue ) were obtained and verified using NBTA agar plates ( 40 mg/L 2 , 3 , 5-triphenyltetrazolium , 25 mg/L bromothymol blue , 8 g/L nutrient agar , supplemented with 0 . 1% ( w/v ) sodium pyruvate ) and double checked with MacConkey Agar plates ( reddish brown ) ( Difco MacConkey Agar , #212123 , supplemented with 0 . 1% ( w/v ) sodium pyruvate ) . HGB315 was cultured in LB broth supplemented with 0 . 1% ( w/v ) sodium pyruvate over night at 28°C and shaking at 220 rpm . 100–150 μl of overnight HGB315 liquid culture was spread on lipid agar plates ( 4 ml/L corn oil , 7 ml/L of corn syrup , 5 g/L of yeast extract , 2 g/L MgCl2 , 8 g/L of nutrient broth , 15 g/L of Bacto Agar , supplemented with 0 . 1% ( w/v ) sodium pyruvate ) and incubated at 28°C overnight to form a thin layer of bacterial lawn . Surface sterilized S . feltiae eggs in a minimal volume of sterile Ringer’s solution ( 172 mM KCl , 68 mM NaCl , 5 mM NaHCO3 , pH 6 . 1 ) was dropped onto the lipid agar plates and allowed to develop into gravid females . This is the first round pass to produce F1 generations of S . feltiae nematodes that were exposed only to the non-colonizing HGB315 . HGB315 is a strain of X . nematophila which is not the native symbiotic bacteria of S . feltiae ( Xenorhabdus bovienii ) , therefore these nematodes develop and become gravid much slower at approximately 5–6 days ( versus ~4 days on X . bovienii ) post seeding . To obtain axenic eggs , gravid females were rinsed in autoclaved 0 . 8% NaCl solution for 3 times followed with rocking in axenizing solution ( 0 . 7% NaOCl ( bleach ) /0 . 5 M NaOH ) for 7 . 5 minutes for 3 times . Brief vortexing was applied 2–3 times in the first two rounds of axenizing to ensure mixing and degradation of adult nematode tissue . After the axenizing treatment , the eggs were rinsed in autoclaved Ringer’s solution for 3 times followed by incubation in a triple antibiotic solution ( Penicillin , Neomycin , Streptomycin ) for 30–45 minutes . The eggs were then rinsed with autoclaved Ringer’s solution for 3 times and centrifuged at 700 rcf for 1 min and the supernatant was removed to create a highly dense egg suspension with minimal liquid volume . Approximately 500 , 000 eggs were gently dispersed onto the lipid agar plates containing the HGB315 bacteria . When the bacteria were depleted , the nematodes were washed off and split into 3–5 new HGB315 bacteria plates . The S . feltiae nematodes were kept on the plates until they reached a high density and IJs can be seen crawling up the sides of the plates . At this point the population was still a mix of different life stages so the nematodes were transferred to White traps to collect axenic IJs . To assay for non-colonization of bacteria inside S . feltiae IJs: approximately 1000 IJs were rinsed 3 times with autoclaved Ringer’s solution , followed by surface sterilization with 4 mM Hyamine 1622 solution ( Sigma , 51126 ) for 30 minutes , and rinsed 3 times with Ringer’s solution . The IJs were then concentrated to a volume of 50 μl and homogenized with a tissue grinder ( Fisher Scientific , 12-141-363 ) . The homogenate was then plated onto LB plates ( supplemented with 0 . 1% ( w/v ) sodium pyruvate ) and incubated at 28°C in the dark . The plates were checked for bacterial growth for 5 days ( S3 Fig ) . This was replicated 3 times for each batch of axenic S . feltiae IJs . To prepare S . feltiae ESPs for mass spectrometry analysis , the proteins were first precipitated with 80% acetone ( -20°C pre-chilled ) at 4:1 acetone to sample volume . The mixture was vortexed for 5 seconds 3 times and stored at -20°C overnight . The mixture was then centrifuged at 15 , 000 rcf for 10 minutes at 4°C to pellet the precipitated proteins . The supernatant was carefully removed , followed by addition of fresh -20°C chilled 80% acetone , and mixing by pipetting . The mixture was then centrifuged at 15 , 000 rcf for another 10 minutes . This process was repeated one more time and after removal of the 2nd 80% acetone wash the protein pellet was allowed to air dry for 5 minutes . The protein pellet was then digested using the Trypsin/Lys-C , Mass Spec Grade kit ( Promega , V5071 ) following the manufacturer’s Two-Step In-Solution Digestion protocol . Briefly , the protein pellet was suspended in 7 M urea/50 mM Tris-HCl ( pH 8 ) , followed by addition of DTT to a final concentration of 5 mM , and incubated at 37°C for 30 minutes . Iodoacetamide was then added to a final concentration of 15 mM , and incubated at room temperature for 30 minutes in the dark . The Trypsin/Lys-C protease mix was added at a ratio of 25:1 ( protein: protease ( w/w ) ) and incubated at 37°C for 4 hours . The mixture was then diluted with 50 mM Tris-HCl ( pH 8 ) to reduce the urea concentration to approximately 0 . 5 M and continued incubation at 37°C overnight . Trifluoroacetic acid ( TFA ) was added to a final concentration of 0 . 5–1% to terminate digestion and the mixture was centrifuged at 15 , 000 rcf for 10 minutes to pellet particulate matter . The supernatant containing digested protein was cleaned using a C18 spin column ( Pierce , 89873 ) following the manufacturer’s protocol . Online 2D-nano LC/MS/MS was used to perform MudPIT mass spec analysis of S . feltiae ESPs . The mass spec apparatus consisted of a 2D nanoAcquity UPLC ( Waters , Milford , MA ) configured with an Orbitrap Fusion MS ( Thermo Scientific , San Jose , CA ) . LC solutions/fractionation and MS parameters were as previously described [5] . The raw mass spec data was processed/analyzed with the Proteome Discoverer 2 . 2 software ( Thermo Scientific , San Jose , CA ) with the Sequest HT search engine running against the S . feltiae protein profile , steinernema_feltiae . PRJNA204661 . WBPS11 . protein . fa ( Parasite . Wormbase . org ) . Duplicate genes were removed and only genes with FDR <5% were considered for further analysis . The raw mass spec data have been uploaded to the ProteomeXchange repository and can be accessed with the following links . 0 hr: ftp://massive . ucsd . edu/MSV000082993 6 hr: ftp://massive . ucsd . edu/MSV000082997 Protein/peptide sequences of S . feltiae ESPs obtained from mass spec and the protein domain families were analyzed using the Pfam database and the hmmscan program ( E-value < 10−5 ) of the HMMER software 3 . 0 as described [85] . Peptidase types based on the catalytic center amino acid ( Serine , Metallo , Aspartic , etc . ) and peptidase inhibitors were identified by BLAST+ against the MEROPS Peptidase database [86] from https://www . ebi . ac . uk/Tools/sss/ncbiblast/ . Only hits with an E-value of <10−5 were further analyzed . S . feltiae single nematode transcriptome sequencing was done as previously described [5 , 6] . In vitro activated IJs were activated as described in the Activation of IJs section of the methods but scaled down to fit in a 6 cm petri dish with 1 ml of insect homogenate , 0 . 08 g of sponge , and 25 , 000 IJs [16 , 17] . The IJs were activated for time points 3 , 6 , and 9 hrs . After activation the IJs were washed out of the sponge with autoclaved 0 . 8% NaCl and transferred to 1 . 5 ml eppendorf tubes . The IJs were cleaned by spinning down and removing/replacing the NaCl supernatant 4 times . We used only IJs that displayed fully activated morphology ( confirmed by microscope ) for each time point . This method , though arguably not highly representative of the entire population , was used in order to consistently select for individuals that were activating the fastest for each time point and minimize variation from nematodes with different levels of activation . The IJs were then transferred to RNase-free water before lysis . Naïve ( 0 hr ) IJs were not exposed to any insect tissue and washed before proceeding to lysis . In vivo activated S . feltiae IJs were activated by infecting live waxworms at 50 IJs/waxworm . After 30 minutes the waxworms were gently rinsed in autoclaved 0 . 8% NaCl to wash off IJs that were on the surface of the waxworms but had not entered the waxworm . The infected waxworms were then stored in the dark at 25°C with 60% relative humidity for 3 , 6 , 9 , 12 , or 15 hrs . After the specified hours , the waxworms were individually placed in 6 cm petri dishes with autoclaved 0 . 8% NaCl and the activated IJs were dissected out . The IJs were washed by transferring them to new 6 cm petri dishes with fresh autoclaved NaCl 5x until being transferred to RNase free water before lysis . Activated IJs for each time point/condition ( 6 hr in vitro , 12 hr in vivo , etc . ) were individually isolated in RNase-free water , cut into 3–4 pieces , and immediately transferred to lysis buffer containing RNAse inhibitor Proteinase K . The sample was placed on ice and observed periodically until the nematode tissue had been digested ( typically 45–60 minutes ) . The sample was then incubated in a thermocycler at 85°C for 3 minutes to deactivate proteinase K . dNTP/ Oligo-dT30VN ( 5′-AAGCAGTGGTATCAACGCAGAGTACT30VN-3′ ) was added to the sample and poly-A RNA was reverse transcribed in a reaction solution of 100 U Superscript II RT ( Thermo Fisher Scientific , 18064014 ) , 10 U RNase inhibitor ( Promega , N2611 ) , 1x Superscript II first-strand buffer , 5 mM DTT , 1 M Betaine , 6 mM MgCl2 , 1 μM TSO ( LNA-modified TSO 5′-AAGCAGTGGTATCAACGCAGAGTACATrGrG+G-3′ , Exiqon . com ) , and RNase-free water . The reverse transcription program was set to 1 ) 42°C 90 min , 2 ) 50°C 2 min , 42°C 2 min ( repeat 14x ) , 3 ) 70°C 15 min , and 4 ) 4°C Hold . The cDNA was then added to a cDNA amplification mix with final concentrations of 1x KAPA HiFi HotStart ReadyMix ( Kapa Biosystems , KK2602 ) , 0 . 1 μM IS PCR primer ( 5′-AAGCAGTGGTATCAACGCAGAGT-3′ , ordered from idtdna . com ) , and RNase-free water . The cDNA amplification program was set to 1 ) 98°C 3 min , 2 ) 98°C 20 sec , 67°C 15 sec , 72°C 6 min ( repeat 17x ) , 3 ) 72°C 5 min , and 4 ) 4°C Hold . To clean the amplified cDNA , it was mixed with Ampure XP beads at a ratio of 1:1 ( v/v ) . The mixture was then placed on a magnetic bead stand to magnetize the cDNA-bound beads to side-wall of the tube and washed with 3 rounds of 80% ethanol . After removal of the final ethanol wash the beads were air dried for 3–4 minutes and observed frequently under a microscope . At the first sign of a dry crack in the beads , 17 . 5 μl of elution Buffer ( EB , 10 mM Tris-Cl , pH 8 . 5 ) was added , and incubated for 2 minutes . The sample was placed back on the magnetic bead stand for 2–3 minutes to separate the beads from the EB solution ( now containing clean cDNA ) and the EB solution was collected . cDNA concentration was measured by Qubit Fluorometer ( Thermo Fisher Scientific ) and the quality was analyzed by BioAnalyzer ( Agilent ) . The cDNA was tagmented using the Nextera DNA library prep kit ( Illumina , FC-121-1030 ) following the protocol in L . Serra , et al 2018 . Briefly , 20 ng of cDNA in 8 μl was mixed with 10 μl of Tagment DNA buffer and 2 . 2 μl of Tagment DNA enzyme from the kit . The mixture was incubated at 55°C for 5 minutes and cleaned up using the QIAquick DNA cleanup column ( QIAGEN , 28104 ) . The tagmented cDNA was then amplified using the Phusion High Fidelity PCR master mix ( New England Biolabs , M0531L ) with 30 μl of tagmented cDNA , 2 . 5 μl of Primer-1 ( Ad1_no MX ) , 2 . 5 μl of Primer-2 ( Ad2 . # ) , and 35 μl of Phusion High Fidelity PCR master mix buffer . The amplification program was set to 1 ) 72°C 5 min , 2 ) 98°C 30 sec , 3 ) 98°C 10 sec , 63°C 30 sec . , 72°C 1 min ( repeat 10x ) , and 4 ) 4°C Hold . The sample was then cleaned up with Ampure XP beads as described above except , scaling up to use 30 μl of EB and collecting 27 . 5 μl of the supernatant . Libraries were prepared and sequenced as paired-end , 43 base pair reads on the Illumina Nextseq 500 . Unstranded , paired-end 43 bp RNA-seq reads for each worm were mapped to the S . feltiae transcriptome downloaded from WormBase ParaSite ( WS263 ) using Bowtie 1 . 0 . 0 with the following options: -X 1500 -a -m 200—S—seedlen 25 -n 2—offrate 1 -p 64 -v 3 [87] . After Bowtie , gene expression was quantified with RSEM with the following options: rsem-calculate-expression—bam—paired-end . Gene expression for S . carpocapsae were performed as previously described [5] and reported in Transcripts Per Million ( TPM ) . We used counts for differential gene expression analysis . Reads for single worm RNA-seq samples were submitted to Gene Expression Omnibus ( GEO ) under the accession number GSE119223 . The Transcript per million ( TPM ) generated by rsem-calculate-expression for S . feltiae samples were normalized according to groups using the R package limma [88] because samples were collected , processed and sequenced in different batches . Samples were batched corrected between 3 and 9 hours in vitro to 6 hours in vitro , 3 , 6 , 9 , 12 , 15 hours in vivo with edgeR package removebatcheffects with log2 of TPM matrix . Normalization and batch correction for S . carpocapsae were done as previously described [5] . Log2 of the average TPM+1 ( transcripts per million , RNA levels ) and Log2 of the emPAI ( protein abundance levels ) for the 266 genes of S . feltiae ESPs was plotted in Rstudio using the package ggplot2 [89] . Pearson’s correlation and Spearman’s rank correlation were calculated using Excel . Differential gene expression was determined using edgeR [32] . Counts were normalized by library size using calcNormFactors . Genes were called differentially expressed if FDR < 0 . 05 and fold change > 2 . The list of genes that were differentially expressed ( DE ) using edgeR were used to create a TPM matrix . Gene expression in TPM were clustered using Cluster 3 . 0 [90] with the following options: log transformed , mean centered , normalized . Then genes were hierarchically clustered with center correlation . Heatmap were visualized with Java TreeView [91] . Heatmap for Fig 5C were done using the R package heatmap . 2 with centroid hierarchical clustering by row . MaSigPro was run as a two-time series to evaluate the differences and similarities of gene expression between in vitro and in vivo time course with 5670 differentially expressed genes found with edgeR between 6 hours in vitro activated and naïve IJs . Gene ontology enrichment analyses was calculated using Blast2GO Fisher’s exact test and considered statistically significant if FDR < 0 . 05 [92] . List of genes used in Blast2GO were differentially expressed according to edgeR or dynamically expressed according to maSigPro . We obtained a list of N:N orthologs and paralogs between S . feltiae and S . carpocapsae from WormBase ParaSite Biomart . List were obtained by choosing S . feltiae genome as query to find orthologs and paralogs in S . carpocapsae . List of venom proteins for S . carpocapsae were obtained from Lu et al . 2017 and compared to list of S . feltiae venom proteins . Orthology analysis was done with edgeR with function “match” . In determining the orthology of S . feltiae L889_g32029 ( Sf-flp-21 ) to C . elegans flp-21 , we relied on the predicted sequence of the mature peptide [34 , 93] . Using this method , we determined that , similar to Sc-flp-21 , Sf-flp-21 has an identical predicted mature peptide as the flp-21 from C . elegans .
|
In this study we found a core set of 52 venom proteins conserved between two insect-parasitic nematodes Steinernema feltiae and Steinernema carpocapsae , that are released when initially exposed to host tissue . Most of these proteins are conserved in mammalian-parasitic nematodes suggesting that this core set of proteins is important for parasitic nematodes in general . We show that the relevance of in vitro model systems to in vivo model systems needs to be optimized and experimentally measured . Using an in vitro model of parasitic nematode activation , we stimulated protein release from S . feltiae and evaluated its activity in vivo . This activation model was previously developed using S . carpocapsae and we conclude that this method is robust and can be generalized to other EPNs ( entomopathogenic nematodes ) . We found notable characteristics of S . feltiae venom including time-dependent decreases in protein amount and toxicity after exposure to host tissue , which differs from what has been previously reported for other EPNs , illustrating diversity in parasitic strategies among EPNs . Additionally , naïve S . feltiae infective juveniles ( IJs ) not exposed to host tissue release considerable amounts of protein . These proteins however are not toxic and differ in composition from those of activated IJs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials/methods"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"toxins",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"enzymology",
"parasitic",
"diseases",
"animals",
"toxic",
"agents",
"toxicology",
"nematode",
"infections",
"toxicity",
"venoms",
"proteins",
"gene",
"expression",
"insects",
"arthropoda",
"biochemistry",
"eukaryota",
"genetics",
"protein",
"domains",
"biology",
"and",
"life",
"sciences",
"proteases",
"organisms"
] |
2019
|
A core set of venom proteins is released by entomopathogenic nematodes in the genus Steinernema
|
Genome-wide association studies ( GWAS ) have begun to identify the common genetic component to ischaemic stroke ( IS ) . However , IS has considerable phenotypic heterogeneity . Where clinical covariates explain a large fraction of disease risk , covariate informed designs can increase power to detect associations . As prevalence rates in IS are markedly affected by age , and younger onset cases may have higher genetic predisposition , we investigated whether an age-at-onset informed approach could detect novel associations with IS and its subtypes; cardioembolic ( CE ) , large artery atherosclerosis ( LAA ) and small vessel disease ( SVD ) in 6 , 778 cases of European ancestry and 12 , 095 ancestry-matched controls . Regression analysis to identify SNP associations was performed on posterior liabilities after conditioning on age-at-onset and affection status . We sought further evidence of an association with LAA in 1 , 881 cases and 50 , 817 controls , and examined mRNA expression levels of the nearby genes in atherosclerotic carotid artery plaques . Secondly , we performed permutation analyses to evaluate the extent to which age-at-onset informed analysis improves significance for novel loci . We identified a novel association with an MMP12 locus in LAA ( rs660599; p = 2 . 5×10−7 ) , with independent replication in a second population ( p = 0 . 0048 , OR ( 95% CI ) = 1 . 18 ( 1 . 05–1 . 32 ) ; meta-analysis p = 2 . 6×10−8 ) . The nearby gene , MMP12 , was significantly overexpressed in carotid plaques compared to atherosclerosis-free control arteries ( p = 1 . 2×10−15; fold change = 335 . 6 ) . Permutation analyses demonstrated improved significance for associations when accounting for age-at-onset in all four stroke phenotypes ( p<0 . 001 ) . Our results show that a covariate-informed design , by adjusting for age-at-onset of stroke , can detect variants not identified by conventional GWAS .
Genome-wide association studies ( GWAS ) in ischaemic stroke have begun to identify the common genetic variants that confer risk of the disease . However , there is considerable heterogeneity present in stroke phenotypes: GWAS analyses have primarily looked at the three main subtypes; cardioembolic ( CE ) , large artery atherosclerosis ( LAA ) and small vessel disease stroke ( SVD ) . Within these subtype analyses , numbers of cases are smaller , but the expectation is that the effects of SNPs identified within the subtypes will be considerably larger . Indeed , all validated GWAS SNPs for ischaemic stroke to date have been stroke subtype-specific [1] , [2] , [3] , [4] , [5] , indicating the importance of subtyping of cases . Clinical risk factors are important in stroke; as many as 77% of first-ever stroke patients are hypertensive [6] , and other factors such as diabetes mellitus and elevated serum cholesterol confer a considerable proportion of disease risk [7] . These risk factors increase in prevalence in older age groups , suggesting older stroke patients may have a reduced stroke-specific genetic contribution . Indeed , IS is uncommon in individuals below middle age , but increases greatly in prevalence beyond the age of 65 [8] , with a lifetime risk of 1 in 5 for women and 1 in 6 for men [9] . Under the assumptions of the liability threshold model , the low prevalence of IS in younger age ranges suggests that individuals who do suffer strokes in this age group are likely to have an increased genetic predisposition . This is supported by family history data; with stronger family history seen in younger onset cases [10] , [11] , [12] , and twin studies [13] , which suggest that early onset cases may have higher heritability . We recently showed stronger effects for all stroke-associated SNPs in younger age groups , found evidence genome-wide that a significant number of SNPs show stronger association p-values when the oldest cases are removed , and showed increased pseudoheritability estimates for younger onset cases in certain stroke subtypes , thereby supporting this hypothesis [14] . However , the question of how best to integrate this information into GWAS analyses of ischaemic stroke remains unanswered . Previous GWAS have analysed younger subsets of ischaemic stroke cases [1] , [15] , but this approach may not be optimal for existing GWAS datasets if the increase in odds ratios for SNPs in younger cases are not sufficient to justify discarding a large proportion of the ascertained cases . All previous young onset analyses have been restricted to all ischaemic stroke cases versus controls; this may be particularly relevant given that all known loci for ischaemic stroke to date are for stroke subtypes [16] . A recent publication [17] , outlined a novel method of informing genetic association analyses on important clinical covariates . Using the liability threshold model in conjunction with estimates of disease prevalence for individuals with specific clinical covariates , the method estimates posterior disease liabilities for each individual in a GWAS , and uses these liabilities in regression analyses to test for association with genome-wide SNPs . This approach avoids issues due to multiple testing across age-at-onset thresholds , and provides a simple solution that is rooted is previous epidemiological research . In the present study , we extend the clinical covariate informed analysis approach to imputed genotypes , informing our analyses on the age-at-onset to identify novel variants associated with IS . We perform a genome-wide analysis with four stroke phenotypes ( IS , CE , LAA , SVD ) , and then determine the utility of the approach in ischaemic stroke GWAS , testing whether SNPs increase in significance .
We performed age-at-onset informed association analysis for a total of 6 , 778 ischaemic stroke cases and 12 , 095 controls across four ischaemic stroke phenotypes; all IS and the three major subtypes: CE , LAA , and SVD ( Table 1 ) ; with 1 , 637 , 1 , 316 , and 1 , 108 cases in the CE , LAA and SVD analyses respectively . With the exception of the young Milanese cohort , the age-at-onset distributions were similar in all cohorts ( Table S3 ) . We identified a group of twenty SNPs proximal to MMP3 and MMP12 on chromosome 11 in the LAA subtype that met our criteria for replication . The strongest associated of these was rs662558 ( p = 1 . 4×10−7 ) , a SNP that is in 1000 Genomes , but not HapMap II . Therefore , to enable replication in existing METASTROKE datasets , which were imputed to HapMap II , we selected the most strongly associated SNP from the HapMap II panel , which was in perfect LD with the lead SNP in our discovery meta-analysis ( rs660599: uninformed , p = 1 . 6×10−6; informed , p = 2 . 5×10−7; Figure 1 ) [16] . We found no evidence of between-study heterogeneity at either SNP ( Cochran's Q p = 0 . 22 and p = 0 . 19 for rs662558 and rs660599 , respectively ) . The evidence of an age-at-onset effect at rs660599 was p = 0 . 011 ( from permutations ) . We calculated age-at-onset quartiles for all large artery stroke cases from the discovery cohorts , and used these to evaluate this region at different age-at-onset thresholds . The median age-at-onset was 71 years , and the interquartile range was between 61 and 78 years . Post-hoc analyses of rs660599 in the discovery cohorts using logistic regression ( full details in Text S2 ) showed considerably stronger associations in younger age-at-onset quantiles ( Q1; OR ( 95% CI ) = 1 . 83 ( 1 . 46–2 . 30 ) , Q1–Q2; 1 . 56 ( 1 . 33–1 . 83 ) , Q1–Q3; 1 . 30 ( 1 . 14–1 . 49 ) , Q1–Q4; 1 . 30 ( 1 . 15–1 . 46 ) ) . No other regions met our criteria for replication . The associated locus was evaluated in a further 1 , 881 large artery stroke cases and ancestry matched controls in 9 cohorts from METASTROKE ( Table 2 ) . We found evidence for replication of the SNP ( rs660599 ) in all large artery stroke cases of European Ancestry ( p = 0 . 0048 , OR ( 95% CI ) = 1 . 18 ( 1 . 05–1 . 32 ) ) . Combining this result with the discovery p-value gave a genome-wide significant p-value of 2 . 6×10−8 ( Table 3 ) . Secondly , we used the Han and Eskin random effects meta-analysis approach to evaluate the association [18] after including a further 355 cases and 1 , 390 controls of Pakistani ancestry . The evidence for replication in this sample was p = 0 . 0063 , giving an overall p-value of 3 . 4×10−8 . Age-at-onset information was available across all age-at-onset quantiles for a subset of the replication studies ( 1 , 240 cases , 9 , 238 controls; ASGC , HVH , ISGS/SWISS , MGH-GASROS , Utrecht ) . We evaluated the SNP ( rs660599 ) in these studies at different age-at-onset quantiles using logistic regression , meta-analysing as previously . We again found the strongest effects in the youngest age quantile , consistent with a stronger effect in younger onset cases ( Q1; OR ( 95% CI ) = 1 . 27 ( 1 . 02–1 . 57 ) , Q1–Q2; 1 . 18 ( 1 . 00–1 . 39 ) , Q1–Q3; 1 . 22 ( 1 . 05–1 . 40 ) , Q1–Q4; 1 . 22 ( 1 . 07–1 . 41 ) ) . mRNA expression of the two proximal genes , MMP3 and MMP12 was analysed from 29 carotid , 15 abdominal aorta , 24 femoral plaques , and 28 atherosclerosis free left internal thoracic artery controls . MMP12 expression was upregulated in carotid plaques compared with left internal thoracic artery controls ( P = 1 . 2×10−15; fold change [FC] = 335 . 6 ) . It was also upregulated in femoral plaques ( P = 3 . 2×10−14; FC = 306 . 0 ) and abdominal plaques ( P = 5 . 0×10−11; FC = 399 . 3 ) compared with controls . Conversely , MMP3 was not significantly overexpressed in carotid , femoral or abdominal plaques versus controls ( p>0 . 05 ) . Eight SNPs were identified that were perfect proxies ( r2 = 1 ) with the associated SNP ( rs660599 ) in the region . Seven of the SNPs were in an intergenic region between MMP3 and MMP12 , while one fell within an intron of MMP12 . We investigated the evidence that any of these SNPs are functional variants using RegulomeDB [19] . Of the eight SNPs , we found strong evidence that one of these SNPs ( rs586701 ) affects binding . The SNP overlaps both CHIP-seq and DNA-seq peaks from ENCODE analyses , indicating that there is open chromatin in the region , and therefore that the SNP is likely to be functional . There is also evidence from a separate CHIP-seq analysis that the SNP affects protein binding [20] , and evidence from multiple sources that the SNP overlaps a predicted motif [21] , [22] , [23] . Histone modifications were observed in CHIP-seq experiments from ENCODE in a number of cells types , including Human umbilical vein endothelial ( Huvec ) cells . Two other SNPs ( rs17368582 , rs2276109 ) in moderate LD with the associated SNP ( r2 = 0 . 64 ) have been previously shown to directly influence MMP12 expression by affecting the affinity of an AP-1 binding site in the MMP12 promoter region [24] , [25] . Using RegulomeDB , we found further evidence from ENCODE that one of these SNPs ( rs2276109 ) is indeed functional , giving evidence that the associated locus in this analysis is likely to affect MMP12 expression through altered transcription . Detailed results for all analysed SNPs are given in Table S1 . Additionally , we investigated if these SNPs ( rs17368582 , rs2276109 , rs586701 ) were associated with MMP12 expression in tissues from the GTEx project [26] . However , we could not confirm an association with MMP12 expression in any relevant tissues ( p>0 . 4 in whole blood , tibial artery , aortic artery ) . Finally , we evaluated the overall utility of the age-at-onset informed approach in permutation analyses for SNPs that met p-value thresholds in the case control discovery data set . We generated 1000 permutations of age-at-onset within each centre , and performed age-at-onset informed analysis and subsequent meta-analysis for these SNPs , in the relevant stroke subtype . We compared the sum of the meta-analysis Z scores from all SNPs with p<0 . 05 in the observed age at onset informed meta-analysis with those from permutations . At this p-value selection threshold , we found strong evidence ( p<0 . 001 ) for genome-wide age-at-onset effects in each of the stroke phenotypes , with consistently increased summed Z scores in the observed age-at-onset informed meta-analysis compared to the permutations ( Figure 2 , red points , right hand axis ) . These results suggest that many of the risk variants for each stroke subphenotype have a higher frequency in younger onset cases . As the p-value selection threshold decreased , the summed Z score statistic became less significant in each stroke type , possibly reflecting lower overall power when fewer SNPs are included , even as these SNPs may have larger average effects . Further details are seen from the median proportion of SNPs more significant in the age-at-onset informed analysis than in the permutations ( Figure 2 , blue points , left hand axis ) . For CE and LAA stroke , the proportions increased with more stringent p-value thresholds ( from 52 . 1% to 56 . 3% for p<0 . 05 and p<0 . 00005 thresholds in CE , and from 51 . 4% to 56 . 0% for p<0 . 05 and p<0 . 00005 thresholds in LAA ) . Interestingly , in the all ischaemic stroke analysis the median proportion of SNPs more significant in the observed results than permutations dropped from 55 . 1% for SNPs with p<0 . 05 to 49 . 2% for only SNPs with p<0 . 00005 . This result may indicate a reduced proportion of true associations at stricter p-value thresholds for all ischaemic stroke compared to the subtypes , which is consistent with the observation that all common variants associated with stroke are for stroke subtypes , rather than for the phenotype of all ischaemic stroke [16] . The previously reported GWAS associations from a recent ischaemic stroke meta-analysis ( 9p21 , HDAC9 , PITX2 , ZFHX3 ) were all found to be more significant using the age-at-onset informed approach than the uninformed analysis ( Figure 3 ) . The increase in significance ranged from over half an order of magnitude ( 7 . 9×10−9 to 1 . 5×10−9 for rs879324 in ZFHX3 , CE ) , to under half an order of magnitude ( 5 . 7×10−9 to 2 . 5×10−9 for rs2107595 in HDAC9 , LVD ) . To ensure these analysis methods were comparable , we calculated genomic inflation factors and plotted QQ-plots . These were similar in the standard and the age-at-onset informed approach ( Table S4 , Figure S1 , S2 ) . For these four associated SNPs , we further used the permuted data sets to assess the observation of increased significance in the age-at-onset informed analysis . We compared the observed meta-analysis p-value to those from the permutations , generating an empirical p-value by dividing the number of permutations more significant than the observed results by the number of permutations . In LAA stroke , we observed a significant age-at-onset effect ( p = 0 . 018 , 0 . 011 and 0 . 002 for the HDAC9 , MMP12 and 9p21-associated SNPs in Figure 3 , respectively ) . Similarly , for CE , we observed a significant age-at-onset effect for rs879324 ( ZFHX3 , p = 0 . 026 ) , and a near-significant effect in rs6843082 ( PITX2 , p = 0 . 081 ) . This result provides further evidence that risk variants associated with ischaemic stroke subtypes have a stronger role in younger onset cases , and suggests that the age-at-onset informed approach will produce improved significance when the magnitude of genetic effects are stronger in younger onset cases .
We used a large GWAS dataset to evaluate the utility of an age-at-onset informed analysis approach to ischaemic stroke , and to identify novel variants associated with ischaemic stroke phenotypes . We identified a novel MMP12 locus that is associated with large artery atherosclerotic stroke , and verified that the age-at-onset informed approach produces improved significance for loci associated with each of the stroke phenotypes studied , as well as demonstrating that it increased the significance of four previous GWAS associations with ischemic stroke , all without systematic inflation of the test statistic . Importantly , the novel associated SNP would not have been identified using a standard logistic regression framework . We identified a group of SNPs proximal to Matrix Metalloproteinase 12 ( MMP12 ) that showed increased significance when using the age-at-onset informed approach . The increase in significance from the equivalent uninformed analysis was of almost an order of magnitude ( from p = 1 . 6×10−6 to p = 2 . 5×10−7 for rs660599 ) . We took a single SNP from this region forward for replication in an independent dataset , finding further evidence that the region is associated with large artery stroke . Two SNPs ( rs17368582 , rs2276109 ) in this LD-block have previously been shown to directly influence MMP12 expression by affecting the affinity of an AP-1 binding site in the MMP12 promoter region [24] , [25] , and another variant in this block ( rs17361668 ) is associated with increased fibrinogen levels , leading to an increased risk of developing advanced carotid atherosclerotic lesions , and an increased risk of myocardial infarction . We identified a second functional candidate ( rs586701 ) , which falls within both CHIP-seq and DNA-seq peaks from ENCODE , and is in complete LD with the associated SNP in our analysis . We investigated mRNA expression of MMP12 and MMP3 in carotid atherosclerotic plaques in individuals from the Tampere Vascular Study . MMP12 was overexpressed in diseased tissue compared to healthy controls , while no significant difference was found for the other nearby gene , MMP3 . MMP12 is a member of the Matrix Metalloproteinase ( MMP ) family of proteases , which are capable of degrading extracellular matrix proteins , and have a prominent role in atherosclerosis . They are thought to promote macrophage invasion [27] , [28] , [29] , promote angiogenesis [30] , and show increased activity in atheromatous plaques [31] . MMP12 deletions are associated with smaller , more stable lesions in the brachiocephalic artery of rabbits [32] , and reduced elastin degradation in the aortic arch [33] , indicating that MMP12 may have a role in destabilising plaques . Studies in humans have found MMP12 is localized to the core of advanced plaques , in macrophages with decreased arginase-I expression [34] , that MMP12 localizes selectively to macrophages at the borders of the lipid core [35] , and that MMP12 is significantly overexpressed in ruptured plaques when compared with thick or thin cap plaques , or with plaques with pathological intimal thickening [36] . This indicates that MMP12 is likely be involved in late-stage plaque instability: our study suggests that genetic variation impacts on this process . Secondly , we performed extensive permutation analyses to assess the utility of the age-at-onset informed approach genomewide . In each phenotype studied we found evidence that SNPs were more strongly associated using the approach than would be expected by chance , indicating that multiple risk variants are likely to be more common in younger onset cases . The significance was strongest when more SNPs were included in the analysis , which likely reflects the cumulative impact of age-at-onset effects on many SNPs . An alternative explanation might be that the increased significance for lower p-value thresholds is the result of the cumulative effects of subtle confounding . However , this is unlikely because any subtle biases will also be present in the permutations , and should therefore not affect the significance of the results . This result supports observations from family history and prospective cohort studies , which have observed stronger effects in younger onset cases [6] , [11] . Furthermore , all known associations with stroke were more significant using the age-at-onset informed approach . The increase in significance was around half an order of magnitude ( e . g from p = 7 . 9×10−9 to 1 . 5×10−9 for ZFHX3 , Figure 2 ) , and was significant in all but one locus , as assessed by permutation . Taken together , these results indicate that age-at-onset is an important measure to stratify stroke cases , and show that , as expected by theory [17] , integrating this information into association studies is likely to increase power to identify novel loci when the relative contribution of genetic is dependent on age-at-onset . Our study has limitations . We used imputed data from the Immunochip platform , meaning we only had access to ∼40% of the genome across all centres . Secondly , cases were drawn from a number of international centres , meaning that despite efforts to standardize phenotyping , we cannot rule out differences in screening and clinical ascertainment . Of complex diseases , IS has a particularly large degree of heterogeneity , exemplified by the fact that all validated associations identified to date have been within subtypes defined by clinical and radiological information . Further heterogeneity by risk factor and clinical covariate profiles is likely to exist , but the optimal method of incorporating this information into analyses remains an unanswered question . Our results indicate that a covariate-informed design , conditioning on age-at-onset of stroke , can unearth further associated variants . We provide evidence for this by identifying an association with a novel MMP12 locus in large artery stroke , supported by increased mRNA expression of the implicated gene in carotid plaques . GWAS in ischaemic stroke have begun to identify the genetic component of the disease , but these results are not yet clinically useful . Our study suggests that a more refined approach to analysis of genetic data , incorporating covariate information , is an important step in this process , and will help to ensure success in future GWAS .
All studies were approved by their local ethics committees; all patients gave informed consent . The initial dataset consisted of 6 , 778 ischaemic stroke cases of European ancestry and 12 , 095 ancestry-matched controls from the Wellcome Trust Case-Control Consortium II project in ischaemic stroke [1] , as well as a cohort from Milan , Italy [16] . These included 2 , 858 cases and 5 , 716 matched controls genotyped using the Immunochip platform; and 3 , 940 cases genotyped using either the Illumina 610 k or 660 k platforms matched with 6 , 379 controls genotyped on the Illumina Human 1 . 2M Duo ( UK ) , Illumina Human 550 k ( German ) and Illumina 610 k platforms ( Italian ) ( Table 1 ) . The Immunochip cases were described in the previous WTCCC2 ischaemic study , where they formed the replication effort [1] , as well as in a recent paper [37] . Genotyping of the five Immunochip case cohorts on the commercially available Immunochip array ( Illumina , San Diego , CA , USA ) was performed at the Sanger Centre , Hinxton , Cambridge UK . Swedish controls were provided and genotyped by the Swedish SLE network , Uppsala , Sweden . Belgian control samples were provided through the efforts of the International Multiple Sclerosis Genetics Consortium ( IMSGC ) . German controls were derived from the PopGen biobank , [38] . UK controls were derived from the 1958 Birth cohort . Any of the 1958 Birth controls overlapping with those from the WTCCC2 datasets , as assessed by IBD estimates , were removed prior to analysis . Standard quality control procedures were undertaken on all centres , before centre-wise imputation to the 1000 Genomes phase 1 integrated variant set ( March 2012 ) , using IMPUTE v2 . 2 . 0 [39] , [40] . SNPs with poor imputation quality ( info<0 . 3 ) or low minor allele frequency ( MAF<0 . 01 ) were discarded . Ischemic stroke was defined as a typical clinical syndrome with radiological confirmation; ascertained cases were classified into individual stroke subtypes using the Trial of Org 10172 in acute stroke ( TOAST ) criteria in all centres [41] . Age-at-onset was defined as age at first hospital admission for stroke; where this information was unavailable , age at blood draw was used ( 7 . 3% of cases ) . The age-at-onset and gender distributions of the populations are given in Table S3 . Age-at-onset quantiles were calculated from all the cases from the discovery datasets in the four stroke phenotypes ( all IS and the three stroke subtypes: CE , LAA , SVD ) and these were used to evaluate associated loci at different age-at-onset thresholds . The prevalence of ischaemic stroke by age was obtained from a recent publication [9]; gender-specific estimates were averaged , and prevalences within each of the stroke subtypes were assumed to be approximately 20% of the overall total , similar to proportions seen in population-based studies [42] . We modeled phenotype data using a continuous unobserved quantitative trait called the disease liability , which we used to approximate the effect of age-at-onset on the liability scale , based on estimates of ischaemic stroke prevalence by age from epidemiological data ( full details in Text S2 ) . We developed two models for our analysis; one based on the prevalence rates for all ischaemic stroke cases , and secondly for the three stroke subtypes . We used these models to calculate posterior mean liabilities after conditioning on age-at-onset for the four stroke phenotypes separately . Controls were modeled in the same way , but were assumed to take the posterior mean from the lower ( unaffected ) portion of the distribution in the liability threshold model . Where age data was missing , individuals were assigned the median age value . Full descriptions of the models used and the formulae used to calculate posterior mean liabilities are given in Text S2 . Regression was then performed on posterior liabilities by multiplying the number of samples by the squared correlation between the expected genotype dosage and posterior mean liabilities for each of the discovery cohorts in the four ischaemic stroke phenotypes ( CE , LAA , SVD , IS ) , following a previous approach [17] . Ancestry-informative principal components were included where appropriate ( 6 of 8 centres ) , using the EIGENSTRAT procedure [43] . All analysis was performed using the R statistical software . The results from each centre were meta-analysed for each of the four phenotypes using Stouffer's Z-score weighted approach , as implemented in METAL [44] . Genomic control was used to correct for any residual inflation due to population stratification [45] . Between-study heterogeneity was assessed using Cochran's Q statistic . We considered only SNPs present in at least 75% of the cases , and with no evidence of heterogeneity ( Cochran's Q p-value>0 . 001 ) . All SNPs analysed were either genotyped or imputed in both the Immunochip and the genome-wide datasets . After meta-analysis , the resulting p-values were compared with the equivalent values from an unconditioned analysis . For SNPs more significant in the age-at-onset informed analysis and with p<5×10−6 , we determined the evidence of a true age-at-onset effect by generating 1000 permutations of age-at-onset and rerunning the age-at-onset informed analysis , meta-analysing as previously . We calculated an empirical p-value by dividing the number of permuted observations showing greater significance in the meta-analysis than the observed results by the number of permutations . Any novel SNP with a meta-analysis p<5×10−6 and evidence of an age-at-onset effect at p<0 . 05 were taken forward for replication . We set the experiment-wide significance threshold at p<5×10−8 . Replication of an associated variant was performed in a further 10 cohorts from METASTROKE . Nine of the centres used a cross-sectional design , while one was a large prospective , population based cohort ( ARIC ) . Nine of the centres were of European ancestry , while one consisted of individuals of Pakistani ancestry ( RACE ) ( Table 2 ) . All centres used a case-control methodology; centres with a cross sectional design used logistic regression to model the association of genotype dosages from imputation with the dichotomous outcome of ischaemic stroke and prospective cohorts used Cox proportional-hazards models to evaluate time to first stroke , fitting an additive model relating genotype dose to the stroke outcome . European ancestry replication centres were meta-analysed using a fixed effects inverse-variance weighted method . To assess the evidence for association of the SNP for replication samples of all ancestries , we performed a trans-ethnic meta-analysis using a random-effects model to control for any resulting heterogeneity [18] . To evaluate the overall evidence for association , the results of the discovery and replication analyses were combined using Fisher's Method . Expression of the two genes proximal to the associated variant was tested in atherosclerotic plaques from the Tampere Vascular study [27] , [46] , [47] , [48] , [49] . Carotid , femoral , and aortic atherosclerotic plaques constituting the intima and inner media were prospectively obtained between 2005 and 2009 from patients fulfilling the following inclusion criteria: ( 1 ) carotid endarterectomy attributable to asymptomatic or symptomatic >70% carotid stenosis , or ( 2 ) femoral or ( 3 ) aortic endarterectomy with aortoiliac or aortobifemoral bypass attributable to symptomatic peripheral arterial disease . Whole thickness left internal thoracic artery samples obtained during coronary artery bypass surgery and identified as being microscopically atherosclerosis free were used as controls . The patients were consecutively recruited and stratified according to indication for surgery . All open vascular surgical procedures were performed at the Division of Vascular Surgery and Heart Center , Tampere University Hospital . Fresh tissue samples were immediately soaked in RNALater solution ( Ambion Inc ) and homogenized using an Ultra-Turrax T80 homogenizer ( IKA ) . RNA was extracted with the Trizol reagent ( Invitrogen ) and miRNEasy Mini-Kit ( Qiagen ) with the RNase-Free DNase Set ( Qiagen ) according to manufacturer instructions . The RNA isolation protocol was validated by analyzing the integrity of the RNA with the RNA 6000 Nano Chip Kit ( Agilent ) . The expression levels were analyzed with an Illumina HumanHT-12 v3 Expression BeadChip ( Illumina ) . In brief , 300–500 ng of RNA was reverse transcribed in cRNA and biotin-UTP labeled using the IlluminaTotalPrep RNA Amplification Kit ( Ambion ) , and 1500 ng of cRNA was then hybridized to the Illumina HumanHT-12 v3 Expression BeadChip . The BeadChips were scanned with the Illumina iScan system . After background subtraction , raw intensity data were exported using the Illumina Genome Studio software . Further data processing was conducted by means of R language and appropriate Bioconductor modules . Data were log2-transformed , and robust multichip average and robust spline normalization ( rma_rsn ) were used . Accuracy of the expression array was validated with qRT-PCR [50] . mRNA Expression levels in the tissues were determined; a fold change statistic was estimated between the two tissues , and significance was calculated using a t test . Recent evidence indicates that a significant proportion of GWAS SNPs fall within regions that are likely to affect binding of nearby proteins , such as transcription factor binding sites [51] , [52] . We used the RegulomeDB database to access regulatory information from ENCODE and other existing publications [19] , investigating the evidence that the SNPs in the associated locus have a regulatory function . First , the linkage-disequilibrium ( LD ) patterns amongst the most strongly associated SNPs were determined . We then used PLINK to determine the LD structure of the associated region , using LD-patterns from the 85 Utah residents from the 1000 Genomes project [53] , [54] . All SNPs with r2>0 . 6 were identified within a 2 , 000 kb window from the index SNP . All of the SNPs identified were then investigated using RegulomeDB to determine the evidence that any of the SNPs have a regulatory function . Permutation analysis was performed to evaluate the age-at-onset informed approach , to show that including age at onset information directly led to the increased significance , due solely to inclusion of age-at-onset information at tested SNPs . First , we identified a set of SNPs enriched for true association in the case control analysis of ischaemic stroke and subtypes . An expanded set of discovery and METASTROKE studies were analysed using standard case control methods and subsequent meta-analysis ( see Table S2 ) . SNPs with p<0 . 05 and no evidence of heterogeneity ( p>0 . 0001 ) were extracted and pruned for LD ( 300 kb window , r2<0 . 25 ) , leaving a set of almost independent SNPs for further analysis . Each retained SNP represented the most significant association in each LD block , as determined by the “clump” procedure in PLINK , based on LD patterns from the CEU individuals from 1000 Genomes . The number of SNPs used in each analysis is given in Table S5 . These SNP subsets were derived for ischaemic stroke , and for each stroke subset and then used in the age-at-onset informed analysis . Analysis was performed as previously for each stroke subtype using the age-at-onset informed method within studies and meta-analysis across studies ( giving observed results , as obtained above ) . We then performed a permutation study to obtain the expected distribution of p-values at these SNPs . Age at onset for cases was permuted within stroke subtypes within each study , and then the data were re-analysed , for 1000 permutations . Two summary statistics were constructed: ( 1 ) within permutations , we compared p-values from analysis of permuted age at onset with p-values from the observed data , and tabulated the proportion of SNPs with increased significance in the observed data set than in the permuted data set; across permutations , we calculated the median proportion of SNPs with increased significance in the observed data; ( 2 ) Within permutations , we converted each SNP p-value to a Z score and summed the absolute value of the Z score across SNPs ( sumZ ) . An empirical p-value for the age-informed analysis was calculated from the proportion of simulated data sets where sumZ exceeded the value in the observed analysis . This analysis was performed at SNP subsets defined from four SNP p-value thresholds in the discovery and METASTROKE studies: p<0 . 05 , p<0 . 005 , p<0 . 0005 , and p<0 . 00005 . Finally , we assessed the evidence of an age-at-onset effect at the four stroke loci identified in the METASTROKE ischaemic stroke collaboration ( 9p21 , HDAC9 , PITX2 , ZFHX3 ) [16] . For each SNP , we generated an empirical p-value from the proportion of permutations showing stronger association than in the observed age-at-onset informed analysis .
|
Ischaemic stroke places an enormous burden on global healthcare . However , the disease processes that lead to stroke are not fully understood . Genome-wide association studies have recently established that common genetic variants can increase risk of ischaemic stroke and its subtypes . In this study , we aimed to identify novel genetic associations with ischaemic stroke and its subtypes by addressing the fact that younger onset cases may have a stronger genetic component , and using this information in our analyses . We identify a novel genetic variant on chromosome 11 ( rs660599 ) , which is associated with increased risk of large artery stroke . We also show that mRNA expression of the nearest gene ( MMP12 ) is higher in arteries with the disease process underlying large artery stroke ( atherosclerosis ) . Finally , we evaluate our novel analysis approach , and show that our method is likely to identify further associations with ischaemic stroke .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"stroke",
"medicine",
"and",
"health",
"sciences",
"cerebrovascular",
"diseases",
"ischemic",
"stroke",
"genome",
"analysis",
"neurology",
"gene",
"expression",
"genetics",
"atherosclerosis",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"vascular",
"medicine",
"cardiology"
] |
2014
|
A Novel MMP12 Locus Is Associated with Large Artery Atherosclerotic Stroke Using a Genome-Wide Age-at-Onset Informed Approach
|
Herpes simplex virus type 1 ( HSV-1 ) is a neurotropic virus causing vesicular oral or genital skin lesions , meningitis and other diseases particularly harmful in immunocompromised individuals . To comprehensively investigate the complex interaction between HSV-1 and its host we combined two genome-scale screens for host factors ( HFs ) involved in virus replication . A yeast two-hybrid screen for protein interactions and a RNA interference ( RNAi ) screen with a druggable genome small interfering RNA ( siRNA ) library confirmed existing and identified novel HFs which functionally influence HSV-1 infection . Bioinformatic analyses found the 358 HFs were enriched for several pathways and multi-protein complexes . Of particular interest was the identification of Med23 as a strongly anti-viral component of the largely pro-viral Mediator complex , which links specific transcription factors to RNA polymerase II . The anti-viral effect of Med23 on HSV-1 replication was confirmed in gain-of-function gene overexpression experiments , and this inhibitory effect was specific to HSV-1 , as a range of other viruses including Vaccinia virus and Semliki Forest virus were unaffected by Med23 depletion . We found Med23 significantly upregulated expression of the type III interferon family ( IFN-λ ) at the mRNA and protein level by directly interacting with the transcription factor IRF7 . The synergistic effect of Med23 and IRF7 on IFN-λ induction suggests this is the major transcription factor for IFN-λ expression . Genotypic analysis of patients suffering recurrent orofacial HSV-1 outbreaks , previously shown to be deficient in IFN-λ secretion , found a significant correlation with a single nucleotide polymorphism in the IFN-λ3 ( IL28b ) promoter strongly linked to Hepatitis C disease and treatment outcome . This paper describes a link between Med23 and IFN-λ , provides evidence for the crucial role of IFN-λ in HSV-1 immune control , and highlights the power of integrative genome-scale approaches to identify HFs critical for disease progression and outcome .
Up to 90% of the global population is infected with the α-herpesvirus Herpes simplex virus type I ( HSV-1 ) . Whilst HSV-1 is largely responsible for outbreaks of vesicular oral skin lesions ( fever blisters , or cold sores ) , it can also cause a variety of more severe diseases including encephalitis , meningitis and keratitis [1] , [2] . Furthermore , the frequency of association with genital lesions ( previously associated mainly with HSV-2 infection ) is increasing . As co-infection with HSV is a significant contributing factor to transmission of the Human Immunodeficiency Virus ( HIV ) , our understanding of HSV disease , and herpesviruses in general , has wide implications for global healthcare . Like all herpesviruses , HSV-1 establishes lytic ( epithelial cells ) and asymptomatic latent infection ( sensory neurons in trigeminal and sacral ganglia ) which undergoes periodic reactivation [3] . The equilibrium between these two infection states requires a fine balance between innate and adaptive immune responses , and viral immune evasion mechanisms [4] . Whilst aspects of the HSV-1 replication cycle have been intensively investigated , there remain gaps in our understanding of the complexity of virus:host interactions . For example , a proteomics study identified over 100 changes in the cellular proteome within the first 6h of infection with HSV-1 [5] , and a recent analysis of virion-incorporated cellular proteins found that about 30% of these directly affected virus growth [6] . To systematically identify host factors ( HFs ) required for viral replication , RNAi screens have been performed with a range of different RNA and DNA viruses including HIV-1 [7] , [8] , [9] , Influenza A virus [10] , [11] , [12] , Hepatitis C virus [13] , West Nile virus [14] , Dengue virus [15] , Enterovirus [16] and Vaccinia virus [17] , [18] . The overlap between the results of these studies is generally very low [19] , reflecting either differences in biology , or different experimental set-ups , cutoff and selection criteria . In addition , microenvironmental effects might also play a role for the differences of the results [20] . Whilst loss-of-function siRNA screens provide functional information on specific genes , protein interaction studies can provide insight into the mechanism of action by identifying physical interaction partners between pathogen and host . Genome-scale virus-host protein interaction screens using the yeast-two-hybrid system have been performed for HCV [21] , Influenza A virus [22] , Epstein Barr virus ( EBV ) [23] , Vaccinia virus [24] , [25] , SARS coronavirus [26] and several non-human viruses [27] . Based on these genome-scale studies and individual interactions found by literature curation , several virus-host interaction databases have been created including the HIV-1 , human protein interaction database at NCBI [28] , VirHostNet [29] , VirusMINT [30] , PIG [31] and HPIDB [32] . Although there is little overlap between individual cellular interactors of different viruses , targeting of a number of cellular processes such as cell cycle regulation , nuclear transport and immune response appears to be conserved [33] . Understanding the complex interplay between viral and host components is critical to the definition of herpesvirus infection and pathogenesis . As herpesviruses encode a large number of proteins , in contrast to small RNA viruses such as HIV and Influenza , many cellular processes may be directly affected by viral proteins , and whilst there exists a wealth of information on individual viral proteins , there remain large gaps in our understanding of the HSV-1 life cycle and its interaction with its host . Here , we present data from the first integrative and systematic screening approach to characterise the role of cellular proteins in the HSV-1 life cycle . A genome-scale RNAi knockdown screen to identify HFs functionally influencing HSV-1 replication was performed in parallel with a yeast two-hybrid ( Y2H ) protein interaction screen to simultaneously gain insight into potential mechanisms of action . Combined analyses confirmed the importance of known cellular proteins involved in processes such as cell cycle , proteins transport and gene expression important for virus replication . Furthermore , we identified a subunit of the Mediator multi-protein complex , Med23 , as a key regulator of IFN-λ induction , which appears to be of crucial significance for the control of HSV-1 both in vitro and in vivo . These data demonstrate the power of a combined screening strategy to investigate pathogen:host interactions and identify novel host factors and cellular pathway targets for the development of essential clinical interventions .
Host factors ( HFs ) which positively or negatively regulate HSV-1 replication were identified by screening a druggable genome siRNA library ( 4 siRNAs per gene ) targeting 7 , 237 human genes against a HSV-1 reporter virus expressing the enhanced green fluorescent protein ( eGFP; HSV-1 strain C12 ) in the epithelial Hela cell line , due to their ease of transfection and susceptibility to HSV-1 infection [34] . To generate a robust and reliable dataset the screen was carried out three times in triplicate , with one replicate used in a cell viability assay to determine any cytotoxic effects of gene depletion and duplicates infected for the virus infection assay . The siRNA library was reverse-transfected into Hela cells before infecting with HSV-1 and monitoring virus growth kinetics as a measure of GFP-fluorescence ( Figure 1b ) . By following virus growth over multiple rounds of replication , host proteins involved in all stages of the virus life cycle can be identified . Replication slopes during linear growth were normalized to controls ( mock-transfected cells , and cells transfected with a siRNA unable to be processed by the RNA Silencing Complex , RSCF ) and the mean of six replicates was calculated . siRNAs found to be cytotoxic ( 81 in total ) were excluded from further analyses , and a hitlist of 358 containing the top 2 . 5% inhibitory and the top 2 . 5% enhancing HFs was generated ( Table S1 in Text S2 ) . The identified HSV-1 HFs were compared to datasets from published siRNA depletion screens aimed at identifying cellular factors affecting HIV-1 [7] , [8] , [9] , West Nile Virus ( WNV ) [14] , Hepatitis C Virus ( HCV ) [13] , Dengue virus [15] and Influenza A virus [10] , [11] , [12] . Of our 358 HFs , 54 cellular proteins ( 15 . 1% ) overlapped with these other virus screens ( Influenza A , 29; HIV-1 , 24; HCV , 6; WNV , 2; Dengue virus , 1 ) ( Figure 1c; Table S2 in Text S2 ) . HSV-1 is currently known to encode at least 84 proteins , expressed sequentially under strict temporal regulation during infection . To gain further mechanistic insight into host factors involved in HSV-1 infection , in parallel to the siRNA depletion screen we carried out a yeast two-hybrid protein interaction screen to identify cellular interaction partners of viral proteins . We generated a collection of 107 partial and full-length HSV-1 cDNA constructs and tested them for interactions with proteins encoded by a library of 12 , 381 human cDNA clones [35] . 231 HSV-1-human protein interactions were detected once ( low confidence ) , and 63 more than once ( high-confidence ) ( Table S3 in Text S2 ) . Using these high-confidence interactions , the previously reported HSV-1 interactome [36] was connected into a human interactome ( 62 , 310 published protein interactions ) to generate a combined pathogen-host interactome ( Figure S1a ) . Both degree centrality ( which indicates the number of interactions a protein has , where high values represent highly interactive ‘hubs’ ) and betweenness centrality ( which indicates the number of shortest paths between any pair of proteins passing through the protein considered ) were significantly increased for HSV-1 interactors , particularly in the high-confidence network ( Figure S1b–e ) . These data suggest HSV-1 proteins preferentially target highly connected central human proteins in the cellular interaction network , similar to other viruses [23] . Analysis of this interactome for HFs identified by RNAi found they were enriched in the fraction of cellular proteins that directly interact with viral proteins or that interact via one intermediate , in comparison to proteins that only interact via 2 or more intermediates ( p = 0 . 036 , Fisher's exact test ) ( Figure S1f ) . A direct comparison of HSV-1 protein interaction partners and the siRNA screen HFs found 215 genes in common . Of those , ten ( 4 . 6% ) were identified as a hit in both screens ( Table S4 in Text S2 ) , suggesting that these technologies identify complimentary yet not necessarily overlapping HFs . An extended literature and database search identified 599 cellular proteins that interact with or are involved in infection with human herpesviruses . The overlap between the high-confidence Y2H cellular interactors ( 63 ) and HFs ( 358 ) with this set was statistically significant ( p = 0 . 008 ) ( Figure 2a; Figure S1h; Table S5 in Text S2 ) . From this combined analysis , a subset of HFs was chosen for further validation . Protein interactions were tested in a mammalian cell system by LUMIER pull-down assay [37] . Of the 45 interactions tested , 26 ( 57 . 8% ) were confirmed , with 15 strongly positive ( z-score >2 ) and 11 weakly positive ( z-score 1–2 ) ( Figure S1g ) . siRNA deconvolution ( 4 siRNAs per gene tested individually ) was used to further validate 72 HFs ( Figure 2b; Figure S2 ) . The replication phenotype could be confirmed ( ≥2 or more siRNAs gave the same or better replication slope than observed in the primary screen ) in a high proportion ( 83 . 3% ) of candidates , highlighting the reliability of the primary screen dataset . Quantitative RT-PCR analysis of mRNA expression levels found a minimum depletion of 60% ( mean 88% ) in a subset of 52 genes ( data not shown; Table S6 in Text S2 ) confirming the observed effects on HSV-1 replication are genuine and not due to ‘off-target’ effects or insufficient gene knockdown . To further investigate the virus-specificity of our identified HFs , we tested this subset for their effect on the replication of an additional α-herpesvirus ( Varicella-Zoster virus , VZV ) , the β-herpesvirus Cytomegalovirus ( CMV ) , and a completely unrelated RNA virus , Semliki Forest Virus ( SFV ) . None of the three proteins which enhanced HSV-1 replication upon knockdown had an effect on either VZV or CMV , and one ( NR3C2 ) was even inhibitory for SFV ( Figure 2c ) . Of the 64 siRNAs which inhibited HSV-1 , 27 ( 42 . 2% ) were also inhibitory for VZV , 60 for CMV ( 93 . 8% ) and 23 ( 35 . 9% ) for SFV replication ( Table S7 in Text S2 ) . Some functional groups ( transcriptional regulators ) were required by most viruses , but there were notable differences between other proteins . For example IFITM-1 , previously identified as an inhibitor of Influenza A , Dengue virus and WNV [10] , inhibited VZV yet had a positive effect on HSV-1 replication . These data suggest that whilst there are some HFs which are broad in their effects on virus replication , a large proportion are species-specific . Functional and pathway analysis of the 358 HSV-1 HFs identified in the siRNA depletion screen ( Figure S3a ) , and of direct and indirect virus-host interactions with multiple interaction partners ( Figure S3b ) , found a significant enrichment of a wide range of cellular processes involved in multiple stages of virus replication ( Table S8 in Text S2 ) . Pathways included those involved in gene expression , transcription , splicing and translational regulation ( RNAi screen ) , and protein transport , cell cycle , and transcriptional repressor activity ( Y2H screen ) . A combined analysis of HFs from both screens found dominant functional categories centred on the regulation of transcription ( RNA polymerase II-associated genes , splicing factors , transcription activation and the Mediator complex ) ( Figure S3c , d ) . The physiological relevance of some HFs and pathways was confirmed by further biological validation . Protein transport pathways ( in the form of dynein microtubule networks ) are exploited by HSV-1 early after infection to shuttle viral capsids to the nucleus . These screens confirmed known interactions between dynein subunits and viral proteins , and identified additional previously unknown interactions ( Text S1 and Figure S4a ) . Several dynein chain subunits were found to be essential for virus replication , whilst the moderate effect of depletion of other subunits demonstrated a level of functional redundancy in HSV-1 capsid transport ( Figure S4b–e ) [38] , [39] . Intrinsic anti-viral host defense mechanisms , in the context of cellular E2 ubiquitin ligases , were also investigated . The immediate-early viral protein ICP0 , an E3 ubiquitin ligase , is crucial for blocking anti-viral defense mechanisms by degrading promyeloctic leukemia ( PML ) nuclear bodies ( ND10 domains ) in the presence of cellular E2-ubiquitin-conjugating enzymes ( E2s ) . Our siRNA screen found multiple E2s were required for this , and suggests that HSV-1 ICP0 is promiscuous in its exploitation of E2s to mediate PML degradation and ensure successful infection ( Text S1 and Figure S5 ) . Combined bioinformatic analyses of protein interaction and siRNA depletion screens found a significant functional enrichment for proteins involved in transcription , and identified multi-protein complexes enriched for pro-viral HFs which strongly inhibited HSV-1 upon depletion , including the RNA-polymerase II , eIF3 and Mediator complexes ( Figure 3a ) . The Mediator complex links the cellular transcription machinery ( RNA polymerase II ) to specific transcription factors , and the identification of many Mediator subunits as HFs in other viral siRNA depletion screens highlights its significant role in viral genome transcription [7] , [9] , [11] , [40] ( Table S2 in Text S2 ) . Further , several Mediator subunits ( Med25 , 29 , 17 and 8 ) are known to interact with the HSV-1 transactivator VP16 ( UL48 ) and other herpesviral proteins [41] ( Figure S6a ) . Consistently , the Mediator complex was found to be strongly required for HSV-1 replication , with depletion of the majority of subunits ( Med 4 , 6 , 7 , 8 , 14 , 16 , 17 , 21 , 25 , 26 , 27 and 28 ) leading to a severe reduction in virus replication in the primary screen ( Figure S6b ) or in confirmatory deconvolution assays ( Figure 3b ) . However , depletion of the Med23 subunit was striking in that it led to a significant enhancement of virus growth ( Figure 3b ) . Flow cytometry quantification found that removal of Med23 not only increased the total number of infected cells ( combination of GFPlo and GFPhi cells; 75 . 5% in comparison to 49 . 6% in mock-transfected cells ) but also the copy number of virus genomes ( GFPhi cells; 44 . 8% in comparison to 24 . 4% in mock-transfected cells ) ( Figure 3c ) . Gain-of-function experiments found overexpression of Med23 led to a corresponding inhibition of two strains of HSV-1 ( Figure 3d; Figure S6c ) , confirming Med23 is a natural anti-viral component of the pro-viral Mediator complex . This anti-viral effect of Med23 was specific for HSV-1 , as replication of VZV ( α-herpesvirus ) , hCMV ( β-herpesvirus ) , Vaccinia virus ( DNA ) and SFV ( RNA virus ) remained unaffected by Med23 depletion ( Figure 3e ) . Med23 could exert anti-viral effects either by having an inhibitory effect on viral transactivators or by interacting with and having a positive effect on an existing anti-viral factor . We first tested whether Med23 directly affects viral gene expression using luciferase reporters with HSV-1 promoters , however observed no inhibitory effect ( data not shown ) . Since the Mediator complex and Med23 in particular is known to be involved in Jak/Stat-mediated interferon signaling [42] , we used the lung epithelial cell line A549 and its Stat-1-deficient derivative A549-V [43] to determine if Med23 influences HSV-1 replication by modulating innate immunity . In the parental A549 cells the phenotype of HSV-1 replication was the same as that observed in Hela cells , where depletion of Med23 enhanced replication and over-expression inhibited virus growth . However , in the Stat1-deficient A549-V cells HSV-1 replication was unaffected by both depletion and over-expression of Med23 ( Fig . 4a ) , indicating that Med23 requires an intact Jak/Stat signalling pathway to exert its anti-viral effects . To determine which interferon may be responsible for the anti-viral effects of Med23 , A549 cells were depleted for Med23 and infected with HSV-1 following pre-stimulation with Type I ( IFN-α or IFN-β ) , Type II ( IFN-γ ) or Type III ( the distinct IFN-λ1 or the almost identical IFN-λ2 and -λ3 , termed IFN-λ2/3 ) interferons . Whilst treatment with IFN-α , -β and -γ significantly decreased HSV-1 replication levels , the observed ∼2-fold enhancement of HSV-1 replication following Med23 depletion was still seen . However , pre-treatment with the both IFN-λ1 and IFN-λ2/3 blocked the enhancing effect of Med23 depletion ( Figure 4b ) . Investigation into the effect of Med23 on interferon induction by qRT-PCR found that whilst Med23 over-expression induced IFN-β ( ∼3-fold increase ) , induction of IFN-λ1 and λ2/3 was considerably and statistically significantly higher ( ∼26-fold induction; p = 0 . 003 and 0 . 002 , respectively ) ( Figure 4c ) . This induction was specific , as levels of other cytokines and interferon-regulatory factors ( IRFs ) were unaffected by Med23 overexpression ( Figure S7a ) . Secretion of IFN-λ2/3 protein was also increased in all cell lines tested , but most significantly to ∼11-fold in A549 cells ( Figure 4d ) , which is consistent with a recent report showing that type III interferons are the dominant type of IFNs expressed by primary airway epithelial cells [44] . Furthermore , qPCR analysis found depletion of Med23 inhibited the induction of IFN-λ expression following HSV-1 infection of A549 cells in comparison to cells transfected with the RSCF siRNA control ( Figure 4e ) . Together , these data suggest that IFN-λ is responsible for the observed inhibitory effect of Med23 on HSV-1 replication . As IFN-λ expression is induced following activation of pathogen recognition receptors ( PRRs ) by virus infection [45] , [46] , [47] , [48] , we tested whether Med23 induced IFN-λ by directly interacting with an interferon-responsive transcription factor ( IRF ) . Y2H and confirmatory co-immunoprecipitation experiments in mammalian cells with a panel of IRFs found that Med23 interacted with IRF4 and IRF7 ( Figure 5a; Figure S7b ) . We also observed a weak interaction with IRF9 , which may explain the previously observed effect of Med23 on Jak/Stat signalling [42] . To determine if this interaction had a functional effect , we looked at whether Med23 influenced IRF-mediated induction of IFN-λ . In a luciferase reporter assay , neither IRF4 nor IRF9 led to a significant induction of the IFN-λ1 promoter , either alone or in conjunction with Med23 ( data not shown ) . IRF7 induced expression from the IFN-β and IFN-λ1 promoters to similar levels ( ∼7-fold and 9-fold higher than background , respectively ) , whilst the ISRE , induced by IRF7 and also present in the IRF7 promoter , was induced ∼15-fold ( Figure 5b ) . Whilst co-expression of Med23 with IRF7 had no further effect on IFN-β expression , a synergistic induction of the IFN-λ1 promoter and , to a lesser extent , the ISRE , was observed ( IFN-λ1 doubled to ∼18-fold , p = 0 . 02 ) ( Figure 5b ) . Interestingly , a Med23 mutant unable to induce immediate early gene expression via jun/fos ( R617Q , or R611Q in Med23 transcript variant 1 used here ) synergistically induced ISRE expression with IRF7 , yet was unable to further enhance IRF7-mediated induction of IFN-λ1 ( data not shown ) . A similar synergistic effect of Med23 and IRF7 was seen at the protein level , where co-expression increased supernatant levels of IFN-λ3 more than 2-fold those seen with Med23 or IRF7 alone ( Figure S7c , d ) . Successful disease and treatment outcome in Hepatitis C virus infection ( demonstration of a sustained virologic response ) is strongly associated with a single nucleotide polymorphism ( SNP ) in the IFN-λ3 promoter ( rs12979860; CC genotype over CT or TT ) and higher plasma levels of IFN-λ3 [49] , [50] . Furthermore , IFN-λ expression is impaired in a cohort of ethnically Italian individuals suffering recurrent HSV-1-related herpes labialis reactivation [51] . To determine if the clinical severity of HSV-1 disease is due to the observed deficiency in IFN-λ expression , we screened a subset of the recurrent herpes labialis ( HL ) cohort and additional subjects for the IFN-λ3 promoter polymorphism . Genotypic analysis found the presence of a T ( CT or TT genotype ) had a dose-dependent association with clinical severity , with the homozygous TT genotype being more prevalent as disease severity increases ( Figure 6 ) . In spite of the relatively small sample numbers in some clinical categories ( Table 1 ) , the association of a CT or TT genotype with the most severe recurrence of herpes labialis ( H+ ) was statistically significant ( p = 0 . 014; Fishers's exact t-test ) . As the CC genotype is directly associated with increased IFN-λ3 levels [51] , these data highlight a previously unknown association between the frequency/severity of recurrence of herpes labialis , the CT/TT genotype and subsequent reduction in secretion of IFN-λ3 . It is of importance to investigate this genotype association with a larger cohort of HL patients , as well as those suffering with other HSV-1-related disease , in order to determine the role of IFN-λ in the full spectrum of HSV-1 pathogenesis . Taken together , these data identify Med23 as a novel anti-viral factor which acts as a key regulator of IFN-λ expression by interacting with and enhancing the activity of IRF7 , a major transcription factor involved in innate immunity . Our observation of a link between the clinical severity of HSV-1 disease and , a CT/TT genotype at a SNP known to regulate IFN-λ3 secretion demonstrates the significance of IFN-λ in the control of HSV-1 replication in vivo . Whilst this study provides no direct link between the IFN-λ3 promoter polymorphism and Med23 , these associations of IFN-λ with HSV-1 disease , combined with our observations that Med23 is required for the induction of IFN-λ following HSV-1 infection , identifies for the first time a link between Med23 and IFN-λ , provides a clinical context for Med23 regulation of IFN-λ expression and underscores the potential biological significance of these data . The use of HSV-1 in a combined genome-scale screening approach has led to the identification of a regulatory axis in anti-viral innate immunity , and this important finding not only highlights the power of such combined genome-scale screening approaches to identify novel host candidates for anti-herpesvirus drug discovery , but provides an invaluable dataset to the herpesvirus and scientific community at large .
By nature of their scale , high-throughput screening technologies have limitations . RNAi technology is limited by technical issues such as off-target effects , where an alternative gene to the intended target is degraded , and insufficient gene knockdown . Similarly , Y2H protein interaction screens can generate both false-positive interactions , due to ‘sticky’ proteins and auto-activation of the reporter gene used , and false-negative interactions . Whereas the number of false positives can be considerably reduced by stringent screening and selection criteria , the low sensitivity of the Y2H assay , which detects 20–30% of known interactions , is inherent to the system and can only be marginally improved . This poor sensitivity is caused by factors such as structural restraints of the Y2H bait and prey fusion proteins , a lack of or existence of distinct protein modification in yeast cells , and cellular localization signals in bait and prey proteins preventing nuclear import [52] . However , as all other high-throughput methods for measuring binary protein interactions possess a similarly low sensitivity , but are considerably more laborious and expensive , the Y2H system is still the most commonly used technology [53] . In this study we have exploited a combined genome-wide screening approach to investigate HSV-1 replication and interaction with its host . This identified 358 functional HFs modulating HSV-1 replication , and 63 cellular interaction partners . In validation experiments , 57 . 8% of the interactions were confirmed by co-immunoprecipitation assays in mammalian cells , and of the 358 functional HFs identified in the siRNA screen , the phenotype of 83 . 3% was confirmed in deconvoluted siRNA experiments . This , combined with qPCR data demonstrating a minimum gene depletion of 69% , suggests that the functional phenotypes on virus replication are genuine , and not due to ‘off-target’ effects . The confirmation of such a high proportion of selected validation candidates , in spite of the potential technical drawbacks , highlights the reliability of our primary screen datasets , and thus provides an invaluable resource for the herpesvirology research community . One interesting outcome of this study was the surprisingly low overlap between hits identified using these different technologies . Of the 215 genes in common between the siRNA and cDNA libraries , only 10 ( 4 . 7% ) were classified as a hit by both methods . This , however , is not unexpected , as even the overlap between studies using the same technology has been reported to be low . For example , the overlap between the three previously published HIV screens was only 7% [19] . Furthermore , the degree of functional redundancy within the siRNA library , and cellular pathways in general , the potential situation-specificity of virus-host interactions , and the possibility of indirect interactions between viral and host proteins , suggest that these methodologies detecting functional outcomes or physical interactions are linked , but complementary rather than confirmatory . The identified HFs were enriched for a range of cellular processes , such as transcription , gene expression , protein transport and cell cycle ( Figure S2 and S3 ) , and involved at different stages of viral infection . We investigated HFs involved in capsid transport and ubiquitination of antiviral intrinsic host defence factors in more detail . Incoming HSV-1 capsids are transported to nuclear pores via the microtubule-organizing centre ( MTOC ) , mediated by capsid proteins VP26 ( UL35 ) and UL46 binding to the dynein light chains DYNLT1 ( Tctex1 ) and DYNLT3 ( rp3 ) [54] , [55] . Our Y2H screen confirmed the known interaction between the capsid protein VP26 and the dynein light chain DYNLT3 ( Text S1 and Figure S4 ) . Combined with the siRNA screen data , which found depletion of multiple light chain subunits had moderate anti-viral effects on HSV-1 , these data confirm propositions of redundancy in the capsid transport process which ensures successful infection in the event of viral mutations [38] , [39] , and provide further evidence that HSV-1 has evolved to be highly promiscuous in its exploitation of cellular pathways to its advantage . To overcome the intrinsic host defence , HSV-1 induces a proteasome-dependent degradation of anti-viral promyelocytic leukemia ( PML ) nuclear bodies ( ND10 domains ) by the RING-finger ubiquitin ligase ICP0 expressed during early infection [56] . In vitro , ICP0 is a biochemically active E3 ubiquitin ligase in the presence of E2 ubiquitin conjugating enzymes ( E2s ) [UBE2D1 ( UbcH5a ) and UBE2E1 ( UbcH6 ) ] [57] , but which E2s are used during infection has remained unclear . We identified 20 cellular E2s that are able to influence HSV-1 replication ( Text S1 and Figure S5 ) . Depletion of UBE2D1-4 , UBE2E1-3 and UBE2N significantly increased the number of PML-positive cells post-infection , in an ICP0-dependent manner , indicating that ICP0 can use multiple E2s to degrade PML [57] , [58] . One of the multi-protein complexes affecting HSV-1 replication was the Mediator complex , a large ( >30 subunits ) complex which links specific transcription factors to the RNA polymerase II transcription machinery [40] . As the requirement of Mediator subunits in the replication of herpes and other viruses is already well-known [7] , [9] , [11] , [41] , it was striking that depletion of the Med23 subunit exerted the opposite phenotype and led to a strong increase in virus growth . The Mediator is composed of four distinct modules termed the head , middle , tail and kinase domains , which provide the Mediator with some degree of active control over transcription [59] . As individual subunits of this large complex interact with and exert functional effects via specific transcription factors , it is not unexpected that the observed anti-viral effects were specific to Med23 [60] . Within the Mediator , Med23 forms a tight sub-complex with Med24 and Med16 [60] , [61] . The increase in virus replication observed upon depletion of Med24 may be caused by the destabilisation of the structure of this sub-complex ( Figure S6b ) . Investigations into the mechanism of action revealed Med23 inhibits HSV-1 replication by preferentially inducing a type III interferon response ( IFN-λ ) at the mRNA and protein level . This induction was mediated via a direct interaction with the transcription factor IRF7 , which resulted in a synergistic increase in IFN-λ expression . Med23 was unable , however , to further enhance IRF7-induced levels of IFN-β , suggesting an additional level of complexity to the regulation of interferon signalling . Interestingly , the inhibitory effect of Med23 was specific to HSV-1 , with replication of a range of other viruses including Vaccinia virus and Semliki Forest Virus being unaffected by Med23 depletion . As Vaccinia virus is resistant to IFN-λ anti-viral activity [62] , this observation further highlights the importance of IFN-λ , as opposed to IFN-β , in the anti-viral effect of Med23 . The R617Q mutation in Med23 ( R611Q in Med23 transcript variant 1 , used here ) was unable to enhance IRF7-induced IFN-λ expression . This mutation causes hereditary dementia [63] , and the failure to induce IFN-λ and thereby control HSV-1 in the brain may be a potential cofactor for the development of dementia , similar to Alzheimer's disease [64] . There is mounting evidence for a role of the IFN-λ family in the regulation of virus pathogenesis [45] , particularly in the case of Hepatitis C infection where a polymorphism in the promoter region of IFN-λ3 ( IL-28B; polymorphism rs12979860 ) , which correlates with plasma levels of IFN-λ3 [50] , is associated with disease and treatment outcome [49] . Individuals with recurrent HSV-1 reactivation have been shown to be deficient in IFN-λ expression [51] , and here we found the similar association between the IFN-λ3 promoter polymorphism and ethnically Italian patients suffering recurrent and severe reactivations of HSV-1-related oral herpes outbreaks , albeit with a small sample group ( n = 58 ) . Furthermore , sporadic mutations and genetic polymorphisms in innate immune receptor and signalling molecules that lead to the induction of type I and III IFNs have also been shown to be associated with Herpes Encephalitis [65] , as well as oral and genital Herpes [66] , [67] . HSV infection controlled by a complex , interconnected and highly regulated network of cytokines expressed by innate immune cells . Type I IFNs mainly produced by HSV-infected keratinocytes [68] and pDCs [69] inhibit the spread from neurons to epithelial cells and between epithelial cells [70] , similar to IFN-γ . Type III IFNs are also able to directly inhibit HSV-1 infection in primary neurons , astrocytes , macrophages and dendritic cells [71] , [72] . IFN-γ levels produced by peripheral blood CD4+ T-cells correlate with the frequency of HSV-1 reactivation [73] . IFN-λ is able to induce expression of both itself and the type I IFNs , and a similar effect has also been observed for type I IFNs which induce both type I and III IFNs [71] , [72] . Type III IFNs are mainly expressed by myeloid dendritic cells ( mDC ) and monocyte-derived macrophages [74] , and signal through the heterodimeric IL10RB/IL28RA receptor complex whose expression is largely restricted to cells of epithelial origin and plasmacytoid dendritic cells ( pDC ) , in contrast to the broadly expressed type I IFN receptor ( IFN-αR1/2 ) [75] , [76] . Since primary HSV-1 infection and reactivation affects skin and mucosa in the majority of cases , IFN-λ may play a much greater role in the control of HSV-1 pathogenesis , likely in a complex network of coregulated type I and II IFNS , than previously thought . We hypothesize that HSV-infected DCs at the site of the lesion ( such as skin Langerhans DCs whose role in IFN-λ production is currently unknown , or intruding myeloid DCs ) in individuals with the rs12979860 T/T or C/T haplotype express reduced levels of type III IFNs , and , in consequence , of type I IFNs , which leads to a reduced inhibition of local HSV-1 replication and the occurrence of fresh skin lesions . However , the relative contribution of IFN-λ1 and λ2/3 to the interferon-mediated control of HSV-1 replication in vivo , and indeed the role of Med23 in this , remains to be seen . In summary , this study provides a comprehensive and robust analysis of HFs that influence HSV-1 replication in vitro , which will benefit many future studies on HSV-1 . The identification of Med23 as a crucial cellular component for IFN-λ expression , and evidence for the significant role of type III IFN in the innate immune control of HSV-1 in vitro and in vivo , demonstrates the power of combined , genome-scale studies to identify physiologically important HFs for virus pathogenesis . Future studies will clarify the role of genetic variations in both Med23 and IFN-λ in HSV-1-related diseases , such as meningitis , keratitis and orolabial/genital reactivations .
siRNA SMARTpools ( 4 siRNAs per gene ) at 0 . 3 µM were dispensed in 10 µl volumes using a Rapidplate384 liquid handler ( Qiagen ) into triplicate black 384-well plates ( Corning ) , sealed with adhesive seals ( ThermoFisher ) and plastic lids . Plates were stored at −80°C until needed ( minimum 24 h , maximum 48 h ) . On the day of transfection , assay plates were thawed at room temperature and 10 µl transfection reagent ( Dharmafect 1 , Dharmacon ) , diluted in Hank's buffered saline solution ( HBSS , ThermoFisher ) to give a final concentration of 0 . 1% , was added using a Multidrop 384 ( ThermoFisher ) . Plates were incubated for 20 min at room temperature to allow formation of transfection complexes . During complex formation , low-passage ( p20–22 ) Hela cells ( ECACC ) from ∼50% confluent flasks were washed in PBS and trypsinised in Trypsin-EDTA ( Lonza ) before diluting in phenol red-free , antibiotic-free transfection medium ( DMEM/F-12 1∶1/5% FCS with 15 mM Hepes and L-glu; Gibco ) . Cells were counted and 3×103 cells in 40 µl were added to each well using the Multidrop 384 . Plates were incubated for 48 h at 37°C in a humidified incubator with 5% CO2 . To infect , media was removed from plates by inversion , and 10 µl media ( as for transfection , but containing penicillin-streptomycin; Lonza ) or virus ( HSV-1-eGFP strain C12 , diluted to MOI 0 . 5 in infection media ) [34] was added using the Multidrop 384 . Plates were incubated at 37°C for 1 h before 50 µl infection media was added and plates returned to the incubator . Replication was monitored as a function of eGFP fluorescence from 24 h to 80 h post-infection using the POLARstar OPTIMA plate reader ( BMG Labtech ) . Virus replication slopes over the linear phase were calculated and normalized to mock transfected wells on individual assay plates , and the mean replication slope from six replicates used for subsequent data analyses . Cells were transfected as described above , and the cytotoxicity of siRNAs was determined using the CellTiter Blue ( CTB , Promega ) reagent , which gives a fluorescent or absorbance signal relative to the number of live cells . Briefly , 5 µl CTB was added per well using the Multidrop 384 . Plates were incubated at 37°C in a humidified incubator with 5% CO2 for 2 h before measuring fluorescence ( POLARstar OPTIMA plate reader ) . Readings were normalized to viability of mock-transfected cells , per plate , and mean cell viability over three replicates was calculated . Distribution analysis of cell viability values identified median viability as 60% , and values <60% were considered cytotoxic . The HSV-1 clone collection was cloned by recombinatorial ( GATEWAY™ , Invitrogen ) and conventional cloning into the bait vector pGBKT7 , and screened against a library pooled from 12 , 381 MGC clones [35] in the pGADT7 prey vector using a semi-automated Y2H assay [77] . Interacting prey cDNAs were identified by sequential blasting of RefSeq , ENSEMBL and Unigene databases . BLAST hits with identical parameters ( score , expectation value , length of alignment ) were considered indistinguishable and counted separately . A high-confidence dataset was generated from interaction pairs isolated at least twice , or where the bait interacted with two highly related , non-promiscuous preys . Interactions between HSV-1 and human proteins were connected to a network of human protein-protein interactions ( a total of 62 , 310 ) taken from the databases HPRD [78] ( Release 9 ) , BioGRID [79] , DIP [80] , MINT [30] and IntAct ( downloaded May 18th 2010 ) . A high-confidence interaction set ( 9 , 829 interactions ) was compiled from interactions identified in at least two studies . Betweenness centrality ( g ( v ) of a protein v was calculated as g ( v ) = ∑s‡v‡t ( σst ( v ) /σst ) , where σst is the total number of shortest paths from protein s to protein t , and σst ( v ) is the number of those shortest paths that contain v . Betweenness centrality was normalized by dividing by the total number of protein pairs in the network . Enrichment for functional annotations from gene ontology ( GO ) [81] , KEGG [82] , [83] , REACTOME [84] , [85] , and BIOCARTA was performed using DAVID [86] . Data on known human protein complexes was retrieved from the CORUM database , and complexes with subunits showing consistently stronger effects ( inhibiting or enhancing ) than expected by chance were detected using Wilcoxon's rank-sum test . Genes included in the RNAi screen were ranked by their distance from the median knockdown , with the most inhibiting and enhancing genes being ranked highest . FDR was used for multiple testing correction . The HSV-1 replication phenotype observed in the primary screen was validated for a subset of candidates by deconvoluting the assay SMARTpools . The four individual siRNAs targeting different regions of each gene , as well as a reconstituted SMARTpool , were diluted to 0 . 3 µM in 1× siRNA buffer and dispensed to black 384-well plates . Transfection and infection was carried out as described above . Replication slopes were calculated and normalized as described , and a phenotype was considered validated if two or more of the four siRNAs resulted in the same , or better , phenotype . For inter-viral comparison , siRNAs were considered inhibitory or enhancing if normalized replication was ±2×STDEV of the controls Hela cells were transfected with selected SMARTpool siRNAs in 96-well plates , in triplicate , as described . After 48 h transfection , medium was removed , cells rinsed in PBS and lysed in 100 µl TRIZOL ( Invitrogen ) . Triplicate wells were combined , and RNA extracted by standard phenol:chloroform extraction methods . mRNA levels were determined by TaqMan qPCR , using the one-step RT-qPCR kit ( Thermofisher ) , with gene-specific primers ( Table S9 in Text S2 ) , and probes from the Universal Probe Library ( Roche ) . Expression levels normalized to the housekeeping cellular gene hypoxanthine phosphoribosyltransferase 1 ( HPRT ) and calibrated to mock-transfected cells . qPCR was carried out in duplicate for each sample , and the mean of normalized expression levels calculated . Proteins were transiently expressed in HEK293 cells as hybrid proteins with the Staphylococcus aureus protein A tag or Renilla reniformis luciferase fused to their amino termini . 20 ng of each expression construct were transfected into 1×104 HEK293 cells using 0 . 05 µl of lipofectamine 2000 ( Invitrogen ) in 96-well plates . After 40 h , medium was removed and cells were lysed on ice in 10 µl of ice-cold lysis buffer ( 20 mM Tris pH 7 . 5 , 250 mM NaCl , 1% TritonX-100 , 10 mM EDTA , 10 mM DTT , Protease Inhibitor Cocktail ( Roche ) , Phosphatase Inhibitor Cocktail ( Roche ) , Benzonase ( Novagen ) 25 units per µl final concentration ) containing sheep-anti-rabbit IgG-coated magnetic beads ( Invitrogen , Dynabeads M280 , 2 mg/ml final concentration ) . Lysates were incubated on ice for 15 minutes . 100 µl of wash buffer ( PBS , 1 mM DTT ) were added per well , and 10% of the diluted lysate was removed to determine the luciferase activity present in each sample before washing . The remaining sample was washed 6 times in wash buffer in a Tecan Hydroflex plate washer . Luciferase activity was measured in the lysate as well as in washed beads . Negative controls were wells transfected with the plasmid expressing the luciferase fusion protein and a vector expressing two copies of protein A . For each sample , four values were measured: the luciferase present in 10% of the sample before washing ( “input” ) , the luciferase activity present on the beads after washing ( “bound” ) , and the same values for the negative controls ( “input nc” , and “bound nc” ) . Normalized interaction signals were calculated as follows: Log ( bound ) /log ( input ) – log ( bound nc ) /log ( input nc ) . Normalized interaction signals were z-transformed by subtracting the mean and dividing by the standard deviation . The mean and standard deviation were calculated from large datasets of protein pairs which were not expected to interact , i . e . from negative reference sets . Selected siRNAs SMARTpools were diluted to 500 nM in HBSS and 40 µl was incubated with 40 µl Dharmafect 1 diluted in HBSS to a final concentration of 0 . 15% . After 20 min incubation , 3×104 Hela cells in 320 µl transfection medium were added , mixed with the transfection complexes and transferred to 8-well glass bottomed chamber slides ( Becton Dickinson ) . Plates were incubated for 48 h at 37°C in a humidified incubator with 5% CO2 before infection by removing medium and adding 100 µl HSV-1-eGFP at a MOI of 1 . After incubation for 1 h at 37°C , virus was removed and replaced with 500 µl growth medium . Images were acquired 48 h post-infection . Select siRNAs SMARTpools were diluted to 500 nM in HBSS and 100 µl was incubated with 100 µl Dharmafect 1 diluted in HBSS to a final concentration of 0 . 15% in individual wells of a 12-well plate . After 20 min incubation , 2×105 Hela cells in 800 µl transfection medium were added . Plates were incubated for 48 h at 37°C in a humidified incubator with 5% CO2 before infection by removing medium and adding 500 µl HSV-1-eGFP at a MOI of 1 . After incubation for 1 h at 37°C , virus was removed and replaced with 2 ml growth medium . After 48 h , medium was removed , cells rinsed in PBS and dislodged by trypsinisation . Cells were washed in PBS and pelleted by centrifugation for 10 min at 199 g . Supernatant was removed and cells fixed in 4% paraformaldehyde before analysing for eGFP expression by flow cytometry ( FACS DiVa , BD Biosciences ) using the CellQuest software package . For transient over-expression , 1 . 5×104 HEK293 cells were seeded in black 96-well plates . The following day , cells were transfected with 100 ng pCR3-Med23 using Lipofectamine™ LTX ( Invitrogen ) and incubated for 48 h before infection with the recombinant HSV-1 reporter viruses C12 and VP26-YFP at MOI 0 . 5 . Replication growth curves were monitored , and endpoint replication ( as determined by fluorescence ) was normalized to untransfected cells . For stable expression , Hela cells were transduced with pLenti-Med23 , generated using the ViraPower™ Lentiviral Expression System ( Invitrogen ) , as per manufacturers' instructions . Stable cells were infected , and replication monitored , as above . For confirmation of overexpression , RNA was extracted ( QIAGEN RNA-easy kit ) and expression quantified by one-step RT-qPCR ( Thermofisher ) , with gene-specific primers ( Table S9 in Text S2 ) , and probes from the Universal Probe Library ( Roche ) . Expression levels were normalized to the housekeeping cellular gene hypoxanthine phosphoribosyltransferase 1 ( HPRT ) and calibrated to mock-transfected cells . qPCR was carried out in duplicate for each sample , and normalized expression levels averaged . The VP26-YFP reporter virus was generated from a KOS BAC kindly provided by David Leib using the Red-mediated recombination system , where the first 4 amino acids of VP26 were replaced with YFP [91] , [92] , [93] . The effect of Med23 over-expression and depletion was investigated in interferon-deficient cells . The human alveolar epithelial cell line A549 , and A549-V , a Stat1-deficient derivative cell line stably expressing the V protein from Simian virus 5 [43] , were seeded at 2×104 cells per well in a 96-well plate , and transfected with Med23 siRNA or pCR3-Med23 and infected as described for Hela cells . A549 cells were transfected with 100 ng pCR3 or Med23 overexpression plasmids , or RSCF or Med23 SMARTpool siRNA ( 50 nM ) in duplicates , in 96-well plates as described . RNA was harvested 20 h post-transfection , and mRNA expression levels quantified by qRT-PCR , as described . Induction of IFN-λ by Med23 was determined in a range of cell types by seeding cells in 96-well plates to be ∼80% confluent the next day . Cells were transfected in duplicates with 100 ng pCR3 or pCR3-Med23 using Lipofectamine LTX with Plus reagent ( Invitrogen ) , in antibiotic-free medium . IFN-λ levels quantified 96–120 h post-transfection . The synergistic effect of Med23 and IRFs on IFN-λ induction was determined by co-transfection of pCR3 or pCR3-Med23 ( 50 ng ) with pCR3-IRF7 ( 50 ng ) in A549 cells , with Lipofectamine LTX with Plus reagent . The effect of Med23 depletion on IFN-λ induction was determined by transfection of A549 cells in 96-well plates with 50 nM RSCF or Med23 siRNA . IFN-λ was quantified 120 h post-transfection . IFN-λ protein expression was quantified in supernatants using a commercial IFN-λ DuoSet ELISA kit ( R and D Systems ) . A point mutation ( R611Q ) was introduced into Med23 ( Transcript Variant 1 ) by PCR with specific primers ( see Table S9 in Text S2 ) and the clone verified by sequence analysis . A549 cells were co-transfected with 60 ng of pCR3 , pCR3-Med23 or pCR3-R611Q , 60 ng of pCR3-IRF7 , 20 ng of IFN-β- , IFN-λ1- or ISRE-responsive luciferase reporter constructs and 10 ng pRL-TK , a Renilla Luciferase transfection control , in antibiotic-free low serum ( 1% ) medium . After 33 h cells were lysed , and firefly ( promoter reporter construct ) and Renilla ( transfection control ) luciferase activity was measured ( Dual-Luciferase Reporter Kit , Promega ) . Relative luminescence activity was normalized to Renilla ( as a transfection control ) . A549 cells were transfected in triplicate in black 96-well plates , as described . After incubation at 37°C for 24 h , cells were serum-starved for 24 h by growing in low serum ( 1% ) medium , before cells were left untreated or stimulated with 50 ng/ml IFN-α , IFN-β or IFN-γ , or 100 ng/ml IFN-λ1 or IFN-λ2/3 in serum-reduced medium . After 6 h cells were infected with HSV-1 C12 diluted to MOI 0 . 5 in IFN-containing serum-reduced medium . After 1 h incubation , virus was removed and media replaced with 100 µl serum-reduced growth medium containing no interferon , IFN-α/-β/-γ at 50 ng/ml , or IFN-λ1/-λ2/3 at 100 ng/ml . Replication was monitored as described and normalized to replication in mock-transfected , unstimulated cells . Potential interactions between Med23 and the IRFs were determined by yeast two-hybrid analysis and confirmed by co-immunoprecipitation ( Co-IP ) in mammalian cells . Ethnically Italian subjects with or without a history of recurrent Herpes labialis ( HL ) gave written voluntary informed consent and were enrolled in this study at the University of Rome Tor Vergata with the approval of the Ethical Committee at the University of Rome Tor Vergata . All subjects were interviewed by medically trained investigators using an appropriate questionnaire and agreed to provide saliva and/or blood samples . Data and blood sample collection was carried out as previously described [51] . A total of 58 healthy immunocompetent individuals ( 17 men and 41 women ) age 22–61 years ( overall median age 38 . 5 yr; male median age 40 yr , range 22–61; female median age 38 , range 23–60 ) participated in the study . None of the patients presented with an active lesion at the time of or in the 3 weeks preceding saliva sample collection . For the purpose of this study , patients were characterized into no recurrence of HL ( NR ) , low recurrence ( L; 1–3 HL episodes/yr , with a maximum extension of 1 cm , mild symptoms and healing time <7 days ) , high recurrence ( H; 4 or more HL episodes/yr , extension of lesions more than 1 cm ) , very high recurrence ( H+; more than 4 HL episodes/yr , extension of lesions >3 cm and/or involving nose or cheek beyond the lip , more severe and long lasting associated symptoms including itch , burning , paresthesias and/or neuralgia , with healing times >7 days and who required antiviral therapy ) . The NR group consisted of 21 individuals ( 16 women [median age ( range ) 36 ( 27–520] and 5 men [median age ( range ) 39 ( 26–49 ) ] . The L consisted of 18 individuals ( 12 women [median age ( range ) 37 years ( 23–60 ) ] and 6 men [median age ( range ) 43 years ( 22–61 ) ] ) . The H group consisted of 9 individuals ( 7 women [median age ( range ) 42 years ( 31–54 ) ] and 2 men [median age ( range ) 44 . 5 years ( 37–52 ) ] ) . The H+ group consisted of 8 individuals ( 5 women [median age ( range ) 38 years ( 28–55 ) ] and 3 men [median age ( range ) 35 years ( 35–46 ) ] ) . Saliva samples ( ∼3 ml ) were obtained from each subject after an overnight fast and after rinsing the mouth twice with water , split into two aliquots and frozen at −20°C . Samples were anonymised and stored with dual code labels before shipping on dry ice to the University of Edinburgh for DNA extraction . Saliva and PBMC samples were thawed and DNA extracted using a QIAamp DNA Blood Mini Kit ( Qiagen ) as per manufacturer's instructions , quantified using a NanoDrop and IL28B genotype determined by melt-curve analysis PCR on a LightCycler480 ( Roche ) using the LightMix® Kit IL28B ( TIB Molbiol ) as per manufacturer's instructions . Significance of genotype association was determined by Fisher's exact test , comparing the frequency of the CC , CT or TT genotype in the NR group versus the L ( p-value = 1 ) , H ( p-value = 0 . 12 ) or H+ ( p-value = 0 . 015 ) clinical groups . Below lists the GeneID numbers for genes and proteins mentioned within the text of this manuscript: IFITM1 ( 8519 ) ; IFN-α ( cluster; 3438 ) ; IFN-β ( 3456 ) ; IFN-γ ( 3458 ) ; IFN-λ1 ( 282618 ) ; IFN-λ2 ( 282616 ) ; IFN-λ3 ( 282617 ) ; IL10RB ( 3588 ) ; IL28RA ( 163702 ) ; IRF4 ( 3662 ) ; IRF7 ( 3665 ) ; IRF9 ( 10379 ) ; MED4 ( 29079 ) ; MED6 ( 10001 ) ; MED7 ( 9443 ) ; MED8 ( 112950 ) ; MED14 ( 9282 ) ; MED16 ( 10025 ) ; MED17 ( 9440 ) ; MED21 ( 9412 ) ; MED23 ( 9439 ) ; MED25 ( 81857 ) ; MED26 ( 9441 ) ; MED27 ( 9442 ) ; MED28 ( 80306 ) ; MED29 ( 55588 ) ; NR3C2 ( 4306 ) .
|
Herpes simplex virus type 1 ( HSV-1 ) infects the vast majority of the global population . Whilst most people experience the relatively mild symptoms of cold sores , some individuals suffer more serious diseases like viral meningitis and encephalitis . HSV-1 is also becoming more common as a cause of genital herpes , traditionally associated with HSV-2 infection . Co-infection with HSV-2 is a major contributor to HIV transmission , so a better understanding of HSV-1/HSV-2 disease has wide implications for global healthcare . After initial infection , all herpesviruses have the ability to remain dormant , and can awaken to cause a symptomatic infection at any stage . Whether the virus remains dormant or active is the result of a finely tuned balance between our immune system and evasion techniques developed by the virus . In this study we have found a new method by which the replication of the virus is counteracted . The cellular protein Med23 was found to actively induce an innate anti-viral immune response in the form of the Type III interferons ( IFN-lambda ) , by binding IRF7 , a key regulator of interferons , and modulating its activity . Interferon lambda is well known to be important in the control of Hepatitis C infection , and a genetic mutation correlating to an increase in interferon lambda levels is strongly linked to clearance of infection . Here we find the same association between this genetic mutation and the clinical severity of recurrent cases of HSV-1 infection ( coldsores ) . These data identify a Med23-interferon lambda regulatory axis of innate immunity , show that interferon lambda plays a significant role in HSV-1 infection , and contribute to the expanding evidence for interferon lambda in disease control .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"systems",
"biology",
"medicine",
"infectious",
"diseases",
"clinical",
"immunology",
"genetics",
"immunology",
"biology",
"genomics",
"molecular",
"cell",
"biology"
] |
2013
|
A Systematic Analysis of Host Factors Reveals a Med23-Interferon-λ Regulatory Axis against Herpes Simplex Virus Type 1 Replication
|
HIV-1 cell-to-cell transmission allows for 2–3 orders of magnitude more efficient viral spread than cell-free dissemination . The high local multiplicity of infection ( MOI ) observed at cell-cell contact sites may lower the efficacy of antiretroviral therapies ( ART ) . Here we test the efficacy of commonly used antiretroviral inhibitors against cell-to-cell and cell-free HIV-1 transmission . We demonstrate that , while some nucleoside-analog reverse transcriptase inhibitors ( NRTI ) are less effective against HIV-1 cell-to-cell transmission , most non-nucleoside-analog reverse transcriptase inhibitors ( NNRTI ) , entry inhibitors and protease inhibitors remain highly effective . Moreover , poor NRTIs become highly effective when applied in combinations explaining the effectiveness of ART in clinical settings . Investigating the underlying mechanism , we observe a strict correlation between the ability of individual drugs and combinations of drugs to interfere with HIV-1 cell-to-cell transmission , and their effectiveness against high viral MOIs . Our results suggest that the ability to suppress high viral MOI is a feature of effective ART regimens and this parameter should be considered when designing novel antiviral therapies .
Highly active antiretroviral therapy ( HAART ) has significantly reduced the mortality rate and has increased the life span of HIV-infected patients by maintaining viral loads below detection levels , thus preventing the onset of AIDS [1] , [2] , [3] , [4] . However , the presence of a stable latent reservoir , poor treatment adherence , and the emergence of drug-resistant HIV-1 variants continue to present challenges for successful treatments [5] . In order to develop more effective therapies , a detailed understanding of the pathogenesis of HIV-1 is necessary . Cell-to-cell transmission of HIV-1 has attracted significant attention as a potential factor influencing the pathogenesis of HIV-1 [6] , [7] . HIV-1 cell-to-cell transmission describes efficient virus spreading via sites of cell-cell contact through formation of virological or infectious synapses [7] , [8] . It provides for 2–3 orders of magnitude more efficient spread than cell-free virus dissemination and it is believed to be the main mode of viral spread in vitro [9] , [10] , [11] , [12] . The formation of virological synapses allows the coordination of viral assembly with viral entry at sites of cell-cell contacts [13] , [14] , [15] , [16] , [17] . These supramolecular structures permit the efficient transfer of large numbers of infectious particles to target cells resulting in a higher viral MOI than cell-free infection [18] , [19] , [20] , consistent with some in vivo observations [21] , [22] . This transfer of high viral MOI can also result in bystander death of CD4+ lymphocytes [23] . Primary cells may undergo pyroptosis and/or apoptosis in response to a high load of viral DNA in the cytoplasm and/or multiple viral integration events in the nucleus [24] , [25] , [26] . The cell death of highly infected cells may result in the positive selection of CD4+ T cells that carry a single provirus [27] , [28] . HIV-1 cell-to-cell transmission also allows HIV-1 to overcome barriers to infection and protects it from immunological and cellular restriction factors [11] , [20] , [29] , [30] , [31] . Finally , it has recently been reported that cell-to-cell transmission may protect HIV-1 from inhibition by antiretroviral therapies [32] . The transfer of large numbers of particles is thought to reduce the effective concentration of antiretroviral drugs within the cell and thus may provide a mechanism for the spread of HIV-1 in the presence of such therapies [32] , [33] . A reduced effectiveness of drugs during HIV-1 cell-to-cell transmission has been reported for tenofovir ( TFV ) , efavirenz ( EFV ) and zidovudine ( AZT ) [32] , [33] , [34] . However , these reports would seem to be in conflict with the clinical observation that HAART is successful at suppressing retroviral replication in millions of AIDS patients . In this study , we tested a panel of antiretroviral drugs that include nucleoside analog reverse transcriptase inhibitors ( NRTI ) , non-nucleoside analog reverse transcriptase inhibitors ( NNRTI ) , entry inhibitors ( Ent-I ) and protease inhibitors ( PI ) for their ability to inhibit HIV-1 cell-to-cell transmission . We found that while some NRTI drugs lost activity when virus was transferred by cell-to-cell transmission , NNRTIs , Ent-Is and PIs remained highly effective . Importantly , we regained potent antiretroviral activity upon combining NRTIs that were ineffective towards HIV-1 cell-to-cell transmission as single therapies . These results explain the effectiveness of antiretroviral combination therapies in clinical settings . Finally , we demonstrate that the effectiveness of ART against HIV-1 cell-to-cell transmission can be recapitulated by testing their effectiveness against high viral MOI . Altogether , our results suggest that the ability to suppress high viral MOI is a defining feature of effective ART regimens and provides a valuable tool to develop novel ART that remain effective against HIV-1 cell-to-cell transmission .
We applied these experimental conditions to systematically test the efficacy of 6 NRTIs , 4 NNRTIs , 4 Ent-Is and 4 PIs against cell-free and cell-to-cell HIV-1 transmission . The NRTI inhibitors TFV , AZT , and stavudine ( d4T ) were profoundly impaired in their ability to interfere with HIV-1 cell-to-cell transmission to primary human CD4+ T cells ( Fig . 2A , Supplementary Fig . S2 ) . Their dose-response curves were right-shifted indicating that ∼200–1000-fold higher drug concentrations were required to interfere with HIV-1 cell-to-cell transmission as compared to cell-free HIV-1 . This observation is consistent with previous observations for TFV and AZT [32] , [34] and translates into poor HIV-1 inhibition at the active drug concentrations detected in the serum of treated patients ( Fig . 2A , gray bar ) . Interestingly , the NRTI inhibitors lamivudine ( 3TC ) , abacavir ( ABC ) and emtricitabine ( FTC ) showed a narrowing of cell-free and cell-to-cell transmission dose-response curves indicating an increased ability to interfere with HIV-1 cell-to-cell transmission relative to other NRTIs ( Fig . 2A and Supplementary Fig . S2 ) . Importantly , most NNRTIs ( nevirapine ( NVP ) , etravirine ( ETR ) and efavirenz ( EFV ) ) interfered with HIV-1 cell-to-cell transmission as efficiently as with cell-free transmission . The Ent-Is enfurvitide ( T20 ) , plerixafor ( AMD3100 ) , and BMS488043 were also very effective consistent with previous results for T20 [30] . Rilpivirine ( RPV ) and BMS626529 exhibited intermediate effects ( Supplementary Fig . S2 ) . The PIs indinavir ( IDV ) , darunavir ( DRV ) , lopinavir ( LPV ) and saquinavir ( SQV ) also retained their effectiveness regardless of the mode of transmission ( Fig . 2A and Supplementary Fig . S3B ) , consistent with recent observations [36] . The effectiveness of most NNRTIs , Ent-Is and PIs is clearly visible when the fold change in the IC90 during cell-to-cell transmission versus cell-free HIV-1 transmission is plotted for each drug ( Fig . 2B ) . The effects could not be attributed to drug toxicity ( Supplementary Fig . S4 ) . A similar pattern was observed for a more physiologically relevant founder virus HIV-1TROJ . c ( Fig . 2C and Supplementary Fig . S5 ) [37] . Cell-to-cell transmission of HIV-1TROJ . c was more resistant to TFV and AZT , albeit to a lesser extent than HIV-1NL4-3 , and remained highly sensitive to NNRTIs , Ent-Is and PIs . To gain a better understanding of the effectiveness of antiretroviral inhibitors in both modes of HIV-1 transmission , we calculated the instantaneous inhibitory potential ( IIP ) [38] , [39] . The IIP incorporates both the IC50 and the slope of the inhibition curve and may provide a more accurate assessment of the effectiveness of an inhibitor . We found that the IIP in co-culture samples was dramatically weakened for TFV and AZT and significantly reduced for most other NRTIs ( Fig . 3A , B and Supplementary Fig . S6 ) . Importantly , the IIP was not affected for most NNRTIs and Ent-Is in agreement with the observations based on IC90 . All data is summarized as the ratio of the IIP at the top drug dose ( ICMax ) for co-culture over cell-free in Figures 3B and C . All curves are shown in Supplementary Fig . S6 and S7 . The IIP could not be computed for PIs because of the limited dynamic range in the signal for HIV-1 cell-to-cell transmission ( data now shown ) . These data demonstrate that while some antiretroviral drugs such as NRTIs are less efficient against HIV-1 cell-to-cell transmission , most NNRTIs and Ent-Is remain highly effective regardless of the mode of viral transmission . The failure of antiretroviral inhibitors such as TFV and AZT to interfere with HIV-1 cell-to-cell transmission stands in conflict with the clinical experience that they are effective in suppressing HIV-1 replication in AIDS patients [1] , [2] , [3] , [4] . However , mono-therapy is not used for the treatment HIV-1-infected patients due to the high risk of emergence of drug-resistant mutants [40] , [41] . Thus , we wondered whether drugs that fail to interfere with cell-to-cell transmission when used individually , are more effective when used in combination . To test drug combinations , we matched drug concentrations according to their IC90 values and treated co-culture and cell-free infections with serially diluted drug combinations . Strikingly , the combination of AZT and TFV potently interfered with HIV-1 cell-to-cell transmission ( Fig . 4A ) . While each drug individually was ∼200–1000-fold less effective against HIV-1 cell-to-cell transmission , this difference was reduced to ∼4 . 1-fold when the drugs were combined ( Fig . 4A ) . Furthermore , the drug combination shifted the effective dose-range required to suppress HIV-1 cell-to-cell transmission to within the drug concentrations detected in the serum of treated AIDS patients ( Fig . 4A , gray bar ) . This observation was reproduced for three additional combinations of NRTIs including the clinically used combinations of 3TC/ABC and 3TC/AZT ( Fig . 4B , Supplementary Fig . S8A ) [42] . The increased effectiveness of combination therapy was also visible when the IC90 values were compared and the IIP was calculated ( Fig . 4C , D and Supplementary Fig . S8B ) . The effectiveness of combination therapies was surprising since drug combinations at most doubled the total drug concentration . If the effectiveness of competitive NRTI inhibitors was reduced due to a high MOI at sites of cell-cell contact [32] , then doubling the drug concentration should be insufficient to inhibit all the incoming particles ( Fig . 2 ) . The observation of synergy in NRTI combination therapies can likely be explained by more efficient inhibition of reverse transcriptase . During reverse transcription , reverse transcriptase is able to excise an incorporated nucleotide analog , thus lowering the potential effectiveness of many NRTIs [43] , [44] , [45] . Combinations of nucleotide analogs have been observed to interfere with this excision process , thus enhancing the ability of NRTIs to terminate the growing DNA chain [46] . To test this hypothesis , we conducted our co-culture and cell-free inoculations using an HIV-1NL4-3 clone carrying the M184V mutation in RT . This mutation renders HIV-1 reverse transcriptase hypersensitive to AZT due to its inability to excise the drug [47] , [48] . We predicted that AZT would efficiently interfere with HIV-1 cell-to-cell transmission of HIV-1 carrying M184V mutant RT . Indeed , the difference in IC90 between cell-free and co-culture infection was dramatically reduced compared to HIV-1 carrying wild-type RT ( Fig . 4E ) . These results suggest that synergy between NRTIs against HIV-1 cell-to-cell transmission is , at least in part , due to a reduction of NRTI excision , which in turn causes more efficient chain termination . Next , we asked how this drug-resistant HIV-1 mutant would behave during combination therapies in both modes of transmission . The M184V mutation was first characterized as a mutation that provides resistance against 3TC [49] , [50] . We hypothesized that if this mutant were to be exposed to a combination of 3TC and TFV , it may be able to resist inhibition by TFV by cell-to-cell transmission . We found that if HIV-1 is resistant to one of the inhibitors used in the combination , the dose-response curve for cell-to-cell transmission was shifted again towards higher drug concentrations , phenocopying the behavior of NRTI mono-therapy ( Supplementary Fig . S9 ) . This suggests that drug-resistant HIV-1 mutants may gain a replicative advantage to amplify by cell-to-cell transmission in the presence of some combination therapies . It has been suggested that the high local MOI observed at sites of cell-cell contact is responsible for the relative resistance of HIV-1 cell-to-cell transmission to antiretroviral inhibitors [32] , [33] . This would suggest that the reason why most NNRTIs and all combination therapies are effective against HIV-1 cell-to-cell transmission is because they are MOI-independent , thus would remain effective despite high viral MOI . To test this hypothesis , we concentrated HIV-1NL4-3 ( GLuc ) and used highly susceptible MT4 cells , which allowed us to use MOIs of up to 25 . An MOI of 25 is close to the highest MOI that can be detected during HIV-1 cell-to-cell transmission [18] , [19] , [20] . We found that 3TC , TFV , FTC and AZT were indeed overpowered by increasing particle numbers ( Fig . 5A , B ) . In other words , higher drug concentrations were required for these NRTIs to inhibit high MOIs . In striking contrast , NNRTIs and combination therapies were largely MOI-independent ( Fig . 5A , C ) . The same drug concentration of NVP or the combination of AZT and TFV inhibited HIV-1 irrespective of the MOI . The strong correlation between non-effectiveness or effectiveness of ART against HIV-1 cell-to-cell transmission and high MOI was best seen when the change in IC90 during co-culture infection was plotted versus the change in IC90 during high MOI ( Fig . 5D ) . This plot shows the clustering of MOI-dependent and MOI-independent treatments . Thus , we predict that those individual and combination therapies that are effective against high MOI will also efficiently interfere with HIV-1 cell-to-cell transmission .
The recent questioning of ART's effectiveness during HIV-1 cell-to-cell transmission [32] stood in conflict with the clinical experience that HAART is effective at suppressing HIV-1 replication in patients [1] , [2] , [3] , [4] . Many clinicians may have concluded that HIV-1 cell-to-cell transmission cannot be relevant in patients and that cell-free spread must dominate . Here we showed that this interpretation is likely incorrect . Rather , we demonstrate that clinically applied ART regimens are effective against HIV-1 cell-to-cell transmission likely because they also remain effective against the high number of particles transferred at sites of cell-cell contacts . By systematically testing the efficacy of commonly used antiretroviral inhibitors against cell-to-cell and cell-free HIV-1 transmission , we demonstrate that while some NRTIs are indeed less effective against HIV-1 cell-to-cell transmission , most NNRTIs , Ent-Is and PIs remain highly effective . Importantly , upon combining of 2 NRTIs that failed as single therapies , HIV-1 cell-to-cell transmission and cell-free infection often became equally inhibited . Therefore , our findings indicate that the ability of HIV-1 cell-to-cell transmission to evade antiretroviral drug inhibition is not a universal phenomenon . Because standard treatment involves the combination of several drugs ( 2 NRTI+1 NNRTI or PI ) it would seem unlikely that HIV-1 cell-to-cell transmission would provide a feasible mechanism for any ongoing viral replication in the presence of suppressive treatment . This observation is consistent with a large body of evidence indicating that suppressive HAART stops any measurable level of viral replication [51] . Our observations that combination therapies of NRTIs can be effective against HIV-1 cell-to-cell transmission indicates that the clinical effectiveness of HAART did not automatically imply that HIV-1 spreads by cell-free virus in patients . Rather we demonstrate that HAART effectively suppresses the high MOI observed during HIV-1 cell-to-cell transmission . The determination of the exact mechanism of HIV-1 cell-to-cell spread in vivo will require the direct in vivo visualization of viral dissemination [52] , [53] . However , our results already provide evidence that HIV-1 cell-to-cell transmission can contribute to the pathogenesis of HIV-1 as a feasible mechanism of viral escape during drug mono-therapy or inadequate treatment regimens . We confirmed the original observation that some NRTIs fail to restrict HIV-1 cell-to-cell transmission during mono-therapy [32] . We also provide evidence that drug-resistant virus may gain a replicative advantage to spread by HIV-1 cell-to-cell transmission in the presence of inadequate combination therapy . Thus , HIV-1 cell-to-cell transmission may contribute to the rise of drug-resistant virus and therapy failure under conditions of poor adherence [54] . Our finding that ART similarly suppresses high viral MOIs and HIV-1 cell-to-cell transmission is consistent with the suggestion that a high viral MOI is a central feature associated with cell-cell contact mediated viral dissemination [18] , [19] , [20] , [32] . High MOIs have been observed in infected cells in tissues in vivo [21] , [22] . This observation appears to be in conflict with the finding that most circulating T cell lymphocytes carry only a single provirus [27] , [28] . However , a high MOI may often result in bystander death of CD4+ lymphocytes , a hallmark of AIDS pathogenesis [23] . Primary cells have been suggested to innately sense the presence of a large number of viral DNA copies ( unintegrated and/or integrated ) and undergo apoptosis and/or pyroptosis [24] , [25] , [26] . The cell death of highly infected cells may result in the positive selection of CD4+ T cells that carry a single provirus [27] , [28] . The ability of ART to suppress the high viral MOI documented in this report confirms the long standing knowledge that effective ART is able to effectively suppress bystander cell death and protect most AIDS patients from further T cell depletion [4] , [55] . A high local MOI of reverse transcriptase can overwhelm drug activity by mass action [32] , [33] . However , the ability of multiple drugs , particularly NNRTIs , to remain effective against the high local MOI observed during HIV-1 cell-to-cell transmission suggests that mass action alone cannot fully explain the mechanism by which antiretroviral inhibitors function under these conditions . In the case of NRTIs , our data suggest that the ability of reverse transcriptase to excise nucleotide analogs plays an important role in this phenomenon . When nucleotide excision was inhibited through mutation of the RT , mono-therapy with a nucleotide analog can inhibit both modes of viral transmission with similar efficiency . Similarly , we observed synergy in combination therapies consistent with more efficient reverse transcript chain termination and less efficient nucleotide analogue excision by RT [46] . In the case of NNRTIs , allosteric inhibition of RT also provides for synergistic effects [46] . Moreover , we hypothesize that other steps in the cellular uptake , metabolism , or secondary binding sites , determine the effective dosage of antiretroviral inhibitors . Said differently , under conditions of high MOI encountered during cell-to-cell transmission , interaction of the drug with RT is not the rate-limiting step for efficient inhibition of reverse transcription . That is , the number of incoming RT molecules alone does not define the effective dosages of drug . These considerations indicate that there is likely no single mechanism that explains whether a drug or drug combination is effective against HIV-1 cell-to-cell transmission . Thus , each drug and drug combination needs to be tested . To this day , therapy outcome in patients has been difficult to predict . Mathematical models have been developed recently that incorporate drug IC50 , and the slopes of inhibition curves as in the IIP , as well as viral fitness , mutations and treatment adherence [56] , [57] . Our data indicate that the effectiveness of ART against HIV-1 cell-to-cell transmission and viral MOI are additional helpful parameters to predict drug efficacy . Moreover , we observed that all drugs effective against HIV-1 cell-to-cell transmission were effective because they are MOI-independent and can efficiently suppress the high local MOI at virological synapses . These data suggest that highly effective drug regimens , either single or in combination therapies , must exhibit MOI-independence . Testing the effectiveness of antiretroviral inhibitors against increasing MOI provides a simple assay and a valuable tool for screening existing and novel individual drugs and combination therapies prior to clinical testing .
All the cells used in this study were anonymized and were obtained from commercially available sources ( ATCC , AIDS Research and Reagents Program , New York Blood Center ) . As such , these samples are exempt from IRB review . Peripheral blood mononuclear cells were purified from blood enriched by leukapheresis ( New York Blood Center ) with the Ficoll-Paque Plus gradient ( GE Healthcare Life Sciences ) . Following this purification step , CD4+ T cells were purified using the EasySep Human CD4+ T Cell Enrichment Kit ( StemCell Technologies ) and were stimulated with PHA ( 10 µg/mL ) ( Sigma-Aldrich ) , IL-2 ( 100 U/mL ) , and IL-7 ( 100 ng/mL ) for 72 hr ( cytokines from Miltenyi Biotec ) at 37°C . After stimulation , cells were maintained in RPMI ( Gibco ) supplemented with 100 U/mL penicillin/streptomycin ( Gibco ) , 2 mM of L-glutamine ( Gibco ) , 10% FBS ( Gibco ) , IL-2 ( 100 U/mL ) , and IL-7 ( 100 ng/mL ) at 37°C . A subclone of Jurkat-inGLuc was selected from the population described by Zhong , et al . [20] . The cell lines Jurkat-inGLuc , MT4 ( NIH AIDS Research and Reagents Program ) , and HEK293 ( ATCC ) were maintained in RPMI supplemented with 100 U/mL penicillin/streptomycin , 2 mM of L-glutamine , and 10% FBS at 37°C . TZMbl cells were obtained from the NIH Research and Reagents Program and were maintained in DMEM supplemented with 100 U/mL penicillin/streptomycin , 2 mM of L-glutamine , and 10% FBS at 37°C . The plasmid encoding the intron-regulated HIV-based Gaussia luciferase pUCHR-inGLuc ( HIVinGLuc ) was kindly donated by Gisela Heidecker , National Cancer Institute . The plasmid encoding the HIV-1 molecular clones NL4-3 [58] and pTRJO . c [37] were obtained from the AIDS Research and Reagents Program . The plasmid encoding the M184V mutation in reverse transcriptase ( pNL4-3ΔEnv ( M184V ) ) was kindly donated by Robert Siliciano , Johns Hopkins University . To generate a wild type version of the M184V mutant , the construct was digested with PspOMI and AgeI ( New England Biolabs ) . The ∼1 . 5 kb fragment generated was then ligated to the ∼13 kb fragment of wild type NL4-3 after digestion with the same enzymes . The plasmid encoding the vesicular stomatitis virus G-glycoprotein ( VSV-G ) was obtained from Michael Marks , University of Pennsylvania . Most antiretroviral drugs tested in this study were obtained from the AIDS Research and Reagents Program . The attachment inhibitors BMS488043 and BMS626529 were donated by Mark Krystal ( Bristol-Myers Squibb ) [59] , [60] , [61] . HIV-1 pseudotyped with VSV-G was generated by co-transfecting HEK293 cells with pVSV-G and pNL4-3 or pTRJO . c at a ratio of 1∶10 . HIVGLuc was generated by co-transfecting HEK293 cells with pNL4-3 ( or pTRJO . c ) and pHIVinGLuc at a ratio of 6∶1 or 10∶1 . For inoculations of MT4 cells , HIVGLuc was generated by inoculating HEK293 cells stably carrying HIVinGLuc and collecting culture supernatant at 36 and 60 hr post-infection . Viral supernatants were concentrated using Lenti-X Concentrator ( Clontech ) or by ultracentrifugation ( ∼20 , 000×g ) over a 20% sucrose ( in PBS ) cushion for 2 hr at 4°C . Primary CD4+ T cells were incubated with serial dilutions of nucleoside analogs at 37°C for 16–24 hr prior to inoculation in a total of 1% DMSO . This is required for the accumulation of sufficient concentrations of active inhibitors within the cells . Cells were incubated at 37°C with non-nucleoside analogs and entry inhibitors for 2 hr prior to inoculation also in a total of 1% DMSO . Cell-free inoculations were conducted by spinoculating 105 primary CD4+ T cells in 96-well plates at 1 , 200×g and at room temperature for 2 hr with 50 µL of concentrated HIVGLuc [62] . Cultures were then incubated at 37°C for 36–40 hr . Co-cultures were conducted by first spinoculating Jurkat-inGLuc cells with full length HIV-1NL4-3 pseudotyped with VSV-G at 1 , 200×g and at room temperature for 2 hr . The Jurkat-inGLuc clone was originally selected to be CD4-low cells to minimize donor-to-donor infection in co-culture experiments with target primary CD4+ T cells . Cells were then washed , stimulated with 6 . 25 ng/mL of PMA for 2 hr at 37°C , washed and incubated in fresh medium for 18 hr at 37°C . A brief PMA treatment was used to stimulate expression of latent HIVin-GLuc for efficient packaging by the incoming wild type HIV . Additionally , PMA treatment causes down-regulation of CD4 expression in the donor Jurkat-inGLuc cells , further preventing donor-to-donor infection [63] . Subsequently , PMA was removed from the culture so that target primary CD4+ T cells were never exposed to the drug . 105 infected Jurkat-inGLuc cells were then washed and co-cultured with 105 primary CD4+ T cells in a total of 50 µL . GLuc accumulated in the culture supernatant was detected using the BioLux Gaussia Luciferase Assay Kit ( New England Biolabs ) and a Berthold Technologies luminometer . To test PIs , this protocol had to be modified to account for the activity of this drug class within the HIV-1 donor cell . To do this , HIV-1 infected Jurkat-inGLuc cells were incubated with increasing concentrations of PIs immediately following stimulation with PMA for 12 hr prior to co-culturing with primary cells ( see Supplementary Fig . S3A ) . Co-cultures were incubated for 42 hr prior to measuring GLuc . To assess the effect of protease inhibitors on the infectivity of cell-free particles , we collected the supernatant of donor cells cultured alone in the presence of PIs 54 hr after exposure to the inhibitors . This supernatant corresponds to the total number of particles released during the co-culture . The supernatant was tittered on 105 target primary CD4+ T cells or on 2×104 TZMbl target cells at a total volume of 60 µL in 96-well plates , spinoculated and incubated at 37°C for 36 hr prior to measuring GLuc activity . TZMbl cells were used to assess the infectivity of the supernatant because they are much more susceptible to cell-free HIV-1NL4-3 than primary CD4 T cells and could detect very low titers of HIV-1NL4-3 produced by donor cells . Prior to infection , target cells were stained with 1 µM of Cell Proliferation Dye eFluor 670 ( eBioscience ) in OptiMEM medium ( Gibco ) at 37°C for 20 min . Cells were washed and incubated in complete medium supplemented with cytokines at 37°C for 30 min , washed and prepared for drug treatment . 24 hr after infection , cultures were harvested and fixed in 100 µL of BD CytoFix/CytoPerm buffer ( BD Biosciences ) for at least 30 min at 4°C . The cells were then washed with BD Perm/Wash buffer ( BD Biosciences ) and stained for 30 min at 4°C in 100 µL of BD Perm/Wash buffer containing the anti-HIV-1 Gag antibody clone KC57 ( Beckman Coulter ) . The cells were washed with BD Perm/Wash buffer , resuspended in PBS supplemented with 0 . 5% BSA and 2 mM of EDTA and analyzed by flow cytometry with a FACSCalibur ( BD Biosciences ) . The same staining protocol was used for sorting HIV-1-positive target cells after cell-free or cell-to-cell transmission . The sort was conducted using a BD FACSAria sorter . Following the sort , cells were spun , resuspended in 200 µL of PBS +200 µL of Buffer AL ( Qiagen ) +20 µL of Proteinase K ( Qiagen ) and incubated at 60°C for 24 h to remove paraformaldehyde . DNA was purified using the DNeasy Blood and Tissue Kit ( Qiagen ) . HIV-1 integration was measured by Alu-PCR as previously described using 2 . 5 U of Platinum Taq ( Life Technologies ) [64] . 36 hr post-infection , a sample of 10 µL of culture was collected for each drug treatment condition and mixed with 10 µL of CellTiter-Glo ( Promega ) . Cells were incubated at 37°C for 10 min and the luciferase signal was measured using a Berthold Technologies luminometer . Inhibitor IC90 and IIP were calculated using MATLAB software . Statistical tests were calculated using Minitab software .
|
HIV-1 cell-to-cell transmission has gained interest due to its potential role in AIDS pathogenesis . It has recently been suggested that antiretroviral therapies fail during cell-to-cell transmission because of the high number of particles transferred at sites of cell-cell contacts . However , these findings stand in contrast with the clinical observation that ART is successful in suppressing retroviral replication in HIV-positive patients . Consequently , many interpreted this observation to suggest that HIV-1 cell-to-cell transmission is not clinically relevant . Here we show that this interpretation is likely incorrect . By systematically testing the efficacy of commonly used antiretroviral inhibitors against cell-to-cell and cell-free HIV-1 transmission , we demonstrate that , while some NRTIs are less effective , most NNRTIs , entry inhibitors and protease inhibitors remain highly effective . Moreover , NRTIs become highly effective when combined , thus supporting the known effectiveness of HAART in clinical settings . Interestingly , the ability of individual drugs and combinations to interfere with HIV-1 cell-to-cell transmission correlates with their effectiveness against high viral MOIs . Our results suggest that the ability to suppress the high viral MOI during HIV-1 cell-to-cell transmission is a critical feature of existing ART regimens that should be tested when designing novel antiviral therapies .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immunodeficiency",
"viruses",
"medicine",
"infectious",
"diseases",
"mechanisms",
"of",
"resistance",
"and",
"susceptibility",
"viral",
"transmission",
"and",
"infection",
"hiv",
"virology",
"retrovirology",
"and",
"hiv",
"immunopathogenesis",
"biology",
"antivirals",
"microbiology",
"viral",
"replication",
"viral",
"diseases"
] |
2014
|
Highly Active Antiretroviral Therapies Are Effective against HIV-1 Cell-to-Cell Transmission
|
The reproductive ground plan hypothesis ( RGPH ) proposes that the physiological pathways regulating reproduction were co-opted to regulate worker division of labor . Support for this hypothesis in honeybees is provided by studies demonstrating that the reproductive potential of workers , assessed by the levels of vitellogenin ( Vg ) , is linked to task performance . Interestingly , contrary to honeybees that have a single Vg ortholog and potentially fertile nurses , the genome of the harvester ant Pogonomyrmex barbatus harbors two Vg genes ( Pb_Vg1 and Pb_Vg2 ) and nurses produce infertile trophic eggs . P . barbatus , thus , provides a unique model to investigate whether Vg duplication in ants was followed by subfunctionalization to acquire reproductive and non-reproductive functions and whether Vg reproductive function was co-opted to regulate behavior in sterile workers . To investigate these questions , we compared the expression patterns of P . barbatus Vg genes and analyzed the phylogenetic relationships and molecular evolution of Vg genes in ants . qRT-PCRs revealed that Pb_Vg1 is more highly expressed in queens compared to workers and in nurses compared to foragers . By contrast , the level of expression of Pb_Vg2 was higher in foragers than in nurses and queens . Phylogenetic analyses show that a first duplication of the ancestral Vg gene occurred after the divergence between the poneroid and formicoid clades and subsequent duplications occurred in the lineages leading to Solenopsis invicta , Linepithema humile and Acromyrmex echinatior . The initial duplication resulted in two Vg gene subfamilies preferentially expressed in queens and nurses ( subfamily A ) or in foraging workers ( subfamily B ) . Finally , molecular evolution analyses show that the subfamily A experienced positive selection , while the subfamily B showed overall relaxation of purifying selection . Our results suggest that in P . barbatus the Vg gene underwent subfunctionalization after duplication to acquire caste- and behavior- specific expression associated with reproductive and non-reproductive functions , supporting the validity of the RGPH in ants .
Division of labor is the cornerstone of insect societies and implies the coexistence of individuals that differ in morphology , reproduction and behavior [1] , [2] . There are usually two levels of division of labor among individuals in social insect colonies . The first relates to a reproductive division of labor , whereby reproduction is monopolized by one or several queens while sterile workers perform all the tasks related to colony maintenance . The second relates to a division of labor among the worker force , whereby workers perform different tasks in an age-dependent sequence: young workers usually perform tasks inside the colony ( e . g . brood care ) while old workers forage outside the nest [3] , [4] . The colony organization of advanced eusocial insects evolved independently in ants , bees , and wasps [5] , [6] . While the ecological constraints favoring social evolution are well studied [7] , it remains largely unknown whether the genetic mechanisms regulating behavior are conserved among species [8]–[12] . The ovarian ground plan hypothesis ( OGPH ) is a theoretical framework that seeks to explain the evolution of reproductive division of labor in social insects [13] , [14] . The OGPH proposes that the physiological pathways regulating the reproductive and behavioral cycles of solitary ancestors have been co-opted and selected to evolve into the queen and worker castes of existing eusocial insects . This hypothesis is based on the observations that the ovarian cycle of alternate development and depletion phases of solitary wasps is associated with reproductive and non-reproductive behavioral traits that resemble the queen and worker castes of eusocial insects: females with developed ovaries lay eggs while females with undeveloped ovaries forage for food and defend the nest [13] . The same was likely true in the solitary ancestors of ants and bees . The reproductive ground plan hypothesis ( RGPH ) extends this concept to explain the evolution of worker division of labor in honeybees [15]–[17] . Indeed , honeybee worker subcastes have two distinctive phases of ovarian activity , with nurses having large ovaries and high titers of the yolk protein vitellogenin ( Vg ) , and foragers small ovaries and low titers of Vg [15] , [18] , [19] . The RGPH suggests that the mechanisms controlling ovarian activity influence the behavioral development and the mechanisms of food collection in worker honeybees . Support for this hypothesis is provided from studies demonstrating that workers with larger ovaries and higher titers of Vg are more likely to forage at younger ages and show a pollen foraging bias compared to workers with smaller ovaries and lower titers of Vg , which are more likely to forage at older ages and show a nectar foraging bias [15] , [16] . These variations in reproductive traits have a genetically inherited component as strains with different ovarian activity and foraging bias have been selected from wild type populations [15] , [20] . Although it has been established that the pleiotropic mechanisms connecting reproduction and division of labor have a genetic component , three lines of evidence suggest that the two processes are linked by an additional nutritional factor . First , in honeybees , reproductive queens and potentially reproductive nurses ( with large and medium-sized ovaries , respectively ) [18] , [21] , [22] consume diets with higher protein content [23] compared to sterile foragers with smaller and presumably non-functional ovaries [18] . Second , pollen consumption in nurses [23] is associated with higher Vg protein levels [19] , compared to foragers that only consume honey [23] . Finally , there is a causal relationship between nutrition , Vg levels and behavior as pollen consumption is required to induce Vg expression [24] , [25] and experimental reduction of Vg expression results in precocious foraging [26] , [27] . To determine whether the co-option of reproductive pathways plays a major role in social evolution would require to investigate the link between reproductive physiology and behavior in other social insects , preferentially in those , such as ants , that evolved sociality independently from bees [5] , [6] , [28] . Ants have two additional characteristics that make them a particularly interesting model to study the predictions of the RPGH . First , in contrast to the honeybee genome that contains a single Vg gene , ant genomes can harbor multiple Vg genes . Indeed , the genome of the fire ant Solenopsis invicta harbors four Vg genes , two of them preferentially expressed in queens ( Si_Vg2 and Si_Vg3 ) and the two others in foraging workers ( Si_Vg1 and Si_Vg4 ) [29] . Vg duplication and subsequent subfunctionalization to acquire caste-specific expression provides a unique opportunity to test whether the genes associated with reproduction were co-opted to regulate worker behavior . Second , also in contrast with bees and wasps , where workers are facultatively sterile , workers are fully sterile in a significant number of ant species , including P . barbatus and S . invicta [30]–[32] , which allows one to test the hypothesis that Vg reproductive function was co-opted to regulate behavior in species with fully sterile workers . The first aim of this study was to determine the number of Vg genes in P . barbatus and other ants and investigate their phylogenetic relationships . This analysis is expected to determine whether the occurrence of multiple Vg genes is a phenomenon specific to S . invicta [29] or shared with other ant species as well as provide information on the origin and evolution of Vg genes in ants . Our second objective was to test whether Vg genes in P . barbatus display caste-specific expression profiles similar to that observed in S . invicta , which will address the question whether Vg gene duplication and subfunctionalization to acquire caste-specific functions is a common feature in ant species . Finally , the third objective of this study was to investigate whether the expression of Vg genes in P . barbatus is associated with task performance as predicted by the RGPH . We carried out these objectives by annotating Vg genes , building a phylogenetic tree , measuring mRNA levels of P . barbatus Vg genes in queens , nurses and foragers and performing molecular evolution analyses .
We identified two adjacent copies of Vg genes ( Pb_Vg1 and Pb_Vg2 ) in the genome of Pogonomyrmex barbatus [33] with predicted lengths of 1742 and 1654 amino acids , respectively ( Table 1 ) . The two genes are separated by a putative mariner-like transposon , a DNA transposable element that has been involved in duplication events [34] . The two Vg genes have an identical number of exons ( Figure 1A ) and share the three structural domains typical of vitellogenins: the lipoprotein N-terminal domain ( LPD-N ) , the domain of unknown function 143 ( DUF1943 ) and the von Willebrand factor type D domain ( VWD ) [35] , [36] ( Figure 1B ) . However , the coding sequence of Pb_Vg2 is truncated compared to Pb_Vg1 because of an earlier stop codon in the last exon of Pb_Vg2 . To determine whether the presence of multiple Vg genes is a general feature of ants , we searched for Vg genes in the five additional recently published ant genomes . These are divided up into four different subfamilies: Myrmicinae ( Atta cephalotes and Acromyrmex echinatior ) [37] , [38] , Formicinae ( Camponotus floridanus ) [39] Dolichoderinae ( Linepithema humile ) [40] and Ponerinae ( Harpegnathos saltator ) [39] . We found that numbers of Vg genes vary between one and five per species ( Table 1 ) , and that when a genome contains multiple Vg genes , they are always adjacent . To determine the evolutionary history of these genes , we subsequently constructed a phylogenetic tree using known hymenopteran Vg sequences . The phylogenetic analysis ( Figure 2 ) revealed that the first duplication of the ancestral Vg gene in ants resulted in two gene subfamilies with different predicted amino acid length ( Table 1 ) . The phylogenetic analysis also showed that additional duplications occurred in the lineages leading to Acromyrmex echinatior , Solenopsis invicta and Linepithema humile . Interestingly , the two Pogonomyrmex barbatus genes ( Pb_Vg1 and Pb_Vg2 ) respectively cluster with the S . invicta genes preferentially expressed in queens ( Si_Vg2 and Si_Vg3 ) and foraging workers ( Si_Vg1 and Si_Vg4 ) [29] . To test the prediction that P . barbatus Vg genes display caste-specific expression profiles similar to their closest S . invicta orthologs , we performed quantitative RT-PCR analysis of Vg genes in P . barbatus queens and workers in five independent colonies ( Figure S1 ) . On average , Pb_Vg1 was 4 . 7 times more highly expressed in queens than in nurses ( pMCMC <0 . 0001 ) and 908 times more highly expressed in queens than in foragers ( pMCMC <0 . 0001 ) . The expression of Pb_Vg2 did not differ between queens and nurses ( pMCMC = 0 . 98 ) , but it was on average 5 . 7 times more highly expressed in foragers than in queens ( pMCMC = 0 . 0028 ) ( Figure 3 ) . Furthermore , we tested whether the expression of Vg genes in P . barbatus is associated to task performance as predicted by the RGPH . We found that Pb_Vg1 was significantly more highly expressed in nurses than in foragers in 5 out of 5 colonies ( Wilcoxon tests; col1: W = 56 , p = 0 . 001; col2: W = 70 , p = 0 . 0001; col3: W = 56 , p = 0 . 0003; col4: W = 49 , p = 0 . 0006; col5: W = 48 , p = 0 . 0007 ) , while Pb_Vg2 was significantly more highly expressed in foragers than in nurses in 4 out of 5 colonies ( Wilcoxon tests; col1: W = 28 , p = 0 . 06; col2: W = 16 , p = 0 . 005; col3: W = 0 , p = 0 . 0003; col4: W = 5 , p = 0 . 02; col5: W = 0 , p = 0 . 0007 ) . On average , Pb_Vg1 was 190 times more highly expressed in nurses than in foragers ( pMCMC<0 . 0001 ) and Pb_Vg2 6 . 5 times more highly expressed in foragers than in nurses ( pMCMC <0 . 0001 ) ( Figure 3 ) . Pb_Vg1 is the predominant transcript in workers as it is expressed in strikingly high levels in nurses compared to Pb_Vg2 in foragers ( Figure 3 ) ; a pattern similar to that observed for the single Vg gene in honeybees . Finally , we performed molecular evolution analyses to determine the relative contributions of selection for novel biochemical functions ( i . e . positive selection ) , selection for the maintenance of existing biochemical functions ( i . e . purifying selection ) and neutral evolution in the evolution of ant Vg genes . In particular , we estimated selective pressures on two basal branches of Vg respectively leading to primitive ants and modern ( Formicoid ) ants , and on the two branches that followed the duplication of Vg in the ancestor of modern ants ( Figure 2 ) . Analyses of the branch common to the ancestor of all ants ( Formicidae ) and the branch common to all modern ants yielded identical results: 60 . 5% of codon sites evolved under purifying selection ( dN/dS = 0 . 24 ) , 39 . 5% neutrally ( dN/dS = 1 ) , and none had evidence for positive selection . The two branches that followed the duplication of Vg are interesting because one branch includes all Vg genes known to be more highly expressed in queens than in workers ( hereafter referred to as subfamily A , which includes Pb_Vg1 , Si_Vg2 and Si_Vg3 ) , while the other branch includes all Vg genes more highly expressed in foraging workers than in queens ( hereafter referred to as subfamily B , which includes Pb_Vg2 , Si_Vg1 and Si_Vg4 ) ( Figure 2 and Table 1 ) . The branch leading to subfamily B shows overall relaxation of purifying selection and no significantly positively selected sites . However , the branch leading to subfamily A shows strong evidence for positive selection ( p = 0 . 008 ) , with a total of 7 . 1% of sites under positive selection . The three sites with the highest posterior probabilities of being under positive selection in this branch ( S44 , E382 and N456; pBEB>95% ) are part of the main vitellogenin N-terminal lipoprotein domain ( LPD-N ) that is likely implicated in the uptake of vitellogenin to the ovary [41] , providing further support that these changes affect the biochemical properties of the protein .
The results of this study suggest that Vg has been co-opted to regulate worker behavior in the ant P . barbatus as in the honeybee . Support for RGPH in groups of insects that evolved sociality independently , demonstrates that the co-option of reproductive pathways to regulate behavior is a major director in the evolution of sociality in insects . On the other hand , the expression of one Vg paralogs in sterile workers reveals that Vg adaptation to regulate worker behavior is not necessarily linked to reproduction but maybe linked exclusively to nutritional functions . Our result suggests that , after the initial duplication in ants , Vg genes underwent neo- or subfunctionalization to acquire caste- and behavioral-specific functions . Overall , our results suggest that even though ants and bees evolved sociality independently , they have conserved similar mechanisms to regulate division of labor .
To determine gene models , we first ran TBLASTN using known hymenopteran Vg sequences against ant genome sequences downloaded from the fourmidable database [63] . Subsequently , ruby/bioruby scripts [64] were used to extract relevant subsets of each genome . Gene predictions were generated on each subset using MAKER2 [65] s65mand subsequently manually refined using Apollo [66] . Conflicts gene predictions were resolved by using EST data when available , splice prediction algorithms ( http://www . fruitfly . org/seq_tools/splice . html ) and manual verification of splicing consensus sequences . ) Inaccurate sequence alignment or phylogeny leads to misleading or incorrect results in molecular evolution analyses . We used an approach centered on the use of phylogenetically aware codon-level aligner PRANK , which is likely to minimize the risks of introducing errors [67] , [68] . This required several steps . We preliminarily aligned the 26 Vg protein sequences with MAFFT linsi [69] and removed ambiguous sections of the alignment using trimAl ( option -gappyout ) [70] . A first tree was built with RAxML ( model GTRGAMMAI ) [71] and rooted with “nw_root” ( Newick Utilities package [72] ) . This tree was used as a guide tree in PRANK [73] to obtain a high-quality codon-level alignment of the 26 Vg coding sequences . Ambiguous sections of the alignment were removed using trimAl ( option -gappyout ) and a final tree was built with RAxML ( GTRGAMMAI model ) ; 10 , 000 bootstraps were generated to assess its confidence . Selective pressures ( dN/dS ) on different parts of the phylogenetic tree were estimated using the branch-site codon-substitution model from CodeML ( PAML 4 . 6 ) [74] . Such dN/dS ratios are obtained by computing the number of non-synonymous changes ( dN ) over synonymous changes ( dS ) ( see Table 2 for more details ) . Vg site coordinates ( S44 , E382 , N456 ) are given as in Apis mellifera Vg ( Uniprot identifier VIT_APIME ) . Pogonomyrmex barbatus founding queens were collected during nuptial flights on July 15th , 2008 in Bowie , Arizona , USA ( N32°18′54″//W109°29′03″ ) . Colonies were then kept in laboratory conditions ( 30°C , 60% humidity and 12 h/12 h light∶dark cycle ) in 15×13×5 cm plastic boxes with water tubes , and were fed once a week with grass seeds and a mixture of eggs , honey and smashed mealworms . Thirty months later , five of these colonies were used to collect samples on December 16th , 2010 . Task performance in workers is age related , thus nurses tend to be younger than foragers [2] . Young ants interacting with the brood in the nest tube were considered as nurses . To collect foragers , each colony was connected with a cardboard-made bridge to a foraging area composed of a plastic box containing grass seeds . Any ant handling a food item in the foraging area was considered as forager . Ant samples were flash-frozen in liquid nitrogen and kept at −80°C for further RNA extraction . Whole body worker samples were used to measure the expression of Pb_Vg1 and Pb_Vg2 genes . RNA extractions were performed using a modified protocol including the use of Trizol ( Invitrogen ) for the initial homogenization step , RNeasy extraction kit and DNAse I ( Qiagen ) treatment to remove genomic DNA traces . For each individual worker , cDNAs were synthesized using 500 ng of total RNA , random hexamers and Applied Biosystems reagents . Levels of mRNA were quantified by quantitative real-time polymerase chain reaction ( qRT-PCR ) using ABI Prism 7900 sequence detector and SYBR green . All qPCR assays were performed in triplicates and subject to the heat-dissociation protocol following the final cycle of the qPCR in order to check for amplification specificity . qRT-PCR values of each gene were normalized by using an internal control gene ( RP49 ) . Paralog-specific primers ( Table S1 ) were designed using sequence alignment [75] and primer analysis [76] programs . Primer sequences overlapped coding regions split by introns , allowing the specific amplification of cDNA levels over eventual genomic DNA contaminations . Transcript quantification calculations were performed by using the ΔΔCT method [77] . All data were analyzed using R ( http://www . r-project . org/ ) and the R packages lme4 [78] and language R [79] . The effect of caste on gene expression relative values was analyzed using linear mixed effects models . To avoid pseudoreplication , the colony was included as a random effect . We checked for normality and homogeneity by visual inspections of plots of residuals against fitted values . To assess the validity of the mixed effects analyses , we performed likelihood ratio tests to confirm that the models with fixed effects differed significantly from the null models with only the random effects . Throughout the paper , we present MCMC ( Markov-chain Monte Carlo ) estimated p-values that are considered significant below the 0 . 05 threshold ( all significant results remained significant after Bonferroni correction ) .
|
One of the main features of social insects is the division of labor , whereby queens monopolize reproduction while sterile workers perform all of the tasks related to colony maintenance . The workers usually do so in an age-dependent sequence: young workers tend to nurse the brood inside the nest and older workers are more likely to forage for food . Previous studies revealed that vitellogenin , a yolk protein typically involved in the regulation of reproduction in solitary insects , has been co-opted to regulate division of labor in the honeybee . In this study , we investigate such a role of vitellogenin in another group of social insects: the ants . We first use phylogenetic analyses to reveal the existence of multiple vitellogenin genes in most of the sequenced ant genomes . Then we compare the expression of the two vitellogenin genes ( Pb_Vg1 and Pb_Vg2 ) among queens , nurses and foragers in the seed-harvester ant Pogonomyrmex barbatus . Our results suggest that , after the initial duplication in ants , the vitellogenin genes acquired caste and behavioral specific expression associated with reproductive and non-reproductive nutritionally related functions . This study also shows that ants and bees , despite having evolved sociality independently , have conserved similar mechanisms to regulate division of labor .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology",
"molecular",
"cell",
"biology",
"genetics",
"population",
"biology",
"biology",
"genomics",
"evolutionary",
"biology"
] |
2013
|
Vitellogenin Underwent Subfunctionalization to Acquire Caste and Behavioral Specific Expression in the Harvester Ant Pogonomyrmex barbatus
|
Birds flying through a cluttered environment require the ability to choose routes that will take them through the environment safely and quickly . We have investigated some of the strategies by which they achieve this . We trained budgerigars to fly through a tunnel in which they encountered a barrier that offered two passages , positioned side by side , at the halfway point . When one of the passages was substantially wider than the other , the birds tended to fly through the wider passage to continue their transit to the end of the tunnel , regardless of whether this passage was on the right or the left . Evidently , the birds were selecting the safest and quickest route . However , when the two passages were of equal or nearly equal width , some individuals consistently preferred the left-hand passage , while others consistently preferred the passage on the right . Thus , the birds displayed idiosyncratic biases when choosing between alternative routes . Surprisingly - and unlike most of the instances in which behavioral lateralization has previously been discovered - the bias was found to vary from individual to individual , in its direction as well as its magnitude . This is very different from handedness in humans , where the majority of humans are right-handed , giving rise to a so-called ‘population’ bias . Our experimental results and mathematical model of this behavior suggest that individually varying lateralization , working in concert with a tendency to choose the wider aperture , can expedite the passage of a flock of birds through a cluttered environment .
A bird flying through a complex and cluttered environment relies heavily on the use of visual cues to rapidly choose between alternative routes and avoid collisions with intervening obstacles . Goshawks , for example , display an impressive ability to fly through dense environments at high speeds [1] . Neural correlates of obstacle detection have been investigated in pigeons , where it was shown that neurons in the nucleus rotundus of the brain respond to a visual stimulus that depicts a moving object on a collision course [2] . While such neurons probably constitute part of an “early warning” system , it remains to be seen how the responses of such neurons contribute to the planning of an efficient and safe flight trajectory . What strategies do birds adopt to fly efficiently through complex environments ? Here we investigate the behavior of budgerigars when they are offered a choice between two passages , presented side by side , through which they can fly . The relative widths of the passages are varied to investigate the rules that govern the birds' choices , and to examine whether these rules can facilitate rapid flight of a flock of birds through dense environments .
We flew birds individually in a tunnel that presented a barrier with two apertures , positioned side by side halfway along its length , as illustrated in Fig . 1 . The frequency with which the birds chose one aperture or the other was recorded , as the relative sizes of the two apertures were varied . Some of the flights were filmed and reconstructed in 3D using high-speed stereo cameras , as described in “Methods” . Three examples of such flights , as viewed from above , are shown in Fig . 2 . In Fig . 2A a bird selects the left-hand aperture ( of width 60 mm ) over the right-hand aperture ( of width 40 mm ) . In Fig . 2B a bird initially flies toward the right-hand aperture ( of width 10 mm ) , but possibly finds it too narrow and then chooses the left-hand aperture ( of width 90 mm ) . In Fig . 2C a bird chooses the aperture on the right ( of width 10 mm ) because the left-hand aperture has zero width ( i . e . is non-existent ) . [Birds can , and do occasionally choose to fly through very narrow apertures because the flanking panels , made of cloth , are compliant . The wings are then folded back to allow the bird to ‘projectile’ through the slit ( unpublished observations ) ] . A preliminary analysis of the birds' choices revealed the following general characteristics . When the apertures were very different in width , the birds tended to prefer the wider aperture , regardless of whether it was on the right side or the left . The birds appeared to be selecting the safest and quickest route . However , when the apertures were of equal ( or nearly equal ) width , some individuals consistently preferred the left-hand aperture , while others consistently preferred the right . Left-biased birds preferred the left-hand aperture , while right-biased birds preferred the aperture on the right . Thus – as we shall demonstrate in greater detail below – each bird had its own , characteristic , side bias . What are the factors that govern the choice of aperture ? We examined this question in greater detail by investigating how the bird's choices changed as the relative sizes of the two apertures were varied systematically . This was done by varying the position of the central panel that separated them , as described in the ‘Methods’ section . As the central panel was moved from its extreme left-hand position to its extreme right-hand position ( in steps of 10 mm ) , the width of the left-hand aperture increased progressively from 0 mm to 100 mm , and the width of the right-hand aperture decreased progressively from 100 mm to 0 mm , as shown in Table 1 . Fig . 3A shows how the choice frequency for the left-hand aperture varied with its width , for one particular bird ( bird One ) . When the two apertures were equally wide ( or nearly so ) , the bird displayed a preference for the right-hand aperture , choosing it with a frequency of 74% . However , as the central panel was moved towards the right , making the right-hand aperture narrower than the left-hand one , the bird exhibited an increased preference for the left-hand aperture , eventually choosing it with 100% probability for left-hand aperture widths of 70 mm or greater . Conversely , when the central panel was shifted progressively towards the left , the bird showed a decreasing preference for the left-hand aperture , eventually not choosing it at all when it was 20 mm or narrower . Overall , the data suggest that bird One has a weak preference for the right-hand aperture , and that this bias is superimposed upon the bird's tendency to choose the larger of the two apertures ( but see below ) . The results of a similar experiment conducted with a different bird ( Casper ) are shown in Figure 3B . This bird was left-biased: when the apertures were of equal width , the bird showed a greater preference for the left-hand aperture , choosing it 87 . 5% of the time . This choice probability was significantly different from the random-choice level of 50% ( p<0 . 02 ) . Figs . 3C–E show data from three additional birds: Two , Drongo and Saras . Bird Two possessed a preference for the right-hand aperture . This bird chose the left-hand aperture only 12 . 5% of the time when the two apertures were equally wide , and this preference was significantly lower than the random choice level of 50% ( p<0 . 02 ) . Drongo ( Fig . 3D ) was also right-biased , but less strongly so than Two . Saras ( Fig . 3E ) was strongly left-biased . The capacity of the birds to discriminate differences in aperture width can be quantified by fitting the choice frequency data to a logistic function that describes the choice frequency FL for the left-hand aperture as ( 1 ) where d is the width of the left-hand aperture . B is a bias parameter that specifies the bias of the bird , as estimated from the fitted function . It is the width of the left-hand aperture at which this aperture is chosen 50% of the time , i . e . the bird chooses randomly between the left- and right-hand apertures . If B = 50 mm , the random choice occurs when the two apertures are equally wide , and the bird is unbiased . If B<50 mm , the bird is left-biased , and if B>50 mm the bird is right-biased . α is a parameter which defines the sharpness of the bird's transition between the left-hand aperture and the right-hand one . The larger the value of α , the steeper the transition , and the sharper the discrimination of aperture width . The parameters B and α , and their confidence intervals were determined by performing a least-squares fit of the logistic function to the data for each bird and for the pooled data from all birds , as described in “Methods” . The computed values of B , and their 95% confidence intervals ( Table 2 ) indicate that bird One has no significant bias , that Casper and Saras are left-biased ( B<50 mm ) , and that Two and Drongo are right-biased ( B>50 mm ) . Table 2 also shows the increase in the width ( Δd , in mm ) of the left-hand aperture that is required for the choice frequency for the left-hand aperture to increase from 25% to 75% . This is derived from the value of α , as described in “Methods” . The smaller the value of Δd , the sharper the discrimination of aperture width . On this measure , bird Two displayed the sharpest discrimination ( Δd = 3 . 2 mm ) and bird One displayed the least sensitive discrimination ( Δd = 19 . 7 mm ) . The mean value of Δd , averaged over all birds , is 13 . 0 mm , which indicates that , on the whole , the birds display an impressively sharp ability to discriminate aperture width . What would be the behavior of the population as a whole ? We examined this question by averaging the data from all of the birds , point by point , for each aperture width . The results are shown in Fig . 3F , which represents the average of the results obtained for all five birds , graphed in Fig . 3A–E . The averaged curve shows no significant bias ( B = 49 . 8; Fig . 3F and Table 2 ) . This is as one might expect , given that in the group of birds that we tested , one bird had no significant bias , three birds displayed a right bias , and two others a left bias . Furthermore , while the relative preferences for the two apertures change sharply as a function of their relative widths for each individual bird ( mean Δd = 13 . 0 mm ) , the relative preferences of the population as a whole change more gradually and smoothly ( Δd for pooled data = 23 . 2 mm ) . The reason for this is that the sharp transition displayed by each bird occurs at a different point along the horizontal axis , because of the different biases possessed by the individual birds . This smoothing effect may have interesting implications for the behavior of a flock of birds , as we shall see in the Discussion section . Fig . 3F suggests that the population , when considered as a whole , does not possess any net bias . A larger sample of birds would need to be examined before this statement can be made with complete confidence . Nevertheless , the spread of left and right biases that we have observed in the five birds that we have investigated suggests that , if there is a net bias at the level of the population ( towards the left or the right ) , it is likely to be small .
The data of Fig . 3 reveal that birds display significant lateralization in their visually guided behavior . To our knowledge , ours is the first report of lateralization in bird flight . The results reveal , furthermore , that the lateralization varies in strength and polarity from bird to bird , but has a value close to zero when averaged across several birds . The pattern of choices that we have observed in the dual-aperture experiments is similar , in some respects , to that reported for tree swallows [3] . They found that tree swallows , when presented with two apertures of different width , tended to choose the wider aperture . However , that study did not examine how the birds' choices varied with changes in the relative widths of the two apertures – their experiments were conducted using two apertures that were either equally wide , or which differed in width by a fixed value . Furthermore , while our findings indicate a clear and strong side bias in most of the individuals that we have tested , Mandel et al . [3] state that they find no individual bias – or any net population bias – in their birds , when they chose between two equally wide apertures . However , their study presents only pooled data from all of the birds ( which reveals no population bias , in agreement with our findings ) , but it does not provide any data or analysis of the performance of individual birds . Their study does not permit a statistical comparison of the performances of the individual birds , because each bird was tested only once in each of their experimental configurations . The question of whether tree swallows differ from budgerigars with respect to individual variations in bias , therefore , remains unresolved . However , there could well be differences between the two species , given that they tend to inhabit somewhat different environments . So far , lateralization of vision in birds has been investigated primarily in the context of tasks that involve object detection and recognition . For example , pigeons memorize visual patterns better when they are viewed with the right eye; whilst chickens use their right eye to detect food , and their left eye to maintain a vigil against predators [4] . It has been suggested that birds that are strongly lateralized are good at multitasking . Parrots that have strongly lateralized brains are better able to use their beak as well as their feet in an experimental task that involves acquiring a food item suspended by a string [5]; Pigeons with strongly lateralized brains are better able to visually discriminate grain from grit [6] . In the above examples , the bias has been observed to occur at the population level . That is , almost all of the individuals display the same direction of bias [7]–[9] . It has been suggested that the presence of a population bias can be beneficial for species that are social: for example , a school of fish would all turn in the same direction when chased by a predator , thus ensuring that individual members do not get singled out for an attack ( e . g . [10] ) . On the other hand , it has also been suggested that individuals of non-social ( solitary ) species would benefit from having individually different biases , because their escape responses would then be less predictable to a predator [7] . However , there are many documented instances of individually varying lateralization , for which the adaptive benefits – if any – remain unexplained: as in the eyeing and picking up of food by parrots [11] , [12] , or the use of twigs to dig out worms from holes by New Caledonian crows [13] , [14] . Our study has revealed that budgerigars display individually varying lateralization when they are required to choose between to apertures . As we shall show in the following discussion and the mathematical model , this behavior can be advantageous when a flock of budgerigars attempts to fly through a densely cluttered environment . What might be the selective advantage of having individually varying biases in the way birds use vision to guide their flight ? One possibility may be an enhancement in the speed and safety with which a flock of birds can fly through dense foliage . It is clear that budgerigars confront this problem often , as do flocks of other bird species . When a flock is faced with a choice of flying through one of two clear passages through a thicket of branches , it would be detrimental if all of the birds were to possess the same bias , say , toward the left . A population bias of this kind would make all of the birds try to fly through the left-hand passage , thus blocking each other , and slowing down as well as endangering the passage of the flock through the thicket . ( The right-hand passage would not be used at all , and therefore would be wasted ) . On the other hand , it would also be detrimental to have no bias at all in each of the birds , because this would tend to make each individual vacillate in front of two equally wide passages before making a decision , again slowing down the progress of the flock through the thicket and increasing the likelihood of bird-to-bird collisions . Furthermore , if the two passages were of unequal size , a flock of unbiased birds would all try to fly through the wider passage , overcrowding it and again slowing down progress and increasing the likelihood of bird-to-bird collisions . The narrower passage would not attract any birds even if it were wide enough to permit safe flight , and it would thus be a ‘waste’ of a potentially useful conduit . On the other hand , if , say , half the population was left-biased and the other half right-biased , two passages of equal width would attract roughly equal numbers of birds , thus speeding up the progress of the flock through the thicket . Furthermore , the left-biased and right-biased birds would choose the left and right-hand passages without any hesitation , leading to a quicker and safer passage of the flock through the thicket . In this case , as the right-hand passage is gradually made wider than the left-hand passage , all the birds would not immediately flock to the right-hand passage: many of the left-biased birds would continue to favor the left-hand passage until it became too narrow for safe transit . Thus , a hybrid flock of left and right-biased birds would make better use of both of the available routes , and fly through the thicket more quickly . In the next section we describe a mathematical model that characterizes the above discussion quantitatively , and demonstrates that the transit of a flock of birds through a two-passage environment will be quickest when individual birds in the flock carry different biases – ranging from an extreme left-bias , through no bias , to an extreme right-bias . We also show that the transit time of the flock would be a minimum when the members of such a flock ( treated as a whole ) choose each passage with a frequency that is proportional to its width . That is , the number of birds that use each passage is proportional to the width of the passage . This would ensure that both passages complete transmitting birds at the same time , and are therefore used optimally . In reality , given that the birds cannot fly through apertures that are narrower than about 30 mm , one would expect that the frequency of choosing an aperture would not increase strictly linearly with its width . Rather , it would vary as a broad sigmoidal function that has a threshold width of about 30 mm , and which attains the 100% level at a width of about 70 mm ( beyond which the other aperture would be narrower than 30 mm and therefore not traversable ) . This is indeed the shape of the discrimination curve that is displayed by the pooled data ( Fig . 3F ) . The above considerations would also apply to a situation where a flock encounters several apertures . For example , let us consider a case of four equal apertures , arranged as shown in Fig . 4 . Let us assume that some individuals in the flock prefer aperture A ( the upper left-hand aperture ) , others aperture B ( the upper right-hand aperture ) , some prefer aperture C ( the lower left-hand aperture ) and yet others prefer aperture D ( the lower right-hand aperture ) . It can then be shown that this individually varying bias , acting in concert with a tendency to prefer the widest aperture when the apertures have different widths , will again lead to an optimal routing of the birds through the various apertures , and expedite the transit of the flock . We conclude that flying budgerigars display characteristic , individual biases when choosing between alternative routes . This is the first report of lateralization in visually guided bird flight . Contrary to most other known instances of lateralization in birds , the bias occurs at the level of the individual , rather than the population . Our mathematical model of this behavior suggests that the individually varying bias , working in concert with a general tendency to prefer routes that are more easily traversable , can expedite the passage of a flock of birds through dense foliage . Our model – which is a simplified , first attempt to characterize choices between navigable apertures – assumes that when the birds make these choices whilst in a flock , they behave independently of each other . That is , our analysis neglects any interactions that might occur among the birds when they are making their choices . Such interactions , if extant , would be an important subject of future experimentation and theory . We note that the behavioral task that we have studied here is fundamentally different in nature from that of flocking or schooling , where interactions among individuals facilitate the coordinated movement of a group of individuals in free space ( e . g . [15]–[17] ) . There the task is not necessarily to select the best aperture through which to fly , because all of the birds are flying at a more-or-less constant speed . Rather , each bird needs to adjust its position and speed to maintain a fixed , short distance to its nearest neighbors , to ensure a tight flock . In addition , there may be a randomly manifested tendency for individuals to move toward the center of the flock , to ensure that all birds experience approximately the same risk of predation . Birds flying through a cluttered environment , on the other hand , are likely to be in a different behavioral state because this situation poses a different set of challenges . Here we present a simple mathematical model that captures the behavior of the birds when they are confronted with the task of choosing between two apertures , and incorporates the factors and tradeoffs that could influence the passage of a flock of budgerigars through the two apertures . We assume that the width of the left-hand aperture is d , and that of the right-hand aperture is ( D-d ) , where D is the total width of the two apertures . When , the two apertures are of equal width . We assume that the time T taken for a single bird to fly though a passage is inversely proportional to the width of the passage . While we do not know if this assumption is exactly true , it is a reasonable first approximation , given that ( a ) the narrower the passage , the greater the difficulty in negotiating it , and the longer the bird will take to pass through it; and ( b ) if visually guided flight dynamics of budgerigars are similar to bees , the speed of their flight through a passage will be proportional to the width of the passage [18] , so that the time required to fly through the passage will be inversely proportional to its width . Indeed , there is recent evidence that budgerigars use optic-flow cues to regulate their flight speed , which would lead to this kind of behavior [19] . Thus , the times TL and TR taken by a bird to fly through the left- and right-hand apertures will be given respectively by ( 2 ) ( 3 ) where K is a constant of proportionality . When the two apertures are of equal width , we have , which leads to ( 4 ) If a flock of N budgerigars encounters the two apertures , and if NL of them choose to fly through the left-hand aperture and NR through the right-hand aperture ( NL+NR = N ) , the time required for the NL birds to transit the left-hand aperture will be , from ( 2 ) , ( 5 ) and the time required for the NR birds to transit the right-hand aperture will be , from ( 3 ) , ( 6 ) Let us now consider , in turn , a number of ways in which the birds might choose between the two apertures and examine , for each case , the time taken by the entire flock to pass through the twin-aperture obstacle . The optimum overall choice probabilities predicted by the model , as shown by the red and blue curves in Fig . 5B , are very reminiscent of the so-called “Matching Law” , which states that an animal's choices between two options tend to be proportional to the relative benefit ( or reward ) that is offered by each option [20] , [21] . Interestingly , while individual birds display “overmatching” and varying degrees of “bias” as characterized by Equation ( 2 ) in [21] , the predicted behavior of the flock as a whole follows the classic , proportional Matching Law , as postulated in the pioneering work of Herrnstein ( 1961 ) . Our study also finds that individual birds display different sensitivities to changes in gap width– that is , the transition from preferring the left-hand gap to the right-hand gap occurs over a smaller change in gap width for some birds , than for others . This variation in sensitivity is captured by the variations in the parameter Δd ( Table 2 ) . The reasons for this variation across individuals are presently unclear . They could arise simply from individual differences in sensory discrimination capacity . Alternatively , evolution may have tailored these differences to fine-tune and better optimize flock performance . For example , in a flock with a large number of birds , the optimum overall choice probability functions ( as shown by the red and blue curves in Fig . 5B ) could be achieved by having individuals with sharp discrimination of gap widths ( small Δd ) , but with a large variety of biases . In a flock with a small number of birds , on the other hand , there can only be a few different biases , and so the optimum overall choice probability functions can be realized only if the discrimination of gap widths is relatively poor ( large Δd ) . Further work , investigating individuals from different-sized flocks , would be required to investigate this possibility .
All experiments were carried out in accordance with Australian Laws on the protection and welfare of laboratory animals and the approval of the Animal Experimentation Ethics Committees of the University of Queensland , Brisbane , Australia ( Permit QBI/646/07/ARC ) . Adult male and female wild type budgerigars ( n = 5 , approximately 1 year old ) served as subjects for the experiments . The birds were obtained from different local breeders . They were reared in aviaries and did not have the opportunity to fly outdoors in flocks . Male budgerigars were identified by a characteristically green plumage and a distinctly blue nasal coloration while the females had a characteristic pink nasal coloration . The birds were housed in pairs in identical cages of length 470 mm , breadth 345 mm and height 820 mm , and were not under acoustic or visual isolation . All of the birds were housed indoors in a room ( length 4740 mm , width 2940 mm , height 3320 mm ) . Indoor lighting was provided by Phillips daylight fluorescent tubes . The lights were controlled by an automatic timer ( HPM , Excel Light Switch and Timer , Cat XL 770 T ) , which provided a 12∶12 L∶D photoperiod . Food and water were provided ad libitum . The food was commercial budgerigar seed mix ( Trill , budgerigar seed mix , Wacol , Queensland , Australia ) containing a mixture of seeds , shell grit and essential vitamins and minerals . The birds were also fed occasionally with apples and greens . Daily , the birds were moved to an adjoining screened patio , of length 5400 mm , width 2300 mm and height 1800 mm , where they were released from their cages and allowed to fly between perches . This enclosure provided the opportunity for regular flight as well as exposure to natural daylight . It also contained a bird bath . The budgerigars were flown in a purpose-built climate-controlled corridor ( temperature: 23–25°C , relative humidity: 35–40% ) of dimensions 7280 mm ( length ) , 2440 mm ( height ) and 1360 mm ( width ) , as illustrated in Fig . 1 . The walls and floor were painted with a white , low sheen acrylic paint . Each wall was decorated with vertically oriented black , machine-cut cardboard stripes , 110 mm wide and separated by 110 mm edge to edge . Illumination was provided by four lamps in the ceiling , each carrying two 36 W fluorescent tubes ( L 36W/880 Osram Skywhite FLH1 ) driven by a 40 kHz ballast to avoid any perception of flicker . Halfway along the tunnel ( 3000 mm from the start ) the birds encountered a barrier , which presented two vertically oriented apertures . The barrier was composed of cloth panels , stretched from the ceiling to the floor , to prevent accidental injury to the birds . The apertures were created by using three panels to create two vertical slits ( Fig . 1 ) . Each panel was composed of a cloth that carried a black-white checkerboard texture , of check size 40 mm×40 mm , printed on it ( SJCLOTH91418 , Studio Jet Instant Dry 180MIC , GBC Australia . Male and female budgerigars were brought individually into one end of the corridor . The birds were induced to take off from a hand-held perch when it was slowly rotated , and were trained initially to fly to the other end of the corridor to receive a food reward . In the later stages of training the food reward was dispensed with: the birds automatically took off when the perch was rotated , flew through the aperture and continued to the other end of the corridor , where they left the tunnel through a door at the far end , to be reunited with their companions . For each bird , this shaping and training procedure required approximately 30 flights , spread over 3 days . Once training was complete , the bird was flown under different experimental conditions and filming was commenced . Flights of individual birds were captured in three dimensions using two high-speed video cameras ( DRS lightning RDT , DRS Technologies Inc , USA ) at a frame rate of 250 frames/sec . The cameras were controlled by a custom configured Pentium 4 computer running special-purpose software ( Midas 2 . 0 , Xcitex , Inc , USA ) . One camera was placed at the center of the ceiling of the corridor , looking downwards . The other camera was placed at the center of the end wall of the corridor from where the birds commenced their flight , and looked horizontally along the axis of the corridor . Each flight yielded two synchronized image sequences , one representing an overhead view of the bird and the other a rear view of the bird during the flight along the corridor . Stereo calibration of the cameras was carried out using a reference checkerboard pattern ( check size 150 mm×150 mm ) in association with the J . Y . Bouguett camera calibration toolbox [22] , [23] . This procedure delivered the calibration parameters for each camera ( including characterization of imaging distortions ) and also determined the precise 3-D position and orientation of one camera with respect to the other . Tests with calibration markers indicated that the system had an absolute positional accuracy of ca . 10 mm×10 mm×10 mm . The video cameras were run at a frame rate of 250 frames per second . The image co-ordinates of the center of the head of the bird in each pair of synchronized frames were digitized interactively using a computer mouse and a custom-designed Matlab program . The head was clearly visible by virtue of the natural yellow patch that it carried ( which appears white in a black and white image ) . The sequences of head co-ordinates obtained from the two image sequences were then used in conjunction with the camera calibration parameters to reconstruct the trajectory of the head in 3-D , through the entire flight sequence . Trajectories were plotted without down-sampling the data . The apertures were presented halfway along the tunnel in a transversely oriented wall . Each aperture was oriented vertically and extended from the floor to the ceiling , as shown in Figure 1 . The two apertures were created by constructing the transverse wall out of three panels . There were two outer panels , each 450 mm wide extending inwards from the side walls . In addition there was a central panel , 340 mm wide . All of the panels carried the checkerboard pattern . The relative widths of the two apertures were varied systematically , in different experiments , by changing the position of the central panel along the width of the tunnel . When the central panel was positioned exactly midway between the two outer panels , each aperture was 50 mm wide . Displacing the central panel to the left caused the left-hand aperture to become narrower and the right-hand aperture to become wider , and vice versa . By varying the position of the central panel in steps of 10 mm , the relative widths of the two apertures were varied systematically from one extreme of 0 mm ( left ) and 100 mm ( right ) , through the symmetrical position of 50 mm ( left ) and 50 mm ( right ) , to the other extreme of 100 mm ( left ) and 0 mm ( right ) , as shown in Table 1 . The 11 different experimental conditions were presented in random sequence , as prescribed by a computer-generated sequence of random numbers generated using Matlab ( Mathworks , USA ) . 5 birds were used in the experiments: One , Casper , Two , Drongo and Saras . At the time of conducting these experiments , our capacity to hold and maintain birds in an ethically proper environment was restricted to 5 birds . Each bird was tested on each of the experimental conditions for between 6 and 14 trials , so that each bird made a total of 106–107 choices . The data were analyzed to obtain the choice frequency ( expressed as a percentage of the total number of choices ) for the left-hand aperture , for each experimental condition and for each bird . Thus , if a particular bird chose the left- hand aperture in 8 out of 11 trials in one particular experimental condition , its choice frequency for the left-hand aperture was calculated as 100× ( 8/11 ) % = 73% . The choice frequency for the right-hand aperture was then 100%-73% = 27% . The choice frequencies for the apertures were analyzed to determine whether they were significantly different from the random-choice level of 50% . If a bird chooses the left-hand aperture n times out of N trials , the probability of choosing the left-hand aperture α is n/N . Assuming that the bird's choice behavior follows a binomial distribution , the standard error of the mean of this distribution , σ , can be calculated as [24] . This value of σ is then used in a standard two-tailed t-test to determine whether α is significantly different from the random-choice level of 50% , as described in [24] and [25] . We quantified the capacity of the birds to discriminate differences in aperture width by fitting the choice frequency data to a logistic function [26] that describes the choice frequency FL for the left-hand aperture as ( 24 ) where d is the width of the left-hand aperture . This function is illustrated in Fig . 6 . B is a bias parameter that specifies the bias of the bird , as estimated from the fitted function . α is a parameter which defines the sharpness of the bird's transition between the left-hand aperture and the right-hand one . The parameters B and α , and their 95% confidence intervals were determined by performing a least-squares fit of the logistic function to the data using the NLINFIT and CI routines of Matlab ( Mathworks , USA ) . The logistic function was chosen to model the data because ( a ) it is a relatively simple function ( b ) it is perfectly anti-symmetrical about the 50% choice frequency level , as is required by the reciprocal relationship between the widths of the left- and right-hand apertures . However , other anti-symmetrical functions could have been used instead , and would have yielded similar results . From the fitted logistic functions we can also estimate the sharpness of each bird's ability to discriminate changes in aperture size by calculating the change in aperture width ( Δd ) that is required for the choice frequency of the bird for the left-hand aperture to increase from 25% to 75% . This is carried out as follows . d1 , the width of the left-hand aperture that elicits a choice frequency of 25% for this aperture , is given by the relationship ( 25 ) from which we can solve for d1: ( 26 ) Similarly d2 , the width of the left-hand aperture that elicits a choice frequency of 25% for this aperture , is given by the relationship ( 27 ) from which we can solve for d2: ( 28 ) Therefore Δd , the change in aperture width required for the preference of the left-hand aperture to increase from 25% to 75% is given by ( 29 ) We note that Δd is inversely proportional to the value of α . It does not depend upon the parameter B , which specifies the bias of the bird . This is appropriate , because Δd is meant to indicate the sharpness of the transition of the bird's preference from the one aperture to the other , irrespective of where this transition occurs . A recent study [27] , published while this paper was under review , has demonstrated that budgerigars also display individually varying lateralization in other tasks such as choice of landing location , or choice of foot used to climb on to a perch .
|
Birds display a clear mastery of the skill of flying rapidly and safely through complex and cluttered environments . An example of this can be viewed at http://www . youtube . com/watch ? v=p-_RHRAzUHM , which shows a bird flying at high speed through a dense forest . Such mastery requires the ability to determine , from moment to moment , which of several possible routes would provide the safest and quickest passage . Our study is one of the first to investigate how birds achieve this . Our experiments reveal that , when flying budgerigars are required to choose between two passages , they tend to favor the wider passage . However , this tendency is superimposed upon a bias that , surprisingly , varies from bird to bird: some individuals show an intrinsic preference for the left-hand passage , and others for the passage on the right . This is very different from handedness in humans , where the majority of humans are right-handed . We develop a mathematical model of the interaction between the birds' individual biases with their tendency to prefer the wider passage . The model reveals that this interplay is actually beneficial – it can expedite the passage of a flock of birds through a complex environment .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"biology"
] |
2014
|
Behavioral Lateralization and Optimal Route Choice in Flying Budgerigars
|
The Global Programme to Eliminate Lymphatic Filariasis ( LF ) aims to eliminate the disease as a public health problem by 2020 by conducting mass drug administration ( MDA ) and controlling morbidity . Once elimination targets have been reached , surveillance is critical for ensuring that programmatic gains are sustained , and challenges include timely identification of residual areas of transmission . WHO guidelines encourage cost-efficient surveillance , such as integration with other population-based surveys . In American Samoa , where LF is caused by Wuchereria bancrofti , and Aedes polynesiensis is the main vector , the LF elimination program has made significant progress . Seven rounds of MDA ( albendazole and diethycarbamazine ) were completed from 2000 to 2006 , and Transmission Assessment Surveys were passed in 2010/2011 and 2015 . However , a seroprevalence study using an adult serum bank collected in 2010 detected two potential residual foci of transmission , with Og4C3 antigen ( Ag ) prevalence of 30 . 8% and 15 . 6% . We conducted a follow up study in 2014 to verify if transmission was truly occurring by comparing seroprevalence between residents of suspected hotspots and residents of other villages . In adults from non-hotspot villages ( N = 602 ) , seroprevalence of Ag ( ICT or Og4C3 ) , Bm14 antibody ( Ab ) and Wb123 Ab were 1 . 2% ( 95% CI 0 . 6–2 . 6% ) , 9 . 6% ( 95% CI 7 . 5%-12 . 3% ) , and 10 . 5% ( 95% CI 7 . 6–14 . 3% ) , respectively . Comparatively , adult residents of Fagali’i ( N = 38 ) had significantly higher seroprevalence of Ag ( 26 . 9% , 95% CI 17 . 3–39 . 4% ) , Bm14 Ab ( 43 . 4% , 95% CI 32 . 4–55 . 0% ) , and Wb123 Ab 55 . 2% ( 95% CI 39 . 6–69 . 8% ) . Adult residents of Ili’ili/Vaitogi/Futiga ( N = 113 ) also had higher prevalence of Ag and Ab , but differences were not statistically significant . The presence of transmission was demonstrated by 1 . 1% Ag prevalence ( 95% CI 0 . 2% to 3 . 1% ) in 283 children aged 7–13 years who lived in one of the suspected hotspots; and microfilaraemia in four individuals , all of whom lived in the suspected hotspots , including a 9 year old child . Our results provide field evidence that integrating LF surveillance with other surveys is effective and feasible for identifying potential hotspots , and conducting surveillance at worksites provides an efficient method of sampling large populations of adults .
Lymphatic filariasis ( LF ) is a mosquito-borne parasitic infection caused by Wuchereria or Brugia species of helminths . Mosquito vectors vary between countries and regions , and include Aedes , Anopheles , Culex and Mansonia species . Globally , it is estimated that 68 million people are affected , comprising approximately 36 million microfilaraemic persons and 36 million with disabling complications such as severe lymphedema , including elephantiasis and scrotal hydroceles [1] . The Global Programme to Eliminate LF ( GPELF ) was launched by the World Health Organization in 2000 , with the aim of eliminating the disease as a public health problem by 2020 . The program consists of two components: i ) to interrupt transmission through mass drug administration ( MDA ) and ii ) to control morbidity and disability of affected populations [2] . As part of GPELF , the Pacific Programme to Eliminate LF ( PacELF ) was formed in 1999 to focus on 22 Pacific Island Countries and Territories ( PICTs ) . PacELF focused on coordinating the elimination efforts in the PICTs , which include >3000 islands and 8 . 6 million people [3 , 4] . Since 2000 , the GPELF has made impressive progress globally , with a total of 6 . 2 billion treatments delivered to >820 million people [2] . Once elimination targets have been reached , effective monitoring and surveillance are critical for ensuring that programmatic gains are sustained in the long-term . The World Health Organization and GPELF have identified some key operational challenges in post-MDA surveillance , including i ) timely identification of residual areas of high-prevalence and/or resurgence , ii ) strategies for managing these high risk areas , and iii ) development of cost-effective surveillance strategies [5 , 6] . In American Samoa , a group of remote islands in the South Pacific , LF is caused by the diurnally sub-periodic W . bancrofti , and the main mosquito vector responsible for transmission is the highly efficient day-biting Ae . polynesiensis . Other vectors include Ae . samoanus ( night-biting ) , Ae . tutuilae ( night-biting ) , and Ae . upolensis ( day-biting ) [7–10] . Two rounds of MDA in 1963 and 1965 reduced microfilaria ( Mf ) prevalence from ~20% to <2% [4 , 11 , 12] . Unfortunately , transmission was not successfully interrupted , and antigen prevalence measured by rapid immunochromatographic test ( ICT ) had risen to 16 . 5% ( N = 3018 ) when the PacELF baseline survey was conducted in 1999 . Since then , American Samoa has made significant progress towards LF elimination . After seven rounds of MDA from 2000–2006 , antigen prevalence dropped to 2 . 3% ( N = 1881 ) in a community cluster survey in 2007 [13–15] . However , the results of the 2007 survey did not meet PacELF’s criteria for stopping MDA ( <1% antigenaemia , upper 95% CI <2% ) , and an additional round of MDA was recommended , but no further effective rounds of MDA were successfully completed after this time [15] . The WHO currently recommends post-MDA surveillance using transmission assessment surveys ( TAS ) , which use critical cut-off values of numbers of antigen-positive children aged 6–7 years to determine whether transmission has been interrupted in defined evaluation units [16] . In areas where W . bancrofti is endemic and Aedes is the principal vector , the target threshold for post-MDA transmission assessment surveys ( TAS ) is <1% antigenaemia [16] . American Samoa passed TAS-1 in 2011–2012 [17] and TAS-2 in 2015 [18] , but recent human seroprevalence studies and molecular xenomonitoring studies of mosquitoes identified epidemiological and entomological evidence of ongoing LF transmission [10 , 19 , 20] . As prevalence drops to very low levels in the end stages of elimination programs , not only will it become more challenging to detect any residual hotspots of ongoing transmission , but funding and resources for programmatic activities will also generally be reduced . The current WHO guidelines therefore encourage cost-efficient methods for post-MDA surveillance , including the integration of LF surveillance activities with other population-based surveys , and opportunistic screening of large groups ( e . g . military recruits , hospital patients , and blood donors ) for microfilaraemia , antigenaemia , or antibodies [16] . For example , in Togo [21] and Vanuatu [22] , nationwide LF surveillance has been successfully conducted by screening blood smears collected for malaria diagnosis . However , there is currently little evidence about the effectiveness of these strategies for identifying infected persons in the post-MDA setting . Lau et al previously reported a study of the seroprevalence and spatial epidemiology of LF in American Samoa after successful MDA [19] . The study used a serum bank collected from adults ( aged ≥18 years ) for a leptospirosis study in 2010 [23] , four years after the last effective round of MDA . The study found epidemiological evidence of possible residual foci of Og4C3 Ag-positive people in two localised areas , with an average cluster size of ~1 . 5 km . One cluster was found in the very small village of Fagali’i , where the seroprevalence of Og4C3 Ag was 30 . 8% ( 95% CI 9 . 1–61 . 4% ) . Another cluster spanned the contiguous villages of Ili’ili , Vaitogi , and Futiga , where overall Og4C3 Ag prevalence was 15 . 6% ( 95% CI 5 . 3–32 . 8% ) ( data derived from [19] ) . However , the findings were not definitive because information on microfilaraemia was not available; the study was conducted using a pre-existing serum bank and the findings were based entirely on serological markers . In this paper , we report results of a follow up study in 2014 to confirm whether there was indeed ongoing transmission and/or higher infection rates in the two suspected ‘hotspot areas’ identified from the previous work . If ongoing transmission was truly occurring in these two suspected hotspot areas , our findings would provide field evidence to support WHO’s recommendations for integrating LF surveillance activities with other population-based surveys , including those that only include adults [16] .
Ethics approvals were granted by the American Samoa Institutional Review Board , and the Human Research Ethics Committees at James Cook University ( H5519 ) and The University of Queensland ( 2014000409 ) . The study was conducted in collaboration with the American Samoa Department of Health , and official permission for village visits was sought from the Department of Samoan Affairs and village chiefs and/or mayors . Verbal and written information were provided to all participants ( or their parent or guardian ) in Samoan or English according to the participant’s preference . Signed informed consent forms were obtained from all participants , or their parent or guardian if under 18 years of age . American Samoa is a United States Territory in the South Pacific , consisting of a group of small tropical islands with a total population of 55 , 519 living in ~70 villages ( average population ~800 per village ) at the 2010 census [24] . Over 90% of the population live on the main island of Tutuila , and the remainder on the adjacent island of Aunu’u and the remote Manu’a group of islands . American Samoa has a tropical climate and is one of the wettest inhabited places in the world , with islands that include mountains , valleys , tropical rainforests , wetlands , fringing reefs , and lagoons . Field data were collected from American Samoa in 2014 from the following groups of participants: Adult workers and village residents were recruited by convenience sampling because probability-based sampling was not logistically possible with the available budget and resources . The field team was stationed at the weekly Department of Health clinic from May to December 2014 ( ~4 hours per visit ) , and invited all clinic attendees to participate . Visits to the tuna cannery , villages , and school were conducted over a 3-week period in October and November 2014 . The team visited the tuna cannery on four occasions ( ~4 hours per visit ) , and all employees on duty were invited to participate . For visits to the suspected hotspot areas , permissions were sought from village chiefs and mayors , who informed residents of the team’s pre-arranged visits . During the three village visits to Ili’ili/Vaitogi/Futiga , and one village visit to Fagali’i ( ~4 hours per visit ) , the field team was positioned in a prominent and central part of the community , and all residents were invited to participate . All children in Grades 3 to 7 who attended the elementary school in Ili’ili were invited to participate; information sheets and consent forms were distributed to parents and guardians about one week beforehand , and all children who returned valid consent forms were tested . The following samples and data were collected from each participant: Data on population demographics were sourced from the 2014 American Samoa Statistical Year Book [24] , and high-resolution GIS data were provided by the American Samoa GIS User Group [25] . Venous or fingerprick blood samples were tested for filarial antigen immediately after collection using the Alere BinaxNOW Filariasis immunochromatographic test ( ICT ) . If an ICT was positive , the result was confirmed by repeating the test , and two Mf slides were prepared , each with 60 μL of blood in 3 lines of 20 μL each per slide . Once thoroughly dried , slides were dehaemoglobinized , fixed with methanol and stained with 2% Giemsa stain for 50 minutes according to WHO guidelines [16] and examined at 100x magnification . Mf densities in 60 μL were determined by counting all Mf on each slide . Each set of Mf slides were read blindly by two or three experienced parasitologists , one at James Cook University Cairns ( PG ) and the other ( s ) at the LBJ Tropical Medical Centre in American Samoa and/or at James Cook University in Townsville , Australia . Counts were converted to Mf/mL and the final Mf density recorded was the average of the counts reported by two or three parasitologists . Venous blood samples were allowed to clot before centrifuging , and serum were aliquoted and stored at -20 degrees Celsius in American Samoa . Frozen sera and dried blood spots ( DBS ) were shipped to Australia for serological analysis at James Cook University , Cairns , Australia . All samples were tested for Og4C3 Ag using the TropBio Og4C3 Filariasis Antigen ELISA test ( Cellabs Pty . Ltd . , New South Wales , Australia ) using dilutions for serum and DBS recommended by the manufacturer . Bm14 Ab was measured using ELISA tests ( CDC in house version ) as previously described [19] . For Wb123 Ab , all the village samples and 109 of the adult worker samples were tested with in-house Wb123 ELISA as previously described [19] . The remaining 552 adult worker samples and 178 of the school children samples were tested using the InBios Wb123 ELISA [26] . Due to insufficient volumes of blood available , 149 children did not have Wb123 ELISA done by either method . All ELISAs used standard curves with kit provided standards ( Og4C3 Ag ) or known strong positives ( Bm14 Ab and Wb123 Ab ) . Cutoffs for positivity were aligned between the two Wb123 Ab methods . All ICT-positive individuals were treated with albendazole ( 400mg ) and diethyl-carbamazine ( DEC ) ( 6mg/kg ) according to WHO recommended dosages , and all treatments were provided free of charge . Children were treated with informed consent from and in the presence of at least one parent or guardian . The outcome measures used were positive results for ICT , Og4C3 Ag >32 units ( weak positive ) , Og4C3 Ag >128 units ( positive ) , ‘antigen’ ( ICT and/or Og4C3 Ag >32 units ) , Bm14 Ab , Wb123 Ab , and Mf . The positivity levels for Og4C3 Ag were chosen based on product information provided by Cellabs , and a previous study in American Samoa [19] . To determine whether residents of suspected hotspot areas had higher infection rates than residents of other villages , seroprevalence of adult workers who resided in other villages were used as the reference group to provide estimates of infection rates in the general population . Simple proportions were compared using Chi-squared tests or Fisher exact tests , and binomial exact 95% confidence intervals . Point estimates of antigen and antibody prevalence were calculated for residents of each of the suspected hotspot areas and adult workers who lived in other villages . Prevalence estimates were standardised for age using American Samoa’s age distribution data from the 2014 Statistical Yearbook [24] , and 95% confidence intervals calculated using the ‘stdize’ option in the ‘proportion’ command in Stata 14 , with ‘stdweights’ as the proportion of the population in each age group . Statistical associations between place of residence and presence of serological markers were quantified using univariable logistic regression ( weighted for age distribution ) . Because adult workers were used as the reference population , village residents aged <15 years were excluded from the logistic regression analyses . Data were managed using Microsoft Excel ( v14 , 2011 ) and Qualtrics ( Qualtrics , Provo , UT ) , an electronic platform for collecting and managing data . Stata 14 ( StataCorp , College Station , TX ) was used for data analyses , and p values of <0 . 05 were considered statistically significant .
The study included a total of 1 , 132 participants , comprising 172 employment clinic attendees , 498 tuna cannery workers , 125 community members from the two suspected hotspot areas , and 337 school children who attended an elementary school in Ili’ili . The school children represented 61 . 6% of the total 547 students enrolled in Grades 3 to 8 at the school . Of the 337 children , 283 ( 84 . 0% ) were residents of Ili’ili/Vaitogi/Futiga , one of the suspected hotspot areas . Children who attended the school but resided in other villages were also tested , but not included in the statistical analyses for residents of suspected hotspot areas because most lived near the school and were therefore not representative of the general population of children in American Samoa . The participants were classified into three groups for statistical analyses: Fig 1 shows the age distributions of residents of suspected hotspots , adult workers from other villages , and American Samoa’s general population . A summary of the representativeness of each study population is shown in Table 1 . Fig 2 shows the distribution of the general population on Tutuila , the locations of the suspected hotspot areas included in this study , the elementary school in Ili’ili where children were tested , and the clinic and tuna cannery where adult workers were tested . Fig 3 shows the residential locations of all adult workers who participated in this study . Although the adult workers were sampled at one clinic and one work site , they resided across 51 villages on the main island of Tutuila and the adjacent island of Aunu’u . Considering that convenience sampling was used in this study , the adult workers provided a reasonably representative sample of the general adult population in terms of place of residence . Overall , 72 . 3% ( 95% CI 61 . 4% to 81 . 6% ) of adult community members and 67 . 6% ( 95% CI 63 . 8% to 71 . 3% ) of adult workers reported taking MDA in the past , either in American Samoa and/or elsewhere , while 3 . 6% ( 95% CI 0 . 8% to 10 . 2% ) of adult community members and 1 . 3% ( 95% CI 0 . 5% to 2 . 5% ) of adult workers reported that a doctor or other health worker had previously diagnosed them with LF . There were no statistically significant differences in MDA participation or LF diagnosis between adult community members and adult workers , suggesting that convenience sampling did not introduce any significant participation biases towards either group being more likely to have LF infection . A total of 29 antigen-positive individuals were identified from ICT and/or Og4C3 Ag ( >32 units ) tests , with a female to male ratio of 1 . 07 and age range of 9 to 73 years . A summary of antigen-positive results for each group is shown in Table 2 . Of the 337 school children tested , three of the 283 residents of Ili’ili/Vaitogi/Futiga were ICT-positive ( 1 . 1% , 95% CI 0 . 2% to 3 . 1% ) , compared to none of the 54 who were residents of other villages ( 0% , one-sided 97 . 5% CI 0% to 6 . 6% ) . Four microfilaraemic individuals were identified out of 20 available slides examined , with Mf densities of 8 , 433 , 2667 , and 3267 Mf/mL . These counts were the average of two or three blind readings of the slides by different parasitologists . The Mf-positive persons were aged 9 , 29 , 35 , and 46 years , with a female:male ratio of 1 . All microfilaraemic individuals lived in the hotspot areas; three in Fagali’i and one in Vaitogi . The age distributions of antigen-positive and Mf-positive individuals are shown in Fig 4 . Of the 15 ICT-positive people who resided in suspected hotspot areas , Mf results were available for 14 , of which 4 ( 28 . 6% ) were Mf-positive . Of the 8 ICT-positive people who resided in other villages , Mf results were available for 6 , of which none were Mf-positive . The difference in proportion of Mf-positive results was not statistically significant , but sample sizes were small . The number of years lived in American Samoa was not significantly associated with seroprevalence for antigen or antibodies . Antigen prevalence was 3 . 2% in those who had lived in American Samoa since 2000 ( when MDA started ) , 2 . 5% in those who arrived after 2006 ( when the last effective round of MDA was conducted ) , and 3 . 3% in those who arrived between 2000–2006 ( Chi-squared test , p = 0 . 89 ) . There were no significant differences in seroprevalence between the three groups for Bm14 Ab ( Chi-squared test , p = 0 . 25 ) or Wb123 Ab ( Chi-squared test , p = 0 . 65 ) . Estimates of the seroprevalence of LF antigen ( positive ICT and/or Og4C3 Ag >32 units ) , Bm14 Ab , and Wb123 Ab were calculated , and adjusted for age based on the population age distribution reported in the 2010 census [24] . Comparisons of the age-adjusted seroprevalence for adult residents ( aged ≥15 years ) of each suspected hotspot area and adult workers who resided in other villages are summarised in Fig 5 . In adults who lived outside of hotspot villages , age-adjusted seroprevalence of LF antigen , Bm14 Ab , and Wb123 Ab were 1 . 2% ( 95% CI 0 . 6–2 . 6% ) , 9 . 6% ( 95% CI 7 . 5%-12 . 3% ) , and 10 . 5% ( 95% CI 7 . 6–14 . 3% ) respectively . Comparatively , adult residents of Fagali’i had significantly higher seroprevalence of antigen ( 26 . 9% , 95% CI 17 . 3–39 . 4% ) , Bm14 Ab ( 43 . 4% , 95% CI 32 . 4–55 . 0% ) , and Wb123 Ab 55 . 2% ( 95% CI 39 . 6–69 . 8% ) than adults living in other areas . Adult residents of Ili’ili/Vaitogi/Futiga also had higher seroprevalence of antigen ( 4 . 0% , 95% CI 1 . 8–8 . 8% ) , Bm14 Ab ( 12 . 5% , 95% CI 7 . 1–21 . 1% ) and Wb123 Ab 18 . 5% ( 95% CI 11 . 4–28 . 5% ) than adults living in other areas , but the differences were not statistically significant . Comparisons of the age-adjusted seroprevalence of antigen and antibodies in children ( aged 2 to 14 years ) from each of the suspected hotspot areas are shown in Fig 6 . Age-adjusted seroprevalence of antigen and antibodies were higher in child residents of Fagali’i compared to those who lived in Ili’ili/Vaitogi/Futiga: 5 . 0% ( 95% CI 0 . 7–29 . 5% ) versus 1 . 3% ( 95% CI 0 . 5–3 . 5% ) for antigen; 30 . 0% ( 95% CI 13 . 8%-53 . 4% ) versus 2 . 0% ( 95% CI 0 . 9–4 . 4% ) for Bm14 Ab; and 21 . 5% ( 95% CI 7 . 9–45 . 5% ) versus 5 . 8% ( 95% CI 3 . 1–10 . 5% ) for Wb123 Ab . The study did not collect data from children in other villages that were sufficiently representative for meaningful comparison with the results from suspected hotspots . It should be noted that for some children , there were insufficient blood samples for all serological tests to be conducted . Wb123 Ab results were only available for 173 of the 305 ( 56 . 7% ) children who lived in Ili’ili/Vaitogi/Futiga , and 19 of the 20 ( 95% ) children who lived in Fagali’i . Odds ratios of the presence of antigens and antibodies in residents of suspected hotspot areas were calculated using univariate logistic regression ( weighted for age distribution ) , using adult workers living in other villages as the reference group . Only participants aged ≥15 years were included in this analysis because the reference group did not include any participants aged <15 years . Table 3 shows that residents of both suspected hotspots were significantly more likely to be seropositive . Residents of Fagali’i had significantly higher odds of being antigen positive ( OR 20 . 4 for ICT , OR 31 . 8 for Og4C3 Ag >32 units , and OR 23 . 5 for any antigen ) and antibody positive ( OR 5 . 7 for Bm14 Ab and OR 9 . 5 for Wb123 Ab ) than adult residents of non-hotspot areas . Compared to this reference adult group , residents of Ili’ili/Vaitogi/Futiga also had significantly higher odds of being positive for Og4C3 Ag of >32 units ( OR 4 . 1 ) and Wb123 Ab ( OR 2 . 3 ) , but not for ICT ( OR1 . 6 ) or Bm14 Ab ( OR 1 . 0 ) .
Our results confirm that adult residents of the two suspected hotspot areas were significantly more likely to be seropositive for Og4C3 Ag and Wb123 Ab compared to adult residents of other villages in American Samoa . Residents of Fagali’i were also significantly more likely to be positive on ICT and seropositive for Bm14 Ab . We confirmed the presence of ongoing transmission in the suspected hotspot areas by identifying microfilaraemic residents , including a 9-year-old child . The results of this study therefore support the previous findings of suspected hotspots using a serum bank collected in 2010 for a leptospirosis study [19] , and provide field evidence that WHO’s recommendations for integrating LF surveillance activities with other population-based surveys are potentially effective and feasible . However , the current study does not allow us to determine whether the hotspots represent areas of persistent transmission that were not successfully interrupted by MDA , or newly formed hotspots after MDA was completed . Furthermore , our study confirmed ongoing transmission even though American Samoa passed TAS-1 of 6 to 7 year old children in 2011–2012 [17] and again passed TAS-2 in 2015 [18] ( conducted a few months after this study ) . Our findings suggest that testing adults is a potentially effective surveillance strategy , particularly if performed in conjunction with TAS and used as baseline data . However , this strategy might require a sampling scheme quite different from the current WHO recommended sampling methods for TAS . In the post-MDA setting , when overall prevalence is very low and typically even lower in young children , testing adults might be more accurate for determining transmission status and more sensitive for identifying hotspots . The age-adjusted estimates of the prevalence of all serological markers were higher in adult residents of Fagali’i compared to those who lived outside of hotspot villages . The seroprevalence of antigen and Wb123 Ab ( but not Bm14 Ab ) were higher in Ili’ili/Vaitogi/Futiga compared to residents of non-hotspot villages , but differences were not statistically significant , either because of the small sample size or the true absence of any difference . Although seroprevalence estimates for each group were standardised for age , the variations in age distribution between the groups could have made it more difficult to identify statistical differences . Adult residents of both hotspots had significantly higher odds of being seropositive for Og4C3 Ag and Wb123 Ab than residents of other ( non-hotspot ) areas . Adult residents of Fagali’i also had higher odds of being seropositive for ICT and Bm14 Ab , compared to residents of other areas , but this was not found in Ili’ili/Vaitogi/Futiga . The reasons for the differences in the patterns of serological markers between the two hotspot areas are not clear , but could potentially be attributed to differences in intensity of transmission; or the timing of possible reintroduction or resurgence; or differences between persistent transmission from before MDA versus reintroduction or resurgence . The age-adjusted estimates of overall seroprevalence of all antigens and antibodies were significantly higher in Fagali’i compared to Ili’ili/Vaitogi/Futiga , suggesting higher transmission intensity in Fagali’i both recently and in the past . Previous studies in children have demonstrated the appearance of antigen and Wb123 Ab earlier in the course of infection than Bm14 Ab [27] . In Ili’Ili/Vaitogi/Futiga , the higher odds of being seropositive to antigen and Wb123 Ab ( but not Bm14 Ab ) , together with ICT-positive 6–7 year olds in the local school in both TAS-1 and TAS-2 , might reflect more recent reintroduction and/or resurgence compared to Fagali’i . Our previous study found that recent migrants to American Samoa ( mostly from Samoa ) had significantly higher antigen and antibody prevalence [19] , but this study did not find any significant difference in seroprevalence and number of years lived in American Samoa . A possible explanation is that more time has lapsed since MDA , and there is less difference in infection risk and/or the impact of MDA between long-term residents and recent migrants . The strengths of our study include the large proportion of the population tested in the hotspot areas as well as the general population , and the wide range of age groups included . Our study population was highly stable and allowed accurate assessment of geographic variations in risk; >70% of each group had lived in American Samoa since the PacELF programme commenced in 2000 . Our results should also be considered in light of the study’s limitations . Because of financial constraints and limited resources , the study was conducted using convenience sampling instead of probability-based sampling . Our reference group ( adult workers who lived in other villages ) were over 15 years of age but children were included in the residents of hotspot areas . Our previous study found that recent migrants who had not lived in American Samoa from the beginning of PacELF were more likely to be seropositive for antigen and antibody [19] . A lower proportion of our reference group ( 70 . 7% , 95% CI 66 . 8–74 . 3% ) had lived in American Samoa since the beginning of PacELF compared to residents of Fagali’i ( 75 . 7% , 58 . 8–88 . 2% ) and residents of Ili’ili/Vaitogi/Futiga ( 75 . 7% , 95% CI 66 . 6–83 . 3% ) . Our study only considered the place of residence , but infection could have occurred at work or elsewhere , especially when efficient day biters are present . Each of these three limitations could have weakened the associations between living in a hotspot and the presence of serological markers , but our study found statistically significant results despite the limitations . The age distributions of the hotspot residents and adult workers were significantly different to that of the general population , but the estimates of population seroprevalence were adjusted for age . Convenience sampling of hotspot residents and adult workers might have introduced bias towards people who have been diagnosed with LF , concerned that they might have the infection , or previously taken MDA . However , there was no evidence that any biases related to previous MDA or LF diagnosis were different between hotspot residents and adult workers . Our findings raise a number of questions regarding current guidelines and targets used in LF elimination programmes , strategies for post-MDA surveillance , and transmission dynamics in the post-MDA setting . Firstly , our findings indicate that the current WHO recommended TAS has limitations in detecting ongoing transmission in the American Samoa setting . Our current study and previous studies in American Samoa [10 , 19] found evidence of ongoing transmission despite the territory passing TAS-1 in 2010/2011 and TAS-2 in 2015 . In American Samoa , the antigen prevalence threshold used in school-based TAS of young children was not sensitive enough to detect ongoing low-level transmission . Future post-MDA surveillance strategies should consider including older children and adults , and/or determining thresholds that are more specific for different ecological settings [28] . In areas with highly efficient vectors ( such as Ae . polynesiensis in American Samoa ) , LF transmission is likely to be more intense , and might therefore require different elimination targets to successfully interrupt transmission . Furthermore , TAS in American Samoa did not provide any indication of the high antigen prevalence in the Fagali’i hotspot even though >90% of elementary schools were included in the surveys . Our study identified heterogeneity in LF transmission at very small spatial scales , and concur with findings from diverse settings including Samoa [29] , Haiti [30] , Sri Lanka [31] , and Zanzibar [32] . Our results also corroborate findings from Sri Lanka that TAS might not be sensitive enough for identifying small hotspots [31] . However , it is currently unclear whether these small residual foci of transmission will pose any significant risk of resurgence in the broader community , but the presence of microfilaraemic young children in these hotspots suggest that transmission is unlikely to disappear without intervention , particularly in areas with highly efficient vectors and strong environmental drivers of transmission . Secondly , our findings raise concerns that in some settings , seven annual rounds of MDA might not be sufficient for interrupting transmission . Persistent transmission has been noted in Ghana [33] after up to 11 rounds of annual MDA , particularly in areas with high baseline Mf prevalence . In Zanzibar , ongoing transmission was detected six years after MDA , despite good coverage rates and Mf prevalence of <1% at sentinel sites after five rounds of MDA [32] . Thirdly , our results suggest that ICT might not be as sensitive as Og4C3 Ag or Wb123 Ab for detecting low-level transmission or resurgence , such as our hotspot in Ili’ili/Vaitogi/Futiga where Og4C3 Ag and Wb123 Ab provided warning signals , but ICT did not . In Mali , where MDA was conducted from 2002 to 2007 , surveys in children in six formerly highly endemic villages found that ICT prevalence decreased from 53% pre-MDA to 0% ( N = 120 ) after 6 rounds of MDA , and all adults tested in these villages were also antigen negative ( N = 686 ) [34] . However , other longitudinal surveys in these villages using Og4C3 Ag and Wb123 Ab showed an increasing trend of antigen and antibody positivity in 6–7 year old children , from 0% in 2009 to 2 . 7% in 2011 and 4 . 5% in 2013 , with one and three Mf positive children in 2009 and 2011 respectively [35] . Paradoxically , antigen prevalence by ICT in older children ( >8 years ) and adults decreased from 4 . 9% in 2009 to 2 . 8% in 2012 . The results suggest that Og4C3 Ag and Wb123 Ab were more sensitive than ICT , and also raise the suggestion that in formerly highly endemic areas , adults might be immunologically protected while young children are susceptible and more rapidly infected . Further research is being conducted in American Samoa to improve understanding of LF transmission in the post-MDA setting by conducting more representative sampling of all age groups in the general population , comparing the sensitivity of school-based versus community-based surveys , identifying risk factors for infection including the role of migrants , determining the spatial distribution and clustering of infected persons , and exploring the use of molecular xenomonitoring of mosquitoes .
|
Lymphatic filariasis ( LF ) is caused by infection with filarial worms that are transmitted by mosquito bites . The Global Programme to Eliminate Lymphatic Filariasis aims to eliminate the disease as a public health problem by 2020 . Once elimination targets have been reached , cost-effective surveillance strategies are required to ensure that any areas of ongoing transmission or resurgence are quickly identified and managed . Potential options include the integration of LF surveillance with other public health activities . In American Samoa , blood samples collected in 2010 for a research project on a different disease ( leptospirosis ) were used to test for evidence of LF infection , and the study found two possible areas of ongoing transmission . We conducted a follow up study in 2014 to verify whether LF infection was truly occurring in these two areas , and found that infection rates in both areas were significantly higher compared to other parts of American Samoa . Our results therefore provide field evidence that integrating LF surveillance activities with other population-based surveys are potentially effective and feasible , and provide a cost-effective method for identifying residual areas of transmission .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"american",
"samoa",
"geomorphology",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"education",
"landforms",
"sociology",
"geographical",
"locations",
"age",
"distribution",
"social",
"sciences",
"topography",
"tropical",
"diseases",
"parasitic",
"diseases",
"animals",
"social",
"systems",
"filariasis",
"neglected",
"tropical",
"diseases",
"population",
"biology",
"insect",
"vectors",
"lymphatic",
"filariasis",
"islands",
"infectious",
"diseases",
"serology",
"schools",
"disease",
"vectors",
"insects",
"arthropoda",
"people",
"and",
"places",
"population",
"metrics",
"helminth",
"infections",
"mosquitoes",
"eukaryota",
"oceania",
"earth",
"sciences",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"organisms"
] |
2017
|
Detecting and confirming residual hotspots of lymphatic filariasis transmission in American Samoa 8 years after stopping mass drug administration
|
Sensory neurons give highly variable responses to stimulation , which can limit the amount of stimulus information available to downstream circuits . Much work has investigated the factors that affect the amount of information encoded in these population responses , leading to insights about the role of covariability among neurons , tuning curve shape , etc . However , the informativeness of neural responses is not the only relevant feature of population codes; of potentially equal importance is how robustly that information propagates to downstream structures . For instance , to quantify the retina’s performance , one must consider not only the informativeness of the optic nerve responses , but also the amount of information that survives the spike-generating nonlinearity and noise corruption in the next stage of processing , the lateral geniculate nucleus . Our study identifies the set of covariance structures for the upstream cells that optimize the ability of information to propagate through noisy , nonlinear circuits . Within this optimal family are covariances with “differential correlations” , which are known to reduce the information encoded in neural population activities . Thus , covariance structures that maximize information in neural population codes , and those that maximize the ability of this information to propagate , can be very different . Moreover , redundancy is neither necessary nor sufficient to make population codes robust against corruption by noise: redundant codes can be very fragile , and synergistic codes can—in some cases—optimize robustness against noise .
Neurons in sensory systems gather information about the environment , and transmit that information to other parts of the nervous system . This information is encoded in the activity of neural populations , and that activity is variable: repeated presentations of the same stimulus lead to different neuronal responses [1–7] . This variability can degrade the ability of neural populations to encode information about stimuli , leading to the question: which features of population codes help to combat—or exacerbate—information loss ? This question is typically addressed by assessing the amount of information that is encoded in the periphery as a function of the covariance structure [6 , 8–24] , the shapes of the tuning curves [25 , 26] , or both [27 , 28] . However , the informativeness of the population responses at the periphery is not the only relevant quantity for understanding sensory coding; of potentially equal importance is the amount of information that propagates through the neural circuit to downstream structures [29 , 30] . To illustrate the ideas , consider the case of retinal ganglion cells transmitting information about visual stimuli to the cortex via the thalamus , as shown in Fig 1 . To quantify the performance of the retina , one must consider not only the informativeness of the optic nerve responses ( Ix ( s ) in Fig 1A ) , but also how much of that information is transmitted by the lateral geniculate nucleus ( LGN ) to the cortex ( Iy ( s ) in Fig 1A ) [31] . The two may be very different , as only information that survives the LGN’s spike-generating nonlinearity and noise corruption will propagate to downstream cortical structures . Despite its importance , the ability of information to propagate through neural circuits remains relatively unexplored [31] . One notable exception is the literature on how synchrony among the spikes of different cells affects responses in downstream populations [32–36] . This is , however , distinct from the information propagation question we consider here , as there is no guarantee that those downstream spikes will be informative . Other work [25 , 29 , 30 , 37 , 38] investigated the question of optimal network properties ( tuning curves and connection matrices ) for information propagation in the presence of noise . No prior work , however , has isolated the impact of correlations on the ability of population-coded information to propagate . Given the frequent observations of correlations in the sensory periphery [6 , 8 , 17 , 39–45] , and the importance of the information propagation problem , this is a significant gap in our knowledge . To fill that gap , we consider a model ( Fig 1B; described in more detail below ) , in which there are two layers ( retina and LGN , for example ) . The first layer contains a fixed amount of information , Ix ( s ) , which is encoded in the noisy , stimulus-dependent responses of the cells in that layer . The information is passed to the second layer via feedforward connections followed by a nonlinearity , with noise added along the way . We ask how the covariance structure of the trial-to-trial variability in the first layer affects the amount of information in the second . Although we focus on information propagation , the problem we consider applies to more general scenarios . In essence , we are asking: how does the noise in the input to a network interact with noise added to the output ? Because we consider linear feedforward weights followed by a nonlinearity , the possible transformations from input to output , and thus the computations the network could perform , is quite broad [46] . Thus , the conclusions we draw apply not just to information propagation , but also to many computations . Moreover , it may be possible to extend our analysis to recurrent , time-dependent neural networks . That is , however , beyond the scope of this work . Our results indicate that the amount of information that successfully propagates to the second layer depends strongly on the structure of correlated responses in the first . For linear neural gain functions , and some classes of nonlinear ones , we identify analytically the covariance structures that optimize information propagation through noisy downstream circuits . Within the optimal family of covariance structures , we find variability with so-called differential correlations [22]—correlations that are proven to minimize the information in neural population activity . Thus , covariance structures that maximize the information content of neural population codes , and those that maximize the ability of this information to propagate , can be very different . Importantly , we also find that redundancy is neither necessary nor sufficient for the population code to be robust against corruption by noise . Consequently , to understand how correlated neural activity affects the function of neural systems , we must not only consider the impact of those correlations on information , but also the ability of the encoded information to propagate robustly through multi-layer circuits .
We consider a model in which a vector of “peripheral” neural population responses , x , is determined by two components . The first is the set of tuning curves , f ( s ) , which define the cells’ mean responses to any particular stimulus ( typical tuning curves are shown in Fig 2A ) . Here we consider a one dimensional stimulus , denoted s , which may represent , for example , the direction of motion of a visual object . In that case , a natural interpretation of our model is that it describes the transmission of motion information by direction selective retinal ganglion cells to the visual cortex ( Fig 1 ) [5 , 6 , 47] . Extension to multi-dimensional stimuli is straightforward . The second component of the neural population responses , ξ , represents the trial-to-trial variability . This results in the usual “tuning curve plus noise” model , x = f ( s ) + ξ , ( 1 ) where ξ is a zero mean random variable with covariance Σξ . The neural activity , x , propagates to the second layer via feed-forward weights , W , as in the model of [38] . The activity in the second layer is given by passing the input , W · x , through a nonlinearity , g ( ⋅ ) , and then corrupting it with noise , η ( Fig 1B ) , y = g ( W · x ) + η , ( 2 ) where the nonlinearity is taken component by component , and η is zero mean noise with covariance matrix Ση . The function g ( ⋅ ) need not be invertible , so this model can include spike generation . While we have , in Fig 1 , given one explicit interpretation of our model , the model itself is quite general . This means that our results apply more broadly than just to circuits in the peripheral visual system . Moreover , while our analysis ( below ) focuses on information loss between layers , this should not be taken to mean that there is no meaningful computation happening within the circuit: because we have considered arbitrary nonlinear transformations between layers , the same model can describe a wide range of possible computations [46] . Our results apply to information loss during those computations . In the standard fashion [6 , 12 , 20–22] , we quantify the information in the neural responses using the linear Fisher information . This measure quantifies the precision ( inverse of the mean squared error ) with which a locally optimal linear estimator can recover the stimulus from the neural responses [48 , 49] . The linear Fisher information in the first and second layers , denoted Ix ( s ) and Iy ( s ) , respectively , is given by I x ( s ) = f ′ ( s ) · Σ ξ − 1 · f ′ ( s ) ( 3a ) I y ( s ) = f ′ ( s ) · [ Σ ξ + ( W eff T · Σ eff , η − 1 · W eff ) − 1 ] − 1 · f ′ ( s ) ( 3b ) where a prime denotes a derivative . Here Weff are the effective weights—basically , the weights , W , multiplied by the average slope of the gain function , g ( ⋅ ) —and Σeff , η includes contributions from the noise in the second layer , η , and , if g ( ⋅ ) is nonlinear , from the noise in the first layer . ( If g is linear , Σeff , η = Ση , so in this case Σeff , η depends only on the noise in the second layer ) . This expression is valid if W eff T · Σ eff , η - 1 · W eff is invertible; so long as there are more cells in the second layer than the first , this is typically the case . See Methods for details ( section titled “Information in the output layer” ) . Eq ( 3b ) is somewhat intuitive , at least at a gross level: both large effective noise ( Σeff , η ) and small effective weights ( Weff ) reduce the amount of information at the second layer . At a finer level , the relationship between the two covariance structures—corresponding to the first and second terms in brackets in Eq ( 3b ) —can have a large effect on Iy ( s ) , as we will see shortly . We begin with an example to highlight the difference between the information contained in neural population codes and the information that propagates through subsequent layers . Here , we consider two different neuronal populations with identical tuning curves ( Fig 2A ) , nearly-identical levels of trial-to-trial neural variability , and identical amounts of stimulus information encoded in their firing-rate responses; the populations’ correlational structures , however , differ . We then corrupt these two populations’ response patterns with noise , to mimic corruption that might arise in subsequent processing stages , and ask how much of the stimulus information remains . Surprisingly , the two population codes can show very different amounts of information after corruption by even modest amounts of noise ( Fig 2B ) . In more detail , there are 100 neurons in the first layer; those neurons encode an angle , denoted s , via their randomly-shaped and located tuning curves ( Fig 2A ) . We consider two separate model populations . Both have the same tuning curves , but different covariance matrices . For reasons we discuss below , those covariance matrices , denoted Σ ξ blue and Σ ξ green ( blue and green correspond to the colors in Fig 2B and 2C ) , are given by Σ ξ blue = Σ 0 + ϵ f ′ ( s ) f ′ ( s ) ( 4a ) Σ ξ green = Σ 0 + ϵ u u ( s ) u ( s ) ( 4b ) where Σ0 is a diagonal matrix with elements equal to the mean response , Σ 0 , i j = f i ( s ) δ i j . ( 5 ) Here δij is the Kronecker delta ( δij = 1 if i = j and 0 otherwise ) , and we use the convention that two adjacent vectors denote an outer product; for instance , the ijth element if uu is ui uj . The vector u has the same magnitude as f′ , but points in a slightly different direction ( it makes an angle θu with f′ ) , and ϵ and ϵu are chosen so that the information in the two populations , Ix ( s ) , is the same ( ϵu also depends on s; we suppress that dependence for clarity ) . In our simulations , both ϵ and ϵu are small ( on the order of 10−3; see Methods ) , so the variance of the ith neuron is approximately equal to its mean . This makes the variability Poisson-like , as is typically observed when counting neural spikes in finite time windows [1–6] . ( More precisely , the average Fano factors—averaged over neurons and stimuli—were 1 . 01 for the “blue” population and 1 . 04 for the “green” one . ) Both model populations also have the same average correlation coefficients , which are near-zero ( see Methods , section titled “Details for Numerical Examples” ) . To determine how much of the information in the two populations propagates to the second layer , we computed Iy ( s ) for both populations using Eq ( 3b ) . For simplicity , we used the identity matrix for the feed-forward weights , W , a linear gain function , g ( ⋅ ) , and independently and identically distributed ( iid ) noise with variance σ2 . Later we consider the more general case: arbitrary feedforward weights , nonlinear gain functions , and arbitrary covariance for the second layer noise . Those complications don’t , however , change the basic story . Fig 2B shows the information in the output layer versus the level of output noise , σ2 , for the two populations . Blue and green curves correspond to the different covariance structures . Although the two populations have identical tuning curves , nearly-identical levels of trial-to-trial neural variability , and contain identical amounts of information about the stimulus , they differ markedly in the robustness of that information to corruption by noise in the second layer . Thus , quantifying the information content of neural population codes is not sufficient to characterize them: recordings from the first-layer cells of the two example populations in Fig 2 would yield identical information about the stimulus , but the blue population has a greater ability to propagate that information downstream . One possible explanation for the difference in robustness is that the information in the green population relies heavily on correlations , which are destroyed by a small amount of noise . To check this , we compared the information of the correlated neural populations to the information that would be obtained with the same tuning curves and levels of single neuron trial-to-trial variability , but no inter-neuronal correlations [11 , 50 , 51] ( Fig 2C ) . We find that removing the correlations actually increases the information in both populations ( Fig 2C; “Trial-Shuffled” ) , and by about the same amount , so this possible explanation cannot account for the difference in robustness . We also considered the case where the correlated responses carry more information than would be obtained from independent cells . We again found ( similar to Fig 2C ) that there could be substantial differences in the amount of information propagated by equally informative population codes ( see Methods , section titled “Details for numerical examples” , and the figure therein ) . These examples illustrate that merely knowing the amount of information in a population , or how that information depends on correlations in neural responses , doesn’t tell us how much of that information will propagate to the next layer . In the remainder of this paper , we provide a theoretical explanation of this observation , and identify the covariance structures at the first layer that maximize robustness to information loss during propagation through downstream circuits . To understand , from a geometrical point of view , why some population codes are more sensitive to noise than others , we need to consider the relationship between the noise covariance ellipse and the “signal direction , ” f′ ( s ) —the direction the mean neural response changes when the stimulus s changes by a small amount . Fig 3A and 3B show this relationship for two different populations . The noise distribution in the first layer is indicated by the magenta ellipses , and the signal direction by the green arrows . The uncertainty in the stimulus after observing the neural response is indicated by the overlap of the green line with the magenta ellipse . Because the overlap is the same for the two populations , they have the same amount of stimulus uncertainty , and thus the same amount of information—at least in the first layer . Although the two populations have the same amount of information , the covariance ellipses are very different: one long and skinny but slightly tilted relative to the signal direction ( Fig 3A ) , the other shorter and fatter and parallel to the signal direction ( Fig 3B ) . Consequently , when iid noise is added , as indicated by the dashed lines , stimulus uncertainty increases by very different amounts: there’s a much larger increase for the long skinny ellipse than for the short fat one . This makes the population code in Fig 3A much more sensitive to added noise than the one in Fig 3B . To more rigorously support this intuition , in Methods , section titled “Analysis behind the geometry of information loss” , we derive explicit expressions for the stimulus uncertainty in the first and second layers as a function of the angle between the long axis of the covariance ellipse and the signal direction . Those expressions corroborate the phenomenon shown in Fig 3 . The geometrical picture in the previous section tells us that a code is robust against added noise if the covariance ellipse lines up with the signal direction . Taken to its extreme , this suggests that when all the noise is concentrated along the f′ ( s ) direction , so that the covariance matrix is given by Σ ξ ( s ) ∝ f ′ ( s ) f ′ ( s ) , ( 6 ) the resulting code should be optimally robust . While this may be intuitively appealing , the arguments that led to it were based on several assumptions: iid noise added in the second layer , feedforward weights , W , set to the identity matrix , and a linear neural response function g ( ⋅ ) . In real neural circuits , none of these assumptions hold . It turns out , though , that the only one that matters is the linearity of g ( ⋅ ) . In this section we demonstrate that the covariance matrix given by Eq ( 6 ) optimizes information transmission for neurons with linear gain functions ( although we find , perhaps surprisingly , that this optimum is not unique ) . In the next section we consider nonlinear gain functions; for that case the covariance matrix given by Eq ( 6 ) can be , but is not always guaranteed to be , optimal . To determine what covariance structures maximize information propagation , we simply maximize information in the second layer , Iy ( s ) , with respect to the noise covariance matrix in the first layer , Σξ , with the information in the first layer held fixed . When the gain function , g ( ⋅ ) , is linear ( the focus of this section ) , this is relatively straightforward . Details of the calculation are given in Methods , section titled “Identifying the family of optimal covariance matrices”; here we summarize the results . The main finding is that there exists a family of first-layer covariance matrices Σξ , not just one , that maximizes the information in the second layer . That family , parameterized by α , is given by Σ ξ ( s ) = α I x ( s ) I η ( s ) Σ y + 1 - α I x ( s ) f ′ ( s ) f ′ ( s ) , ( 7 ) where Σy is the effective covariance matrix in the second layer , Σ y ≡ ( W eff T · Σ η - 1 · W eff ) - 1 , ( 8 ) and Iη ( s ) is the information the second layer would have if there were no noise in the first layer , I η ( s ) = f ′ ( s ) · Σ y - 1 · f ′ ( s ) ( 9 ) ( see in particular Methods , Eq ( 46 ) ) . For this whole family of distributions—that is , for any value of α for which Σξ is positive semi-definite—the output information , Iy ( s ) , has exactly the same value , I y ( s ) = I x ( s ) 1 + I x ( s ) / I η ( s ) ( 10 ) ( see Methods , Eq ( 76 ) ) . This is the maximum possible output information given the input information , Ix ( s ) . Two members of this family are of particular interest . One is α = 0 , for which the covariance matrix corresponds to differential correlations ( Eq ( 6 ) ) ; that covariance matrix is illustrated in Fig 4A . This covariance matrix aligns the noise direction with the signal direction . Accordingly , as for the geometrical picture in Fig 3 , it makes the encoded information maximally robust . The other family member we highlight is α = 1 , for which Σξ ∝ Σy . For this case , the covariance matrix in the first layer matches the effective covariance matrix in the second layer; we thus refer to this as “matched covariance” . To understand why this covariance optimizes information in the second layer , we start with the observation that the population activities can be decomposed into their principal components: each principal component corresponds to a different axis along with the population activities can be projected . The information contained in each such projection ( principal component ) adds up to give the total Fisher information ( see Methods , Eq 71 ) . The most informative of these projections are those that have low noise variance , and which align somewhat with the signal curve—like the blue line in Fig 4B . When Σξ ∝ Σy , the projections that are most informative in the first layer are corrupted by relatively little noise in the second layer . Consequently , this configuration enables robust information propagation . In contrast , when the covariance structures in the first and second layers are less well matched , all projections are heavily corrupted by noise at some point ( i . e . , either in the first or the second layer ) , and hence very little information propagates ( Fig 4C ) . The family of optima interpolates between the two configurations shown in Fig 4A and 4B ( see also Eq ( 7 ) ) . Almost all members of this optimal covariance family depend on the details of the downstream circuit: for α ≠ 0 in Eq ( 7 ) , the optimal noise covariance at the first layer depends on the feed-forward weights , W , and the structure of the downstream noise . The one exception to this is the covariance matrix given by Eq ( 6 ) : that one is optimal regardless of the downstream circuit . These are so-called “differential correlations”—the only correlations that lead to information saturation in large populations [22] , and the correlations that minimize information in general ( see Methods , section titled “Minimum information” , for proof ) . The fact that correlations can minimize information content and at the same time maximize robustness highlights the fact that optimizing the amount of information in a population code versus optimizing the ability of that information to be transmitted put very different constraints on neural population codes . The existence of an optimum where the covariance matrices are matched across layers emphasizes that not all optimally robust population codes are necessarily redundant . ( By redundant we mean the population encodes less information than would be encoded by a population of independent cells with the same tuning curves and levels of single neuron trial-to-trial variability [12 , 21]; see Fig 2 ) . Notably , if the effective second layer covariance matrix , Σy , admits a synergistic population code—wherein more information is encoded in the correlated population versus an uncorrelated one with the same tuning curves and levels of trial-to-trial response variability—then the matched case , Σξ ∝ Σy , will also admit a synergistic population code , and be optimally robust . Optimally robust , however , does not necessarily mean the majority of the information is transmitted; for that we need another condition . We show in the Methods section titled “Variances of neural responses , and robustness to added noise , for different coding strategies” that for non-redundant codes , a large fraction of the information is transmitted only if there are many more neurons in the second layer than in the first . This is typically the case in the periphery . For differential correlations , that condition is not necessary—so long as there are a large number of neurons in both the input and output layers , most of the information is transmitted . So far we have focused on linear gain functions g ( ⋅ ) ; here we consider nonlinear ones . This case is much harder to analyze , as the effective covariance structure in the second layer , Σeff , η , depends on the noise in the first layer ( see Methods , Eq ( 22 ) ) . We therefore leave the analysis to Methods ( section titled “Nonlinear gain functions” ) ; here we briefly summarize the main results . After that we consider two examples of nonlinear gain functions—both involving a thresholding nonlinearity to mimic spike generation . For linear gain functions we were able to find a whole family of optimal covariance structures , for nonlinear ones we did not even try . Instead , we asked: under what circumstances are differential correlations optimal ? Even for this simplified question a definitive answer does not appear to exist . Nevertheless , we can make progress in special cases . When there is no added noise in the second layer ( e . g . , η = 0 for the model in Fig 1B ) , differential correlations maximize the amount of information that propagates through the nonlinearity , so long as the tuning curves are sufficiently dense relative to the steepness of the tuning curves ( meaning that whenever the stimulus changes , the average stimulus-evoked response of at least one neuron also changes; see Methods ) . If there is added noise at the second layer , differential correlations tend to be optimal in cases where the addition of noise at the first layer , ξ , causes reductions in information , Ix ( s ) . ( This means that , so long as there are no stochastic resonance effects causing added noise to increase information , then differential correlations are optimal . ) We first check , with simulations , the prediction that differential correlations are optimal if there is no added noise . For that we use a thresholding nonlinearity , chosen for two reasons: it is an extreme nonlinearity , and so should be a strong test of our theory , and it is somewhat realistic in that it mimics spike generation . For this model , the responses at that second layer , yi , are given by y i = Θ ( x i - θ i ) ( 11 ) where Θ is the Heaviside step function ( Θ ( x ) = 1 if x ≥ 0 and 0 otherwise ) , and θi is the spiking threshold of the ith neuron . This is the popular dichotomized Gaussian model [52–56] , which has been shown to provide a good description of population responses in visual cortex , at least in short time windows [54] , and to provide high-fidelity descriptions of the responses of integrate-and-fire neurons , again in short time windows [57] . In our simulations with the step function nonlinearity , as for all of the other cases we considered above , the first layer responses are given by the tuning curve plus noise model ( Eq ( 1 ) ) . The tuning curves , f ( s ) , of the 100-neuron population are again heterogeneous ( similar to those in Fig 2A but with a different random draw from the tuning curve distribution ) , and the trial-to-trial variability is given by Σ ξ = γ u Σ 0 + ϵ u u ( s ) u ( s ) ( 12 ) with Σ0 given by Eq ( 5 ) . This is the same covariance matrix as in Eq ( 4b ) , except that we have included an overall scale factor , γu , chosen to ensure that the information in the input layer is independent of both ϵu and u ( s ) ( see Methods , Eq ( 99 ) ) . Because these ( step function ) nonlinearities are infinitely steep , the tuning curves are not sufficiently dense for our mathematical analysis to guarantee that differential correlations are optimal for information propagation . However , we argue in Methods ( section titled “Nonlinear gain functions” ) , that this should be approximately true for large populations . And indeed , that’s what we find with our numerical simulation , as shown in Fig 5B . When θu = 0 ( recall that θu is the angle between u ( s ) and f′ ( s ) ) , so that u ( s ) = f′ ( s ) , the second term in Eq ( 12 ) corresponds to differential correlations; in this case , information increases monotonically with ϵu . In other words , information propagated through the step function nonlinearity increases as “upstream” correlations become more like pure differential correlations . In contrast , when θu is nonzero ( as in Fig 3A ) , information does not propagate well: information decreases as ϵu increases . This is consistent with our findings for the linear gain function considered in Fig 2 . Thus , differential correlations can optimize information transmission even for a nonlinearity as extreme as a step function . The lack of explicit added noise at the second layer makes this case somewhat unrealistic . In neural circuits , we expect noise to be added at each stage of processing—if nothing else , due to synaptic failures . We thus considered a model in which noise is added before the spike-generation process , y i = Θ ( x i + ζ i - θ i ) ( 13 ) where ζi is zero-mean noise with covariance matrix Σζ . We computed information for this model using the same input tuning curves , spike thresholds , and covariance matrix , Σξ , as without the additional noise ( i . e . , as in Fig 5 ) . To mimic the kind of independent noise expected from synaptic failures , we chose the ζi to be iid , and for simplicity we took them to be Gaussian distributed with variance σ ζ 2 . We computed the amount of stimulus information , Iy ( s ) , for several different levels of the added input noise σ ζ 2 . We found that for all levels of noise , differential correlations increase information transmission ( Iy ( s ) increases monotonically with ϵu in Fig 6A , for which θu = 0 ) . And we again found that when the long axis of the covariance ellipse makes a small angle with the signal direction , information propagates poorly ( Fig 6B , for which θu = 0 . 1 rad . ) . These numerical findings for a spike-generating nonlinearity with added noise are similar to the previous cases of a linear transfer function , g ( ⋅ ) , with added input noise ( Figs 2 and 3 ) , for which we have analytical results , or a spike generating nonlinearity with no added input noise ( Fig 5 ) , for which we do not . We further argue in Methods ( section titled “Nonlinear gain functions” ) , that for nonlinear gain functions differential correlations are likely to be optimal if the tuning curves are optimal ( in the case of Eq ( 13 ) , if the thresholds θi are chosen optimally ) . Taken together , our findings demonstrate that differential correlations in upstream populations generally increase the information that can be propagated downstream through noisy , nonlinear neural circuits .
Much work in systems neuroscience has investigated the factors that influence the amount of information about a stimulus that is encoded in neural population activity patterns . Here we addressed a related question that is often overlooked: how do correlations between neurons affect the ability of information to propagate robustly through subsequent stages of neural circuitry ? The question of robustness is potentially quite important , as the ability of information to propagate determines how much information from the periphery will reach the deeper neural structures that affect decision making and behavior . To investigate this issue , we considered a model with two cell layers . We varied the covariance matrix of the noise in the first layer ( while keeping the tuning curves and information in the first layer fixed ) , and asked how much information could propagate to the second layer . Our main findings were threefold . First , population codes with different covariance structures but identical tuning curves and equal amounts of encoded information can differ substantially in their robustness to corruption by additional noise ( Figs 2 , 5 , 6 and 7 ) . Consequently , measurements of information at the sensory periphery are insufficient to understand the ability of those peripheral structures to propagate information to the brain , as that propagation process inevitably adds noise . For instance , populations of independent neurons can be much worse at transmitting information than can populations displaying correlated variability ( Fig 5B ) . Thus , to understand how the brain efficiently encodes information , we must concern ourselves not just with the amount of information in a population code , but also with the robustness of that encoded information against corruption by noise . Second , for linear gain functions , or noise-free nonlinear ones with sufficiently dense tuning curves , populations with so-called differential correlations [22] are maximally robust against noise induced by information propagation . This fact may seem surprising given that differential correlations are the only ones that lead to information saturation in large populations [22] , and the correlations that minimize information in general . However , in hindsight it makes sense: differential correlations correspond to a covariance ellipse aligned with the signal direction ( see Fig 3B ) , and added noise simply doesn’t make it much longer . For nonlinear gain functions combined with arbitrary noise , differential correlations are not guaranteed to yield a globally optimal population code for information propagation . However , for the spike-generating nonlinearity we considered here , differential correlations were at least a local optimum ( see Figs 5 and 6 ) . Third , while differential correlations optimize robustness , for linear gain functions that optimum is not unique . Instead , there is a continuous family of covariances that exhibit identical robustness to noise ( see Fig 4 and Eq ( 7 ) ) . However , within this family , only differential correlations yield population codes that are optimally robust independent of the downstream circuitry . Thus , they are the most flexible of the optima: for all other members of the family , the optimal covariance structure in the first layer depends on the noise in subsequent layers , as well as the weights connecting those layers . The existence of this family of optimal solutions raises an important point with regards to redundancy and robust population coding . Populations with differential correlations—which are among the optimal solutions in terms of robustness—are highly redundant: a population with differential correlations encodes much less information than would be expected from independent populations with the same tuning curves and levels of trial-to-trial variability ( Fig 2C ) . It is common knowledge that redundancy can enhance robustness of population codes against noise [58] , and thus it is worth asking if our robust population coding results are simply an application of this fact . Importantly , the answer is no: as discussed in Methods , section titled “A family of optimal noise structures” , within the family of optimal correlational structures are codes with minimal redundancy . Moreover , as is shown in Fig 2B , a code can be redundant without being robust to added noise . In other words , redundancy in a population code is neither necessary , nor sufficient , to ensure that the encoded information is robust against added noise . However , there is an important caveat: unless the number of neurons in the second layer is large relative to the number in the first , and/or the added noise in the second layer is small relative to the noise at the first layer , non-redundant codes tend to lose a large amount of information when corrupted by noise . This contrasts sharply with differential correlations , which can tolerate large added noise with very little information loss ( see Methods section titled “Variances of neural responses , and robustness to added noise , for different coding strategies” ) . In the case of real neural systems , there will always be a finite amount of information that the population can convey ( bounded by the amount of input information that the population receives from upstream sources [59] ) , and so the question of how best to propagate a ( fixed ) amount of information is of potentially great relevance for neural communication . Our results suggest that the presence of differential correlations serves to allow population-coded information to propagate robustly . Thus , an observation of these correlations in neural recordings might indicate that the population code is optimized for robustness of the encoded information . At the same time , we note that weak differential correlations might be hard to observe experimentally [22] . Moreover , our calculations indicate that there exists a whole family of possible propagation-enhancing correlation structures , and so differential correlations are not necessary for robust information propagation . This means that observations of either differential correlations , correlation structures matched between subsequent layers of a neural circuit ( Fig 4 ) , or a combination of the above would indicate that the system enables robust information propagation . How might the nervous system shape its responses so as to generate correlations that enhance information propagation ? Recent work identified network mechanisms that can lead to differential correlations [60] . While it is beyond the scope of this work , it would be interesting to explicitly study the network structures that allow encoded information to propagate most robustly through downstream circuits . Relatedly , [38] and [29 , 30] asked how the connectivity between layers affects the ability of information to propagate . While we identified the optimal patterns of input to the multi-stage circuit , they identified the optimal anatomy of that circuit itself . Note that we have used linear Fisher information to quantify the population coding efficacy . Other information measures exist , and it is worth commenting on how much our findings generalize to different measures . In the case of jointly Gaussian stimulus and response distributions , correlations that maximize linear Fisher information also maximize Shannon’s mutual information [20] . In that regime our findings should generalize well . Moreover , whenever the neural population response distributions belong to the exponential family with linear sufficient statistics , the linear Fisher information is equivalent to the ( nonlinear ) “full” Fisher information [29] . In practice , this is a good approximation to primary visual cortical responses to oriented visual stimuli [61 , 62] , and to other stimulus-evoked responses in other brain areas ( see [22] for discussion ) . Consequently , our use of linear Fisher information in place of other information measures is not a serious limitation . For encoded sensory information to be useful , it must propagate from the periphery to the deep brain structures that guide behavior . Consequently , information should be encoded in a manner that is robust against corruption that arises during propagation . We showed that the features of population codes that maximize robustness can be substantially different from those that maximize the information content in peripheral layers . Moreover , by elucidating the set of covariances structures that optimize information transmission , we found that redundancy in a population code is neither necessary , nor sufficient , to guarantee robust propagation . In future work , it will be important to determine whether the nervous system uses the class of population codes that maximize information transmission . Finally , while our main focus was on information propagation , the model we used—linear feedforward weights followed by a nonlinearity—is known to have powerful computational properties [46] . It is , in fact , the basic unit in many deep neural networks . Thus , our main conclusion , which is that differential correlations are typically optimal , applies to any computation that can be performed by this architecture .
Our analysis focuses on information loss through one layer of circuitry; to compute the loss , we need expressions for the linear Fisher information in the first and second layers . Expressions for those two quantities are given in Eqs ( 3a ) and ( 3b ) . The first is standard; here we derive the second . To make the result as general as possible , we include noise inside the nonlinearity as well as outside it; if nothing else , that’s probably a reasonable model for the spiking nonlinearity given in Eq ( 13 ) . We thus generalize slightly Eq ( 2 ) , and write y = g ( W · x + ζ ) + η ( 14 ) where ζ is zero mean noise with covariance matrix Σζ , and here and in what follows we use the convention that g is a pointwise nonlinearity , so for any vector v , the ith element of g ( v ) is g ( vi ) . When Σζ = 0 , we recover exactly the model in Eq ( 2 ) . Using Eq ( 1 ) for x , Eq ( 14 ) becomes y = g h ( s ) + W · ξ + ζ + η ( 15 ) where , recall , ξ and η are zero mean noise with covariance matrices Σξ and Ση , respectively , and h ( s ) is the mean drive to neuron i , h ( s ) ≡ W · f ( s ) . ( 16 ) To compute the linear Fisher information in the second layer , we start with the usual expression , I y ( s ) = ∂ E ( y | s ) ∂ s · Cov [ y | s ] - 1 · ∂ E ( y | s ) ∂ s ( 17 ) where E and Cov denote mean and covariance , respectively . The mean value of y given s is , via Eq ( 15 ) ) , E [ y | s ] = E ξ , ζ g h ( s ) + W · ξ + ζ ≡ g ¯ h ( s ) . ( 18 ) Like g ( ⋅ ) , g ¯ ( · ) is a pointwise nonlinearity . To compute the covariance , we assume , as in the main text , that ξ and η are independent; in addition , we assume that both are independent of ζ . Thus , the covariance of y is the sum of the covariances of the first and second terms in Eq ( 15 ) . The covariance of the second term is just Ση . The covariance of the first term is harder . To make progress , we start by implicitly defining the quantity δΣg ( s ) via Cov g h ( s ) + W · ξ + ζ ≡ δ Σ g ( s ) + W eff ( s ) · Σ ξ · W eff T ( s ) + G ¯ ′ ( s ) · Σ ζ ( s ) · G ¯ ′ ( s ) ( 19 ) where Weff ( s ) is the actual feedforward weight multiplied by the average slope of g , W eff , i j ( s ) ≡ g ¯ ′ h i ( s ) W i j , ( 20 ) and G ¯ ′ ( s ) is the a diagonal matrix with entries corresponding to the average slope of g , G ¯ i j ′ ( s ) ≡ g ¯ ′ h i ( s ) δ i j . ( 21 ) As in the main text , δij is the Kronecker delta and a prime denotes a derivative . The above implicit definition of δΣg is motivated by the observation that when g is linear , δΣg vanishes . Below , in Sec . , we show that if ξ is Gaussian , δΣg is positive semi-definite . Here we assume that the noise is sufficiently close to Gaussian that δΣg remains positive semi-definite , and thus can be treated as the covariance matrix of an effective noise source . This last assumption is needed below , in the section titled “Nonlinear gain functions” , where we argue that information loss is small when δΣg is small ( see text following Eq ( 64 ) ) . Making the additional definition Σ eff , η ( s ) ≡ G ¯ ′ ( s ) · Σ ζ ( s ) · G ¯ ′ ( s ) + δ Σ g ( s ) + Σ η , ( 22 ) and using Eqs ( 15 ) and ( 19 ) and the fact that η is independent of both ξ and ζ , we see that Cov [ y | s ] = W eff ( s ) · Σ ξ · W eff T ( s ) + Σ eff , η ( s ) . ( 23 ) Combining this with the expression for the mean value of y , Eq ( 18 ) , the linear Fisher information , Eq ( 17 ) becomes I y = f ′ · W eff T · W eff · Σ ξ · W eff T + Σ eff , η - 1 · W eff · f ′ ( 24 ) where we used Eqs ( 16 ) and ( 20 ) to replace ∂sE ( y|s ) with Weff · f′ and , to reduce clutter , we have suppresed any dependence on s . To pull the effective weights inside the inverse , we use the Woodbury matrix identity to write W eff T · [ W eff · Σ ξ · W eff T + Σ eff , η ] − 1 · W eff = W eff T · Σ eff , η − 1 · W eff − W eff T · Σ eff , η − 1 · W eff · [ Σ ξ − 1 + W eff T · Σ eff , η − 1 · W eff ] − 1 · W eff T · Σ eff , η − 1 · W eff . ( 25 ) Then , using the fact that [A + B]−1 = A−1 · [A−1 + B−1]−1 · B−1 , and applying a very small amount of algebra , this becomes W eff T · [ W eff · Σ ξ · W eff T + Σ eff , η ] − 1 · W eff = W eff T · Σ eff , η − 1 · W eff · [ I − Σ ξ · [ Σ ξ + ( W eff T · Σ eff , η − 1 · W eff ) − 1 ] − 1 ] ( 26 ) where I is the identity matrix . It is then straightforward to show that W eff T · W eff · Σ ξ · W eff T + Σ eff , η - 1 · W eff = Σ ξ + ( W eff T · Σ eff , η - 1 · W eff ) - 1 - 1 . ( 27 ) Inserting this into Eq ( 24 ) , we see that the right hand side of that equation is equal to the expression given in Eq ( 3b ) of the main text . Here we address the question: what noise covariance matrix optimizes information transmission ? In other words , what covariance matrix Σξ maximizes the information given in Eq ( 3b ) ? That is hard to answer when g is nonlinear , because in that case Σeff , η depends on Σξ via δΣg ( see Eqs ( 19 ) and ( 22 ) ) . In this section , then , we consider linear gain functions; in the next we consider nonlinear ones . To make our expressions more readable , we generally suppress the dependence on s . Our goal is to maximize Iy with Ix fixed . Using the definition of Σy given in Eq ( 8 ) , for linear gain functions the information in the second layer ( Eq ( 3b ) ) is written I y = f ′ · Σ ξ + Σ y - 1 · f ′ . ( 35 ) We use Lagrange multipliers , ∂ ∂ Σ ξ f ′ · [ Σ ξ + Σ y ] - 1 · f ′ - λ f ′ · Σ ξ - 1 · f ′ - I x = 0 , ( 36 ) where λ is a Lagrange multiplier that enforces the constraint f ′ · Σ ξ - 1 · f ′ = I x . Taking the derivative and setting it to zero yields [ Σ ξ + Σ y ] - 1 · f ′ f ′ · [ Σ ξ + Σ y ] - 1 = λ Σ ξ - 1 · f ′ f ′ · Σ ξ - 1 . ( 37 ) In deriving this expression we used the fact that the gain functions are linear , which implies that Σy does not depend on Σξ . Multiplying by Σξ + Σy on both the left and right , we arrive at f ′ f ′ = λ [ I + Σ y · Σ ξ - 1 ] · f ′ f ′ · [ I + Σ ξ - 1 · Σ y ] . ( 38 ) This is satisfied when Σ y · Σ ξ - 1 · f ′ ∝ f ′ . ( 39 ) There are two ways this can happen , Σ y · Σ ξ − 1 ∝ I ( 40a ) Σ y · Σ ξ − 1 ∝ f ′ a ( 40b ) where a is an arbitrary vector . Combining these linearly , taking into account that Σξ is a covariance matrix and thus symmetric , and enforcing equality in Eq ( 38 ) , we arrive at Σ ξ - 1 = I x α I η Σ y - 1 + ( α - 1 ) I x α I η 2 Σ y - 1 · f ′ f ′ · Σ y - 1 + P · Ω - 1 · P ( 41 ) where Iη is the information the output layer would have if there was no noise in the input layer , I η ( s ) = f ′ ( s ) · Σ y - 1 · f ′ ( s ) ( 42 ) ( this is the same expression as in Eq ( 9 ) , it’s repeated here for convenience ) , Ω is an arbitrary symmetric matrix , P is a projection operator , chosen so that P · f′ = 0 , P ≡ I - f ′ f ′ f ′ · f ′ , ( 43 ) and α is arbitrary ( but subject to the constraint that Σξ has no negative eigenvalues ) . Note that P is a linear combination of the right hand sides of Eqs ( 40a ) and ( 40b ) ) , with a = f′ in the latter equation . It is straightforward to verify that when Σξ is given by Eqs ( 41 ) and ( 38 ) is satisfied . To find an explicit expression for Σξ , not just its inverse , we apply the Woodbury matrix identity to Eq ( 41 ) ; that gives us Σ ξ = Σ α - Σ α · P · Ω + P · Σ α · P - 1 · P · Σ α ( 44 ) where Σ α ≡ I x α I η Σ y - 1 + ( α - 1 ) I x α I η 2 Σ y - 1 · f ′ f ′ · Σ y - 1 - 1 = α I η I x Σ y - ( α - 1 ) f ′ f ′ α I η . ( 45 ) Inserting this into Eq ( 44 ) , we arrive at Σ ξ = α I η Σ y I x + ( 1 - α ) f ′ f ′ I x - α I η I x 2 Σ y · P · Ω + α I η I x P · Σ y · P - 1 · P · Σ y . ( 46 ) This is the same as Eq ( 7 ) in the main text , except in that equation we let Ω go to ∞ , so we ignore the projection-related term . Ignoring that term is reasonable , as it just puts noise in a direction perpendicular to f′ , and so has no effect on the information . By choosing different scalars α and matrices Ω , a family of optimal Σξ is obtained . These all have the same input information , Ix , and the same output information , Iy , after corruption by noise . An especially interesting covariance matrix is found in the limit α = 0 , in which case Σ ξ = f ′ f ′ I x . ( 47 ) These are so-called differential correlations [22] . Importantly , the choice α = 0 is the only one for which the optimal correlational structure is independent of the correlations in the output layer , Σy . Note that pure differential correlations don’t satisfy Eq ( 39 ) . As such , they represent a singular limit , in the sense that Σξ in Eq ( 46 ) satisfies Eq ( 39 ) with alpha arbitrarily small , but not precisely zero . The other covariance that we highlight in the text is found for α = 1 and Ω → ∞ , in which case Σ ξ = α I η I x Σ y . This is the matched covariance case . We now focus on differential correlations , and determine conditions under which they are optimal for information propagation when the gain function , g ( ⋅ ) , is nonlinear . In this regime , the effective noise in the second layer ( the second term in brackets in Eq ( 3b ) ) depends on Σξ . This greatly complicates the analysis , and to make headway we need to reformulate our mathematical description of differential correlations . This reformulation is based on the observation that differential correlations correspond to trial-to-trial variability in the value of the stimulus , s[22] . Consequently , the encoding model in the input layer can be written as a multi-step process , s = s 0 + δ s ( 48a ) x = f ( s ) + ξ ( s ) ( 48b ) y = g ( W · x ( s ) + ζ ) + η ( s ) . ( 48c ) Here s0 is the value of the stimulus that is actually presented . However , the neurons in the input layer , x , encode s—a corrupted version of s0 . This is indicated by Eq ( 48a ) , which tells us that s deviates on a trial-to-trial basis from s0 , with deviations that are described by a zero-mean random variable , δs . To see that this model does indeed exhibit differential correlations , we Taylor expand Eq ( 48b ) around s0 , yielding a model of the form x ≈ f ( s 0 ) + f ′ ( s 0 ) δ s + ξ ( s 0 ) , ( 49 ) for which the covariance matrix is Cov [ x ] = Var [ δ s ] f ′ ( s 0 ) f ′ ( s 0 ) + Cov [ ξ | s 0 ] . ( 50 ) The first term corresponds to differential correlations . Eqs ( 48b ) and ( 48c ) correspond exactly to our previous model ( Eq ( 4a ) ) . Consequently , the information about s in the first and second layers are still given by Eqs ( 3a ) and ( 3b ) of the main text . However , we can’t use those equations for the information about s0 . For that , we focus on the variance of its optimal estimator given x , which we denote s ^ 0 . Because of the Markov structure of our model ( s0 ↔ s ↔ x ) , we can construct s ^ 0 by first considering the optimal estimator of s0 given s , and then the optimal estimator of s given x . The variance of s ^ 0 given x is then simply the sum of the variances of these two ( independent ) noise sources . The optimal estimator of s0 given s is simply s , with conditional variance Var [ s ^ 0 ( s ) | s 0 ] = Var [ δ s ] . The optimal estimator of s given x is s ^ ( x ) , with variance Var [ s ^ ( x ) | s ] . Consequently , Var [ s ^ 0 | s 0 ] = Var [ δ s ] + ∫ d s P ( s | s 0 ) Var [ s ^ ( x ) | s ] . ( 51 ) As usual , we approximate the variance of s ^ ( x ) given s by the linear Fisher information , yielding an approximation for the total Fisher information about s0 given x , 1 I x tot ( s 0 ) = Var [ δ s ] + ∫ d s P ( s | s 0 ) I x ( s ) . ( 52 ) Similarly , the Fisher information about s0 given y is approximated by 1 I y tot ( s 0 ) = Var [ δ s ] + ∫ d s P ( s | s 0 ) I y ( s ) . ( 53 ) Note that we are slightly abusing notation here: above , Ix ( s ) and Iy ( s ) referred to the total information about the stimulus; now they refer to the information about the stimulus that is encoded in the first layer , which is different from the actual stimulus , s0 . However , it is a convenient abuse , as it allows us to take over our previous results without introducing much new notation . Our first step is to parametrize the covariance matrix , ξ , and Var[δs] , in a way that ensures that the information in the first layer I x tot ( s 0 ) remains fixed while we vary ξ and Var[δs] . A convenient choice is Var [ δ s ] = 1 I x tot ∫ d s P ( s | s 0 ) ϵ I 0 ( s ) 1 + ϵ I 0 ( s ) ( 54a ) Σ ξ ( s ) = 1 I x tot I 0 ( s ) Σ 0 ( s ) 1 + ϵ I 0 ( s ) , ( 54b ) where I 0 ( s ) ≡ f ′ ( s ) · Σ 0 - 1 ( s ) · f ′ ( s ) . ( 55 ) Inserting Eq ( 54 ) into Eq ( 52 ) , we see that I x tot ( s 0 ) = I x tot , independent of Σ0 ( s ) . The information in the second layer about s , Iy ( s ) , is given by Eq ( 3b ) , with Σeff , η given in Eq ( 22 ) . It is convenient to make the definition Σ eff , y ≡ ( W eff T · Σ eff , η - 1 · W eff ) - 1 . ( 56 ) This is the analog of Eq ( 8 ) , but for nonlinear gain functions . It is clear from Eqs ( 22 ) and ( 19 ) that Σeff , η depends on Σξ; consequently , it depends on ϵ . To maximize information with respect to ϵ , we take a two step approach . We write I y ( s ; ϵ , ϵ 0 ) ≡ f ′ T ( s ) Σ ξ ( s , ϵ ) + Σ eff , y ( s , ϵ 0 ) - 1 f ′ ( s ) . ( 57 ) Here Σξ ( s , ϵ ) and Σy ( s , ϵ0 ) are the same as in Eqs ( 54b ) and ( 56 ) ; we have just made the dependence on ϵ explicit . The two steps are to maximize first with respect to ϵ , then with respect to ϵ0 . If the two maxima occurr in the same place , then we have identified the covariance structure that optimizes information transmission . In the first step we differentiate I y tot ( s ; ϵ , ϵ 0 ) with respect to ϵ . To simplify the expressions , we make the definition Σ tot ( s , ϵ , ϵ 0 ) ≡ Σ ξ ( s , ϵ ) + Σ eff , y ( s , ϵ 0 ) . ( 58 ) Combining Eqs ( 53 ) , ( 54 ) and ( 57 ) , we have ∂ ∂ ϵ 1 I y tot ( s 0 ; ϵ , ϵ 0 ) = 1 I x tot ∫ d s P ( s | s 0 ) I 0 1 + ϵ I 0 2 + ∫ d s P ( s | s 0 ) I y 2 f ′ · Σ tot - 1 · ∂ Σ ξ ( s , ϵ ) ∂ ϵ · Σ tot - 1 · f ′ ( 59 ) where we used the fact that for any square matrix A ( x ) , ( d/dx ) A−1 = −A−1 · d A/dx · A−1 , and we suppressed much of the s , ϵ and ϵ0 dependence for clarity . Using Eq ( 54b ) for Σξ ( s , ϵ ) , the derivative with respect to ϵ in the second term is straightforward , ∂ ∂ ϵ 1 I y tot ( s 0 ; ϵ , ϵ 0 ) = 1 I x tot ∫ d s P ( s | s 0 ) I 0 ( 1 + ϵ I 0 ) 2 − ∫ d s P ( s | s 0 ) I y 2 f ′ · Σ tot − 1 · I 0 2 Σ 0 I x tot ( 1 + ϵ I 0 ) 2 · Σ tot − 1 · f ′ = 1 I x tot ∫ d s P ( s | s 0 ) I 0 2 I y 2 ( 1 + ϵ I 0 ) 2 [ I y 2 I 0 − f ′ · Σ tot − 1 · Σ 0 · Σ tot − 1 · f ′ ] ( 60 ) Then , applying the definition I y ( s ) = f ′ · Σ tot - 1 · f ′ ( see Eqs ( 57 ) and ( 58 ) ) , and making the new definition V ≡ f ′ · Σ tot - 1 · Σ 0 1 / 2 , ( 61 ) we arrive at the expression ∂ ∂ ϵ 1 I y tot ( s 0 ; ϵ , ϵ 0 ) = 1 I x tot ∫ d s P ( s | s 0 ) I 0 2 I y 2 ( 1 + ϵ I 0 ) 2 V · Σ 0 - 1 / 2 f ′ f ′ T Σ 0 - 1 / 2 I 0 - I · V . ( 62 ) The right hand side of Eq ( 62 ) is negative or zero if the term in brackets is negative semi-definite; that is , if all its eigenvalues are non-positive . Since the term in square brackets is a rank one matrix minus the identity , all but one of its eigenvalues are equal to -1 . The remaining eigenvalue is 0 , with corresponding eigenvector Σ 0 - 1 / 2 · f ′ ( see Eq ( 55 ) ) . Thus , ∂ ( 1 / I y tot ( s 0 ; ϵ , ϵ 0 ) / ∂ ϵ ) ≤ 0 , and I y tot ( s 0 ; ϵ , ϵ 0 ) must have a global maximum at ϵ = ∞ . If g is linear , Σeff , y doesn’t depend on ϵ0 , and ϵ = ∞ corresponds to pure differential correlations . We have , therefore , recovered the α = 0 limit of Eq ( 46 ) . When ϵ = ∞ , Σξ vanishes , and so the expression for the information in the second layer simplifies considerably . Combining Eqs ( 53 ) and ( 54a ) , we have , in the ϵ → ∞ limit , 1 I y tot ( s 0 ; ∞ , ϵ 0 ) = 1 I x tot + ∫ d s P ( s | s 0 ) I y ( s ; ∞ , ϵ 0 ) ( 63 ) where I y ( s ; ∞ , ϵ 0 ) = f ′ ( s ) · W eff T ( s ; ϵ 0 ) · Σ η + G ¯ ′ ( s ) · Σ ζ ( s ) · G ¯ ′ ( s ) + δ Σ g ( s ; ϵ 0 ) - 1 · W eff ( s ; ϵ 0 ) · f ′ ( s ) . ( 64 ) The latter equation follows by combining the fact that Σξ ( s , ∞ ) = 0 ( Eq ( 54b ) ) with the definitions of Σeff , y and Σeff , η ( Eqs ( 56 ) and ( 22 ) , respectively ) . The total information in the output layer is maximized when Iy ( s0; ∞ , ϵ0 ) is maximized . That quantity depends on ϵ0 via Σξ ( s , ϵ0 ) , the noise covariance in the input layer . As can be seen from Eq ( 54b ) , larger ϵ0 implies smaller Σξ ( s , ϵ0 ) . That has two effects . First , when Σξ ( s , ϵ0 ) is small enough , the covariance matrix δΣg becomes small ( see Eq ( 19 ) , and note that δΣg is positive definite , as shown in Sec . ) . This tends to increase Iy ( s ) . However , the effective tuning curves , Weff ( s;ϵ ) · f ( s ) , also depend on Σξ ( s , ϵ0 ) ( see Eq ( 20 ) ) . It is possible that increasing Σξ ( s , ϵ0 ) modifes the tuning curves such that Iy ( s ) increases . Consequently , it is impossible to make completely general statements . Nevertheless , we can identify two regimes . First , if there is no added noise in the output layer ( η = ζ = 0 ) , then Iy ( s; ∞ , ϵ ) goes to ∞ as ϵ0 goes to ∞ , thus maximizing the total information . This holds , however , only if the tuning curves are sufficiently dense relative to the steepness of the tuning curves; otherwise , the Fisher information is no longer a good approximation to the true information . For smooth tuning curves this is generally satisfied , but it is not satisfied for the noise-free spike generating mechanism we consider in the main text ( Eq ( 11 ) ) , since for that nonlinearity f′ ( s ) = 0 with probability 1 . We expect , though , that in the absence of noise , this particular nonlinearity introduces an error that is O ( 1 / n ) , implying that Iy ( s; ∞ , ϵ ) ∝ n2 . Numerical simulations corroborated this scaling . Thus , for sufficiently large populations , differential correlations are optimal for the noise-free spike-generating nonlinearity . Note , though , that the thresholds must be chosen so that there are always both active and silent neurons; otherwise , in the limit that Σξ vanishes , the activity will contain no information at all about the stimulus . The second regime is one in which the tuning curves have been optimized . In this case , modifying the tuning curves by adding noise decreases information , and again differential correlations optimize information transmission . To summarize , we have analyzed the scenario considered in the main text ( section titled “Nonlinear gain functions” ) —namely , the neural activities at the second layer , y , are given by a nonlinear function of the neural activities at the first layer , x , with noise added both before and after the nonlinearity . In this case , whether or not differential correlations in the first layer optimize information transmission depends on the details . They do if g is linear , the tuning curves are optimal , or there is no added noise in the second layer and the tuning curves are sufficiently dense relative to the steepness of the tuning curves . If none of these are satisfied , however , differential correlations may be sub-optimal . Our goal in this section is to make more rigorous the geometrical arguments in Fig 3 . We start with the observation that , for Gaussian distributed neural responses , the 1 standard-deviation probability contours for the responses in the first layer ( magenta ellipses in Fig 3 ) are defined by Δ r · Σ ξ - 1 · Δ r = 1 , ( 65 ) where Δr ≡ f ( s ) − r represents fluctuations around the mean response to stimulus s . In two dimensions , which we’ll focus on here , Eq ( 65 ) becomes Δ r 1 2 σ 1 2 + Δ r 2 2 σ 2 2 = 1 ( 66 ) where σ1 and σ2 are the lengths of the principal axes of the covariance ellipse ( so σ 1 2 and σ 2 2 are the eigenvalues of Σξ ) and Δr1 and Δr2 are distances spanned by the magenta ellipses along those axes . As shown in Fig 3 , the intersection between the magenta ellipse ( the one defined in Eq ( 66 ) ) and the signal curve tells us the uncertainty in the value of the stimulus . To quantify this uncertainty , we simply set Δr to f′ ( s ) Δsx ( the subscript x indicates that this is the uncertainty in the input layer ) , insert that into Eq ( 66 ) , and solve for Δsx . Defining θ to be the angle between f′ ( s ) and the long principal axis ( see Fig 3 , and note that θ = 0 in panel B ) , and letting σ1 correspond to the length of the ellipse’s major axis ( so σ1 > σ2 ) , we have | f ′ ( s ) | 2 cos 2 θ σ 1 2 + sin 2 θ σ 2 2 = 1 Δ s x 2 . ( 67 ) The left hand side is the linear Fisher information in the first layer [10] , a fact that is useful primarily because it validates our ( relatively informal ) derivation . More importantly , we can now see how iid noise affects information . The addition of iid noise simply increases the eigenvalues by σ2 , so the ratio of the information in the output layer to that in the input layer is I y I x = Δ s x 2 Δ s y 2 = cos 2 θ σ 1 2 + σ 2 + sin 2 θ σ 2 2 + σ 2 cos 2 θ σ 1 2 + sin 2 θ σ 2 2 . ( 68 ) We can identify two limits . First , if θ = 0 ( as it is in Fig 3B ) , this ratio reduces to I y I x θ = 0 = σ 1 2 σ 1 2 + σ 2 . ( 69 ) Second , if tan θ ≫ σ2/σ1 ( which essentially means the green line in Fig 3 intersects the covariance ellipse on the side , as in panel A , rather than somewhere near the end , as in panel B ) , the ratio of the informations becomes I y I x tan θ ≫ σ 2 / σ 1 ≈ σ 2 2 σ 2 2 + σ 2 . ( 70 ) Because σ1 > σ2 , the information loss is larger in the second case than in the first . And the longer and skinnier the covariance ellipse , the larger the difference in information loss . Thus , this analysis quantifies the geometrical picture given in Fig 3 , in which there is larger information loss in panel A ( where θ > 0 ) than in panel B ( where θ = 0 ) . Here we ask: what correlational structure minimizes linear Fisher information ? To answer that , we use the multi-dimensional analog of Eq ( 67 ) , I x ( s ) = | f ′ ( s ) | 2 ∑ k cos 2 θ k σ k 2 ( 71 ) where σ k 2 is the kth eigenvalue of the noise covariance matrix and θk is the angle between f′ ( s ) and the kth eigenvector [10] . We would like to minimize Ix ( s ) with respect to the angles , θk , and the eigenvalues , σ k 2 . Without constraints , this problem is trivial: information is minimized by having infinite variances for the neural activities . To make the problem better-formulated , we add a constraint that prevents the optimization procedure from simply identifying that trivial solution . We’ll come to the constraint shortly , but first we’ll minimize information with respect to the angles , θk . That minimum occurs when the eigenvector corresponding to the largest eigenvalue is parallel to f′ ( s ) ; ordering the eigenvalues so that σ 0 2 is the largest eigenvalue , we have cos θ0 = 1 and cos θk > 0 = 0 . Consequently , the information at the minimum is I x ( s ) = | f ′ ( s ) | 2 σ 0 2 . ( 72 ) The next step is to minimize Ix ( s ) with respect to the eigenvalues , subject to a constraint on the covariance matrix . We consider constraints of the form C ( σ 0 2 , σ 1 2 , . . . ) ≤ C 0 ( 73 ) where , to avoid the trivial solution ( of infinite neural variances ) , C is an increasing function of each of it’s arguments: for all k , ∂ C ( σ 0 2 , σ 1 2 , . . . ) ∂ σ k 2 ≥ 0 . ( 74 ) Examples of C ( σ 0 2 , σ 1 2 , . . . ) are the trace of the covariance matrix ( the sum of the eigenvalues ) and the Frobenius norm ( the square root of the sum of the squares of the eigenvalues ) . Because of Eq ( 74 ) , the information , Eq ( 72 ) , is minimized and the constraint , Eq ( 73 ) , is satisfied when all the eigenvalues except σ 0 2 are zero . At this global minimum , the covariance matrix , Σξ , displays purely differential correlations , Σ ξ = σ 0 2 v 0 v 0 ∝ f ′ ( s ) f ′ ( s ) ( 75 ) where v0 is the eigenvector associated with the largest eigenvalue . The last term in this expression follows because the above minimization with respect to the angles forced v0 to be parallel to f′ ( s ) . Thus , for a broad , and reasonable , class of constraints on the covariance matrix , differential correlations minimize information . Throughout most of our analysis we focused on optimality of information transmission . However , also important is how much information is transmitted at the optimum . That’s the subject of this section . For simplicity we consider a linear gain function , which we set , without loss of generality , to the identity . That allows us to use the analysis above , in the section titled “Identifying the family of optimal covariance matrices” , and in particular Eq ( 46 ) , which links the noise in the input and output layers . Our starting point is the derivation of an expression for the ratio of the information in the output layer to that in the input layer . To do that , we dot both sides of Eq ( 37 ) by f′ on the left and right sides and solve for λ; we then do the same , except we dot with f ′ · Σ y - 1 · [ Σ ξ + Σ y ] on the left and its transpose on the right . This yields , after a small amount of algebra , I y I x = I η I η + I x = 1 1 + I x / I η ( 76 ) where Ix , Iy and Iη are given by Eqs ( 3a ) , ( 3b ) and ( 9 ) , respectively . For information to be transmitted efficiently , Ix , the information in the input layer , must be small compared to Iη , the information associated with the added noise in the output layer . Below , we investigate the conditions under which Ix ≪ Iη , and thus when information loss is small . Our strategy is to express Ix/Iη in terms of the single neuron variability , quantified as the average variance—something that has an easy interpretation . We consider two cases: the weights are set to the identity ( W = I ) , and the weights are more realistic ( each neuron in the input layer connects to a large number of neurons in the output layer ) . The first case , identity weights , is not very realistic; we include it because it is much simpler than the second . While the analysis is straightforward , it is somewhat heavy on the algebra , so we summarize the results here . We consider two extremes in the family of optimal covariance structures: the “matched” case ( α = 1 in Eq ( 46 ) , and , for simplicity , Ω = ∞ ) and differential correlations ( α = 0 ) . For matched covariances , near complete information transfer ( Ix ≪ Iη ) requires the effective variance of the noise in the second layer to be small . For identity feedforward weights , the effective variance in the input and output layers is about the same , so information loss is large . However , identity feedforward weights are never observed in the brain; instead , each neuron in the input layer connects to a large number of neurons in the output layer . Using Nx and Ny to denote the number of neurons in the input and output layers , respectively , and K the average number of connections per neuron , the effective noise is reduced by a factor or K N x 2 / N y 2 ( see Eq 95 below ) . Thus , if the number of neurons in the output layer is larger than the number in the input layer by a factor much larger than K1/2 , near complete information transmission is possible . For pure differential correlations , the story is much simpler: so long as the number of neurons in both layers is large , and the added noise doesn’t have a strong component in the f′ ( s ) direction , near complete information transmission always occurs . In the analysis of nonlinear gain functions in the main text ( section titled “Nonlinear gain functions” ) , it was necessary to construct a covariance matrix such that the information in the first layer was independent of ϵu and u . For that we included a prefactor γu in the definition of the covariance matrix , Σξ ( see Eq ( 12 ) ) . Here we determine how γu should depend on ϵu and u . Our starting point is an expression for the inverse of Σξ . As is straightforward to show , via direct substitution , that’s given by Σ ξ - 1 = γ u Σ 0 + ϵ u u u - 1 = 1 γ u Σ 0 - 1 - ϵ u Σ 0 - 1 · u u · Σ 0 1 + ϵ u u · Σ 0 - 1 · u . ( 97 ) Thus , the information in the input layer , f ′ · Σ ξ - 1 · f ′ , is given by f ′ · Σ ξ - 1 · f ′ = 1 γ u f ′ · Σ 0 - 1 · f ′ - ϵ u ( f ′ · Σ 0 - 1 · u ) 2 1 + ϵ u u · Σ 0 - 1 · u . ( 98 ) To ensure that this information is independent of γu , we let γ u = 1 I x f ′ · Σ 0 - 1 · f ′ - ϵ u ( f ′ · Σ 0 - 1 · u ) 2 1 + ϵ u u · Σ 0 - 1 · u . ( 99 ) Note that γu depends on s as well as ϵu and u . In this section we provide details for the numerical simulations for each relevant figure .
|
Information about the outside world , which originates in sensory neurons , propagates through multiple stages of processing before reaching the neural structures that control behavior . While much work in neuroscience has investigated the factors that affect the amount of information contained in peripheral sensory areas , very little work has asked how much of that information makes it through subsequent processing stages . That’s the focus of this paper , and it’s an important issue because information that fails to propagate cannot be used to affect decision-making . We find a tradeoff between information content and information transmission: neural codes which contain a large amount of information can transmit that information poorly to subsequent processing stages . Thus , the problem of robust information propagation—which has largely been overlooked in previous research—may be critical for determining how our sensory organs communicate with our brains . We identify the conditions under which information propagates well—or poorly—through multiple stages of neural processing .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"statistical",
"noise",
"ellipses",
"nervous",
"system",
"random",
"variables",
"geometry",
"neuroscience",
"covariance",
"mathematics",
"statistics",
"(mathematics)",
"algebra",
"computational",
"neuroscience",
"neuronal",
"tuning",
"coding",
"mechanisms",
"gaussian",
"noise",
"animal",
"cells",
"neural",
"pathways",
"probability",
"theory",
"cellular",
"neuroscience",
"neuroanatomy",
"cell",
"biology",
"linear",
"algebra",
"anatomy",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"computational",
"biology",
"eigenvalues"
] |
2017
|
Robust information propagation through noisy neural circuits
|
Congenital heart disease ( CHD ) has a complex genetic etiology , and recent studies suggest that high penetrance de novo mutations may account for only a small fraction of disease . In a multi-institutional cohort surveyed by exome sequencing , combining analysis of 987 individuals ( discovery cohort of 59 affected trios and 59 control trios , and a replication cohort of 100 affected singletons and 533 unaffected singletons ) we observe variation at novel and known loci related to a specific cardiac malformation the atrioventricular septal defect ( AVSD ) . In a primary analysis , by combining developmental coexpression networks with inheritance modeling , we identify a de novo mutation in the DNA binding domain of NR1D2 ( p . R175W ) . We show that p . R175W changes the transcriptional activity of Nr1d2 using an in vitro transactivation model in HUVEC cells . Finally , we demonstrate previously unrecognized cardiovascular malformations in the Nr1d2tm1-Dgen knockout mouse . In secondary analyses we map genetic variation to protein-interaction networks suggesting a role for two collagen genes in AVSD , which we corroborate by burden testing in a second replication cohort of 100 AVSDs and 533 controls ( p = 8 . 37e-08 ) . Finally , we apply a rare-disease inheritance model to identify variation in genes previously associated with CHD ( ZFPM2 , NSD1 , NOTCH1 , VCAN , and MYH6 ) , cardiac malformations in mouse models ( ADAM17 , CHRD , IFT140 , PTPRJ , RYR1 and ATE1 ) , and hypomorphic alleles of genes causing syndromic CHD ( EHMT1 , SRCAP , BBS2 , NOTCH2 , and KMT2D ) in 14 of 59 trios , greatly exceeding variation in control trios without CHD ( p = 9 . 60e-06 ) . In total , 32% of trios carried at least one putatively disease-associated variant across 19 loci , suggesting that inherited and de novo variation across a heterogeneous group of loci may contribute to disease risk .
Congenital heart disease ( CHD ) is the most common congenital malformation and the most common cause of mortality during the first year of life in the United States [1 , 2] . Most cases occur sporadically without a strong family history or identifiable genetic syndrome , and the primary heritable basis of most non-syndromic congenital heart disease has yet to be identified [3 , 4] . Studies of affected kindreds and syndromic disease have revealed high-penetrance mutations at a small number of key loci [5] . Exome sequencing and studies of structural variation of mixed cardiac phenotypes focusing on de novo events have identified novel disease loci in 4–10% of participants [6 , 7] . However the remaining majority of non-syndromic subjects in exome and CNV studies are without an identified genetic cause . Atrioventricular septal defects ( AVSD ) are a rare cardiac malformation associated to date with a handful of canonical genes in cardiac development ( NKX2-5 , GATA4 , GATA6 , CRELD1 ) and may co-occur with certain rare syndromes . A recent study discovered causal variation in the nuclear receptor NR2F2 in 4% of 125 subjects with AVSD , pinpointing a single additional disease-associated gene [7] . However the 4–10% discovery rate in studies of CHD highlight the observation that for any individual gene , highly penetrant de novo coding mutations may only account for a small portion of disease incidence , a phenomenon similar to the sporadic occurrence and complex genetics of neurodevelopmental disorders [8] . Therefore , expanding the scope of analysis in studies of CHD to include both inherited and de novo variation in multiple genes could increase the sensitivity of genetic studies of this heterogeneous group of oligogenic diseases [9–13] . To this end we assembled a multi-institutional cohort combining a discovery cohort of 59 trios with non-syndromic AVSD and a replication cohort of 100 single affected individuals and performed a genetic survey by exome sequencing and array-CGH . In a primary analysis we identified a novel candidate gene for AVSDs using inheritance modeling and prior knowledge of early cardiac gene expression ( Fig 1 ) . In secondary analyses , we searched protein interaction networks to identify the contribution of additional loci to this rare cardiovascular malformation . Finally we explored the contribution of rare inherited variation in genes related to other types of human and mouse cardiac malformations to AVSD .
We determined the sensitivity of our informatics approach using a recently described consensus standard dataset of exonic 24 , 734 variants from NA12878 [14] . Raw exome data from a well-characterized individual was analyzed with BWA/GATK 3 . 2 best practices and the RTG version 3 . 3 . 2 software pipeline ( Real Time Genomics Inc . , Hamilton , New Zealand ) . RTG displayed greater sensitivity detecting 84 . 5% consensus standard variants compared to 79 . 9% for BWA/GATK ( S1 Table ) . Using the RTG pipeline we analyzed exome sequencing data on 159 individuals with AVSD but without a syndrome ( a discovery cohort of 59 trios , and a replication cohort of 100 singletons derived largely from a published study of AVSDs ( S2 Table ) , along with 710 controls without congenital heart disease ( 59 trios , 533 singletons ) . The affected patients were situs solitus with a simple AVSD . Patients with other cardiac defects , heterotaxy , anatomical malformations , or developmental delay were excluded . Across all individuals , called variants displayed a median Ti/Tv ratio of 3 . 10 , and a median of 89 . 6% phased genotypes within trio probands , which suggested an overall highly sensitive and accurate variant call set ( S3 Table ) . All de novo variants and insertions/deletions of interest were confirmed by Sanger sequencing . Within the 59 AVSD trios and 59 control trios we analyzed variants with minor allele frequency of 0 . 03 or less in the 61 , 468 multiethnic individuals in the EXaC consortium [15] . With the remaining protein-altering variants , we applied a rare-disease inheritance model to select only variants displaying classical inheritance patterns associated with sporadic presentation of a rare disease ( de novo , homozygous , compound heterozygous ) ( Fig 1 ) [16]; this filtering process yielded a list of 710 variants in 399 genes in the 59 AVSD trios ( S4 Table ) . To identify novel genes involved in cardiac development and disease among the 399 loci , we reanalyzed 72 digital gene expression datasets derived from 22 tissues during mouse embryonic development ( www . mouseatlas . org ) [17] ( S1 Fig ) . The tissue types included the AV-canal along with 5 other cardiac tissues , and 16 other tissues from other organs or structures . After constructing co-expression modules using unsupervised weighted-gene coexpression network analysis , we observed that one of the co-expression modules expressed in mouse AV-canal tissue included four of six genes known to cause AVSDs ( GATA4 , GATA6 , NKX2-5 , and CRELD1 , p = 6 . 57e-05 , one-tail hypergeometric test ) along with 69 of 756 genes related to other human or mouse cardiac malformations ( S5 Table ) ( p = 7 . 81e-08 , one-tail hypergeometric test ) . These observations suggested the discovery of a co-expression module highly enriched for genes related to heart development and cardiac malformations . Intersecting the genes in this unique coexpression module ( S6 Table ) with the 399 genes identified by the rare-disease inheritance model , two probands from the 59 AVSD trios displayed de novo mutations . One proband had a missense mutation in a non-conserved residue of KCNJ3 and a second proband had a missense variant in NR1D2 . The KCNJ3 variant has been observed in low frequencies in European and African populations in the ExAC database and the available literature on Kcnj3 knockout animals did not suggest occult cardiovascular malformations [18 , 19] . By contrast , the NR1D2 variant causes an arginine to tryptophan mutation at position 175 ( p . R175W ) in a highly conserved DNA binding domain ( Fig 2a and 2b ) . NR1D2 is a transcriptional co-repressor and modulator without a described role in cardiac development [20] . De novo mutations in NR1D2 or any gene in the cardiac malformation module were absent from 59 control trios without congenital heart disease . Among the 61 , 468 putatively healthy individuals cataloged in the ExAC database there was only a single non-synonymous mutation in the surrounding five protein residues surrounding the p . R175W allele . Overall the data were suggestive that this de novo allele might impact the function of NR1D2 . To characterize the transcriptional activity of the p . R175W mutation we developed an in vitro transactivation assay performed in HUVEC cells , which employed a murine Nr1d2 expression vector co-transfected with an Nr1d2 response element ( RE ) consisting of 5 tandem repeats of a conserved binding site upstream of a minimal CMV promoter driving GFP . As NR1D2 is thought to act as a transcriptional co-repressor , a positive change in transcriptional activity of the RE vector may represent a decrease in the transcriptional co-repressor activity of Nr1d2 . In this in vitro assay , the p . R175W mutation displayed an increased transcriptional activity relative to the wild-type allele ( Fig 2c ) , which suggested that the p . R175W mutation in a conserved region of the DNA binding domain might functionally impair the native co-repressor function of NR1D2 . Though reports of a previous characterization of an Nr1d2 knockout allele did not show cardiovascular malformations [20] , a percentage of homozygous knockout animals have been reported to die within hours of birth consistent with the presence of hemodynamically significant cardiac malformations [21] . Two pairs of heterozygous founder animals were bred , yielding 17 pups ( 2 wild type , 7 Nr1d2tm1-Dgen +/- , and 8 Nr1d2tm1-Dgen -/- ) which did not deviate obviously from expected Mendelian allelic ratios . Upon careful histological examination of two spontaneously deceased Nr1d2tm1-Dgen -/- pups at P0 we detected a previously undescribed AVSD phenocopy ( Fig 2d ) . We performed additional matings of +/- and -/- animals , and sacrificed mothers to obtain embryos at e16 . 5 and e17 . 5 . A single -/- animal at e16 . 5 displayed an AVSD and a single -/- animal at e17 . 5 displayed an inlet ventricular septal defect which is closely related to AVSDs ( Fig 2d ) . In total , 4 out of 15 -/- hearts assayed displayed a cardiac defect . Thus by combining inheritance modeling with a gene-coexpression network enriched for genes causing CHD , we identified a variant in NR1D2 which impacts transcriptional activity in vitro , and uncovered previously unrecognized cardiovascular malformations in an Nr1d2 knockout allele . Together these data suggest a new role for the transcriptional repressor NR1D2 in cardiac development and human disease . The relatively low discovery rate in rare-variant association studies of CHD suggests that alternative analytical approaches may be necessary to distinguish the contribution of novel loci to disease risk [6 , 7] . Protein interaction networks have successfully integrated known disease genes to discover the impact of novel loci in neurodevelopmental disorders and cancer , thus in a secondary analysis we employed an algorithm which searches protein-interaction data for over-representation of genetic variation within interacting proteins in the AVSD trios [22] ( Fig 3 ) . Including protein altering single nucleotide mutations and CNVs derived from the discovery cohort of 59 AVSD trios , the algorithm identified 86 enriched subnetworks of interacting proteins containing 2 to 7 genes . By comparison , applying the algorithm to 59 control-trios identified 26 subnetworks of interacting proteins containing only 2 to 4 genes ( Fig 4a ) . Using a procedure where the protein interaction network is randomly permuted , the genes within the AVSD-trio subnetworks were found to be enriched for true protein-protein interactions ( median p = 0 . 01 , network permutation procedure ) , while true protein-protein interactions were not observed within the control trio subnetworks ( median p = 1 . 0 , network permutation procedure ) [23] ( Fig 4a ) . To further characterize the protein interaction networks detected , we compared the subnetworks to gene expression in mouse cardiac development ( S8 Table ) . Genes within the AVSD-trio subnetworks were strongly overrepresented during mouse heart development ( p = 9e-09 , one-tailed hypergeometric test ) while genes within the control-trio subnetworks were not ( p = 0 . 34 , one-tailed hypergeometric test ) ( Fig 4b ) . Thus , mapping genetic variation in the AVSD trios to protein interaction networks identifies 86 enriched subnetworks with deleterious variation in 231 genes that are preferentially expressed during cardiac development , a phenomenon not seen in the control trios without CHD . To validate the genetic associations suggested by the discovered AVSD-trio subnetworks , we assembled a separate replication cohort of 100 singleton individuals ( S2 Table ) and performed burden testing for each of the 86 protein networks ( Fig 3 ) . After Bonferroni correction for multiple hypothesis testing , a single subnetwork from the AVSD trios composed of a pair of interacting collagen genes ( COL2A1 , COL9A1 ) ( Fig 4a and 4c ) displayed an elevated burden of rare coding variation in 100 affected individuals with AVSD compared to 533 controls without congenital heart disease ( p = 8 . 37e-08 , SKAT linear weighted test ) ( Fig 4c ) . Interestingly , the two genes COL2A1 and COL9A1 have evolutionarily conserved roles in cardiac development in both zebrafish and mouse [24–26] . One mouse knockout allele of Col2a1 displays cardiac valve abnormalities [27] , and mutations in both genes are associated with Stickler syndrome where 46% of patients are affected with congenital dysfunction of the mitral valve [28] . Together the functional data on COL2A1 and COL9A1 accompanied by the identification of these genes with two separate methodologies in two separate cohorts of AVSD patients , support a potential association with other congenital structural malformations of cardiac valve tissue such as AVSDs . Outside of novel genes identified by developmental coexpression and protein interaction analyses , we wished to examine the impact of genes known to play a role in cardiovascular development or CHD in our cohort . Interestingly , de novo single nucleotide mutations in genes previously associated with AVSD ( NKX2-5 , GATA4 , GATA6 , EVC , CRELD1 , NR2F2 ) were absent and we did not detect genes with recurrent de novo mutations . In an effort to categorize and catalog variation at known CHD loci within the 710 variants in 399 genes identified by the rare-disease inheritance model in the 59 AVSD-trios ( S4 Table ) , we assembled a predefined group of 756 loci associated with any human or mouse cardiac malformation ( S5 Table ) . Among the genes identified by the rare-disease inheritance model in the 59-AVSD probands ( S4 Table ) , we observed inherited and de novo variation in 16 genes ( Table 1 and S7 Table ) [29] . Four of the affected probands displayed variation in more than one gene . In a set of 59 control trios without congenital heart disease we applied the identical variant calling pipeline and rare-disease inheritance model . Comparing the number of AVSD probands with inherited mutations in the identified 16 genes to unaffected controls , we observed 16 mutations in the AVSD-cases and only a single variant in controls ( p = 9 . 60e-06 , Fisher's exact test ) , and additional simulations confirmed an unusual distribution of mutations in the AVSD-cases compared to controls was unlikely to be a chance occurrence ( p = 1 . 23e-06 , Monte Carlo simulation ) . As an additional negative control we compared mutations in a list of 43 genes associated with congenital ocular malformations between the 59 AVSD cases and 59 controls; there was only a single inherited variant among the AVSD trios and none within the control trios ( p = 1 . 0 , Fisher’s exact test ) . Together these results suggest a preponderance of de novo and inherited variation in genes associated with human or mouse cardiac malformations detected in the AVSD trios which greatly exceeded similar variation in control trios . Within the discovery cohort of 59 trios we also assessed structural variation by array CGH and read-depth analysis from exome studies . One patient was observed to have a 3 . 7 Mb de novo deletion at 8p23 . 1 encompassing 43 genes including GATA4 ( Table 2a ) . An additional paternally inherited duplication at chr22:21 , 989 , 140–23 , 627 , 391 partially overlapping the congenital heart disease-associated 22q11 . 2 duplication syndrome region was also detected [30] . Additional CNVs with a previous association to CHD were identified in the singleton subjects ( Table 2b ) [31] . CNVs in these regions were absent from the trio and singleton controls . Thus , including inherited variants , de novo mutations , and structural variation , rare deleterious variants or CNVs in genes with strong prior evidence for association with congenital heart disease were observed in 14 of 59 or 23% of affected trios .
When combined with inheritance analysis , a gene-coexpression network derived from mouse development allowed us to identify a previously unrecognized role for the transcriptional repressor NR1D2 in cardiac development and human disease . Our experimental studies suggested that the observed p . R175W mutation impacts the transcriptional activity of murine Nr1d2 , and we observed previously unrecognized cardiac malformations in the Nr1d2tm1-Dgen knockout mouse . Yet within a cohort of 159 affected individuals , there was only a single patient with a de novo mutation in NR1D2 , which highlights the underlying genetic heterogeneity of CHD and the utility of applying orthogonal datasets to pinpoint causal variation . The -/- animals for the Nr1d2tm1-Dgen allele display incomplete penetrance; a majority of animals do not display cardiovascular malformations and develop normally to adulthood displaying phenotypes related to circadian rhythm and abnormal lipid metabolism [20] . Nr1d2 may retain multiple roles in modulating transcription but is most clearly described as a transcriptional repressor , therefore a mutation in the DNA binding domain might impact Nr1d2 to binding to target sequences resulting in a “de-repression” transcriptional targets of Nr1d2 . A knockout allele such as Nr1d2tm1-Dgen might similarly “de-repress” targets of Nr1d2 . From the standpoint of transcriptional repression , increased transcription of the p . R175W mutant observed in vitro may represent a decrease in transcriptional repression relative to the wild-type protein , and as such could be entirely consistent with the phenotype of a mouse knockout allele . Interestingly , NR1D2 is a well-characterized component of the molecular clock , and further studies would be necessary to investigate NR1D2 as a link between the molecular clock and timing of cardiac development . Chromatin remodeling factors have recently been implicated as primary and secondary causal factors in CHD [6 , 32] , and both newly discovered factors NR1D2 and NR2F2 may play integrated roles in chromatin remodeling during cardiac development . The key histone deacetylase HDAC1 is directly activated by NR1D2 binding and indirectly activated by NR2F2 via PROX1 [33–35] . Additionally NR1D2 may function upstream of NR2F2 , modulating the auto-regulatory activity of NR2F2 via HDAC1 and the glucocorticoid receptor GR complex [36 , 37] . Further experiments are necessary to delineate the tissue localization , timing , expression , and functional roles of these two transcription factors and their role in chromatin modulation and transcriptional regulation during cardiac development . Within a complex cellular or tissue signaling pathway , capturing the genetic variation in one or more interacting proteins has yielded novel candidate genes in cancer and neurodevelopmental disorders [22 , 38] . Adapting a tool designed to search protein interaction networks in cancer , we identified a small number of variants in genes within the AVSD trios , two of which ( COL2A1 and COL9A1 ) were subsequently validated in burden testing of a separate replication cohort of 100 individuals at a statistically significant threshold . Independent experimental data implicates these genes in the development of the cardiac valve structures , and links these genes to a genetic syndrome that includes abnormalities of the cardiac valves among a host of other phenotypes . Though our observations are not firmly conclusive of a causal role for COL2A1 and COL9A1 in the pathogenesis of AVSDs , they are supportive of such a role , and we believe , consistent with the idea that network-based approaches may be fruitfully applied to gene discovery in CHD phenotypes . Surprisingly , with the exception of a deletion encompassing GATA4 seen in one trio subject and one singleton subject ( Table 2 ) , we did not discover de novo coding mutations or gene dosage alterations within the 59 trios in canonical AVSD genes ( NKX2-5 , EVC , CRELD1 , GATA6 ) or at the newly discovered CHD risk locus NR2F2 . This finding is consistent with studies of CHD examining candidate genes [39] and exome sequencing where protein-altering variants in any single gene are reported in no more than 1–4% of patients [7] . The absence of recurrent de novo variants in a cohort of 59 patients with AVSD is in striking contrast with other cardiac conditions such as long QT syndrome where pathogenic coding variation in only 5 genes accounts for disease in 70% patients [40] . We hypothesize that the absence of de novo variation observed at canonical loci in a cohort of this size reflects the complexity of CHD genetics and highlights the utility of considering alternative inheritance patterns to detect disease-associated variation . Among a list of 756 genes with either a clinical or experimental association to cardiac malformations ( S5 Table ) we observed rare inherited variants in the AVSD trios that were not seen in control trios [29] . In genes clinically associated with CHD we observed compound heterozygous variants inherited in trans in ZFPM2 , NSD1 , NOTCH1 , VCAN , and MYH6 and rare homozygous variants in MYH6 . Variation was observed in 9 additional genes including compound heterozygous variants in ADAM17 , CHRD , IFT140 , PTPRJ , and RYR1 and rare homozygous variants in ATE1 , and the presence of heart defects in mouse knockout models for these loci supports their association with human cardiac malformations . In an independent forward mutagenesis screen Ift140 causes AVSD among a variety of congenital malformations [41] . The calcium channel RYR1 is associated with skeletal myopathies and malignant hyperthermia , but primum atrial septal defects in one mouse allele suggest a role in early cardiac development [42] . The metalloproteinase ADAM17 may link NOTCH1 signaling in cardiac valve development to the left-right patterning of the heart [43 , 44] . Individual knockout alleles of CHRD , PTPRJ , and ATE1 each show defects in heart development recapitulating different human malformations [45–47] . Importantly , we observed rare inherited variation in genes with experimental or clinical evidence for a role in cardiac development and CHD within the AVSD trios , but rare inherited variation in these same genes was largely absent from the control-trios . Despite excluding syndromic features and developmental delay from our patients at the time of recruitment , we observed inherited and de novo variation in genes causing syndromic disease that include heart malformations . A de novo mutation was detected in an unknown protein domain of EHMT1 the causal gene in Kleefstra syndrome , compound heterozygous variants inherited in trans were observed in SRCAP which was recently associated with Floating-Harbor syndrome , two individuals showed compound heterozygous mutations in BBS2 which causes Bardet-Biedel syndrome , a fifth individual displayed compound heterozygous mutations in NOTCH2 which causes Alagille syndrome , and a sixth individual displayed a compound heterozygous mutation in KMT2D the locus implicated in Kabuki syndrome . Each of these multi-organ syndromes is frequently accompanied by AVSD or another related form of congenital heart disease [48–52] . Within a single locus associated with a genetic syndrome , different alleles may vary in their expressivity . We hypothesize that these variants represent hypomorphic alleles of syndromic genes , where the patients affected present only one aspect of the phenotype associated with the syndrome , in this case a phenocopy of a syndromic associated cardiovascular malformation [53] . Indeed on secondary followup , none of the included probands with EHMT1 , KMT2D , SRCAP , or NOTCH2 variants displayed other characteristics of their associated syndromes , while the patients with BBS2 mutations were not available for review ( additional phenotypic information on patients carrying syndromic alleles is detailed in the S1 Text ) . Supporting the possibility of hypomorphic alleles , there was a striking absence of de novo or inherited variants in these syndromic genes within the control trios suggestive that the discovered variants may confer risk for AVSD . These findings have limitations . Although we excluded patients with a family history of cardiac malformations , in an earlier era of surgical care the parents of the study participants would have been less likely to survive to reproductive age [54] . In our study the parents received only a questionnaire and did not receive screening echocardiogram , thus we cannot rule out that a parent in an included trio may have a forme fruste of an AVSD-related malformation such as a cleft mitral valve or ostium primum ASD . Additionally , there is emerging evidence that maternal risk factors ( both genetic and environmental ) which confer risk for CHD that were not considered in our study design [55–57] . Genetic studies of CHD are challenged by the fact that specific individual malformations are quite rare ( AVSD is approximately 0 . 3 per 10 , 000 live births ) , and that any substantial group of patients with a single malformation will contain some population stratification . The unexplored role of non-coding gene regulatory variation in congenital heart disease is not surveyed by our exome-sequencing approach [58 , 59] . The power of SKAT tests are likely limited by a small cohort size , the heterogeneous genetic backgrounds of the case and control populations , the differences in exome sequence capture and sequencing chemistries employed , the absence of an inheritance model , and most importantly the underlying complex oligogenic architecture of cardiac malformations [12 , 60 , 61] . Finally there are no statistical models that account for ethnicity in models of rare-variant transmission , therefore the influence of population stratification or ethnicity upon our rare-inheritance model of disease is not known . Overall our analysis suggests locus heterogeneity in the pathophysiology of a single cardiac developmental malformation . We observed recurrent variation within three genes ( GATA4 , MYH6 , and BBS2 ) in only 6 of 59 trios . Including inherited , de novo , and discovered loci , 32% of trios displayed one or more putatively contributory mutations in the 19 genes identified . Supported by both experimental and clinical evidence , we suggest that inherited rare variants with a moderate effect size across multiple loci may impact the risk of congenital heart disease in addition to de novo variation . Taken together our catalog of 19 loci with experimental evidence for disease among 159 patients is consistent with the long-hypothesized oligogenic inheritance of congenital heart disease [12] .
The guidelines of the Declaration of Helsinki were followed and the study was approved by the institutional review board of Stanford University ( IRB-23637 , IRB-23572 ) along with each institution from which participants were recruited . Written informed consent was obtained from each participant . Trios or single patients with an AVSD and situs solitus were recruited . We excluded patients with other major congenital malformations , developmental delay , or other types of CHD ( excluding patent ductus arteriosus or secundum atrial septal defect ) . All participants were assessed clinically by pediatric cardiologists or clinical geneticists to exclude other syndromes associated with AVSD or CHD . Participants were obtained from Seattle Children’s Hospital [62] , the Pediatric Cardiac Genomics Consortium , the University of Iowa , and reanalyzed from a published study of AVSDs drawn from patients at the University of Toronto [7] ( S2 Table ) . Exome sequences for 531 control subjects of Caucasian and African American descent without CHD were derived from the atherosclerosis risk in communities ( ARIC ) consortium warehoused at dbGAP , and from local recruitment efforts . An additional 59 control trios ( healthy parents with a healthy child ) were obtained from the Simons Foundation . Raw sequence data in the form of bam or fastq files was re-aligned and re-analyzed with the below-described bioinformatic pipelines . DNA was isolated by standard techniques from either whole blood , saliva samples , or immortalized lymphoblasts . Exome sequencing was performed for all complete trios and single affected individuals at two academic centers ( University of Washington or Yale University ) and two commercial sequencing providers with the SeqCap EZ Human Exome Library v2 . 0 ( Roche NimbleGen , Madison , Wisconsin , USA ) , SureSelectXT Human All Exon V4 ( Agilent Technologies Inc . , Santa Clara , California , USA ) , or a proprietary capture library based on the Agilent SureSelectXT Human All Exon V5 platform ( Personalis , Corp , Menlo Park , California , USA ) . Paired-end sequencing was performed on Illumina HiSeq 2500 machines with 75- , 100- , or 150-bp read lengths in all but two trios , which were sequenced with 33 bp paired end reads . For all included samples Median Ts/Tv was 3 . 10 while coverage depth was 43 . 8x ( S3 Table ) . Exome sequencing on single individuals from the Toronto cohort was performed as described [7] , and for subset of unrelated individuals raw fastq files were obtained and reanalyzed via the below described bioinformatics platform for the purposes of reanalysis . AVSD trios and the singleton/replication patients not originating from Toronto were assayed for CNVs by array comparative genomic hybridization ( CGH ) by using a custom chip described previously or the SurePrint G3 Human CGH Microarray Kit , 8x60K ( Agilent Technologies Inc . , Santa Clara , California , USA ) with described protocols [31 , 62] . When a CNV was detected in an affected participant , a dye-swap experiment was repeated to ensure reproducibility and the parents were then assayed if available to determine inheritance status . For detection of smaller variation , two exome CNV detection assays using read depth data were also employed for all participants [63 , 64] . When there was agreement in a structural variant call between at least two calling methods for any variant of interest , we employed an orthogonal CNV genotyping assay when DNA was available ( for all of the AVSD trios and the singleton/replication patients not originating from Toronto ) . The CNV genotyping assay was carried out with 10 ng of DNA per manufacturer instructions against the RNaseP CNV reference assay ( Life Technologies , Carlsbad , California , USA ) . The assay was run on a ViiA 7 Real Time PCR System ( Life Technologies , Carlsbad , California , USA ) for 40 cycles under standard reaction conditions , and CNV genotypes were called with the copycaller software [65] . Two genotyping pipelines were tested . A rapid and sensitive commercial software package , rtg-core version 3 . 3 . 2 was applied to the raw exome sequence data for mapping , pedigree-aware variant calling , and genotype filtration ( Real Time Genomics Inc . , Hamilton , New Zealand ) [66] to the UC Santa Cruz human genome reference sequence ( hg19 ) ( S1 Script ) . A second pipeline based on the HugeSeq BWA/GATK HaplotypeCaller pipeline was also employed for purposes of comparison . To determine the accuracy of our genotyping pipeline we re-genotyped available exome data from NA12878 for comparison to a recently described gold standard dataset of 24 , 734 variants from this individual [14] . This set of variants from the consensus standard dataset was limited to the exome capture region of the nextera kit all human exome v2 ( Illumina Corp , San Diego , California , USA ) and regions of suspicious variant quality [67] were excluded to yield a true positive dataset of 24 , 734 true positive variants from NA12878 . The vcfeval tool from RTG was used for all comparisons of vcf files as it robustly handles the different possible textual representations of insertions , deletions , and substitutions that may be produced by different genotyping algorithms . Using raw fastq files generated by the Garvan Institute , comparing the unfiltered output of both pipelines the two algorithms both called 19 , 566 variants ( 79 . 9% ) of the NA12878 true positive dataset ( TP ) in common . RTG called an additional 1 , 469 TP , compared to BWA/GATK which only called an additional 198 unique TP variants . Overall RTG displayed a greater unfiltered sensitivity at 84 . 5% compared to 79 . 9% for BWA/GATK; therefore we selected the RTG pipeline for further analysis of the cohort ( S1 Table ) . To classify variants we selected an AVR score of 0 . 5 to balance a sensitivity of 99 . 4% and positive predictive value of 90 . 3% for the purposes of variant discovery . Alignments for all variants of interest were manually inspected with the IGV software , and all de novo coding variants and any insertion/deletion of interest was confirmed by direct Sanger sequencing of PCR amplicons , or alternately clonal Sanger sequencing of 12 colonies from cloned PCR amplicons . For burden testing , pooled simultaneous variant calling for all included cases and controls was performed with the RTG population caller and variants filtered for a read depth of 8 and AVR score of 0 . 5 . Analyses were limited to the regions of intersection of the bed files of the exome capture kits obtained from the respective manufacturers . Known artifactual variants arising from exome sequencing were removed at the time of variant filtration [67 , 68] . All statistical analyses employed the R language for statistical computing version 3 . 1 unless otherwise specified . For population analyses we selected 8 , 940 snps from the Affymetrix Genome-Wide Human SNP Array 6 . 0 ( Affymetrix , Santa Clara , California , USA ) common to the five exome capture protocols employed in the study . All included probands were re-genotyped for this limited set of variants , and combined with individual level data from 1032 diverse samples of known ethnicity from the 1000 genomes project . The MDS and kinship modules from the KING software were used to estimate ethnic background and five individuals displaying cryptic familial relationships from the Toronto cohort among the singletons were identified and excluded from further analysis [69] . A multinomial linear model was built for each population and the presence of admixture from the populations of the 1032 known samples , and used to infer ethnic background and the presence of admixture in the included probands . Self-reported ethnicity was available for 547 individuals , which was 97 . 6% concordant with predicted ethnicity . Protein altering variants were sorted for minor allele frequency less than 0 . 03 and prioritized by inheritance patterns consistent with rare disease ( de novo , rare homozygous , and compound heterozygous ) using the trioTools module from the STMP package [16] . Protein-altering variants from the 59 trios were selected , haplotypes constructed , and variants phased with the PLINKseq software package [70] . Imputation was not performed thus all variants analyzed originated from the primary genotyping pipeline . After sorting for inheritance consistent with rare disease ( de novo , rare homozygous , and compound heterozygous ) , variants were collated and assembled into lists . A blacklist of genes with low prior likelihood of causality was excluded from further analysis , which included genes with a residual variation intolerance score greater than 90 and genes with copy number polymorphisms [68 , 71 , 72] . For statistical comparisons between the groups of trios , the number of individuals with rare inheritance ( de novo , rare homozygous , or compound heterozygous ) in these genes in the 59 trios and 59 controls was counted and compared with a Fisher’s Exact test . As the underlying distribution combining de novo , rare homozygous , and compound heterozygous inheritance into a unified “rare inheritance” model is not readily estimated from available genotyping data , and because variant detection ( particularly de novo variation ) may vary systematically with the technical aspects of sequencing ( variant calling algorithm , read depth , exome capture platform , and sequencing chemistry ) we estimated empiric p-values by Monte Carlo simulation . To simulate an underlying distribution we performed permutations drawing 16 genes randomly from a list of 18 , 495 protein-coding genes , the rare inheritance events for each random list counted within 59 cases and 59 controls by individual , and a Fisher’s exact test applied . The p-value was estimated with the formula ( r+1 ) / ( n+1 ) where n is the number of simulated replicate samples and r is the number of test statistics exceeding the calculated test statistic from the observed data ( p = 9 . 60e-06 ) . This simulation procedure suggested an empiric p-value of 1 . 23e-06 , a similar order of magnitude as the p-value calculated from the observed data . Gene modules or subnetworks identified by protein interaction networks in the HotNet2 algorithm ( see below ) , were subjected to burden testing with the SKAT linear weight test and a Bonferroni correction for multiple hypothesis testing employed [73 , 74] . The first four principal components from a principal components analysis of all variants were used as covariates for burden testing . Within the SKAT algorithm variant weighting was derived from a beta density function ( pi , 1 , 25 ) where pi is the minor allele frequency . For variant weighting , minor allele frequencies were derived from the 61 , 428 individuals in the multiethnic EXaC dataset ( version 0 . 2 http://exac . broadinstitute . org/ ) , while for variants not observed or reported in the EXaC dataset minor allele frequencies were calculated from frequency observed among the genotyped individuals . The quantile-quantile plots for the SKAT linear weight test of 86 subnetworks ( inclusive of combinations of 231 single genes—see below ) suggest that the test-statistics derived are controlled for ethnicity or other systematic differences between cases and controls such as coverage ( S2 Fig ) . Unique tag-count data from 74 SAGE libraries representing 22 tissues constructed as a part of the mouse atlas of gene expression project were downloaded from www . mouseatlas . org [17] . SAGE tags were mapped with the Burrows-Wheeler aligner to the 115 , 746 unique mouse RefSeq transcripts downloaded from UC Santa Cruz website ( http://genome . ucsc . edu ) and tag data converted to a digital gene expression format constituting a tag counts per transcript using custom Perl scripts . Tags matching pseudogenes were removed . Using the R environment for statistical computing , tag counts by transcript were normalized by library size with the EdgeR package and collapsed to gene by connectivity with the WGCNA package [75 , 76] . A standard WGCNA workflow for digital gene expression was applied to the normalized and collapsed data for coexpression module construction followed by correlation to the tissue of origin and annotation with the human orthologous gene name when available ( S2 Script ) . To independently assess the predictive capacities of our unsupervised network building procedure , we observed that the developmental expression of 13 well characterized congenital heart disease genes are accurately localized by the network model to their appropriate cardiac tissue , suggesting excellent specificity for detecting developmental cardiac related gene expression ( S9 Table ) . Comparison of variant and gene lists for under- and over-representation were performed with a one-tailed hypergeometric test . Gene expression networks were visualized with the Gephi software package [77] . Because the DMP is a key structure in development of the atrioventricular septum [78] and was not explicitly included in the original set of micro-dissected tissues from the mouse atlas of organ development ( www . mouseatlas . org ) , we added the DMP gene expression data to the cardiac development gene expression set . Among genes expressed in the DMP we selected the most highly expressed 1000 genes across 6 datasets generated from the posterior second heart field including the dorsal mesenchymal protrusion . The microdissection of the posterior heart field was performed in the laboratory of Dr . Moskowitz at embryonic mouse tissue at E9 . 5 , subject to reverse-transcription , amplification , and sequencing by The University of Chicago Genomics Core ( GSE75077 ) . Variant data annotated with the STMP package [16] including SNVs and CNVs from the 59 AVSD trio probands and 59 control trio probands was converted and formatted with custom python scripts , and included all protein altering variants with a minor allele frequency in EXaC of 0 . 02 or less were included in the analysis using four included protein interaction networks . For each of three protein interaction networks ( Multinet , IrefIndex9 , and HINT [22] ) four delta values were derived with the network permutation procedure and applied to identify subnetworks . The resulting subnetworks identified by the HotNet2 algorithm were manually inspected for validity , and the subnetworks from the MultiNet protein interaction networks were chosen for further analysis . For each set of subnetwork sizes ranging from 2 to 10 , the HotNet2 algorithm derives a p-value from the hypergeometric distribution comparing the observed number of subnetworks of size n within in a dataset compared to an expected number of subnetworks . The expected number of subnetworks is derived from a computationally intensive network permutation procedure; in this case the MultiNet protein interaction network was subject to 100 permutations limiting the range of p-values to a minimum p-value to 0 . 01 and maximum p-value to 1 . For the AVSD-trio subnetworks the median p-value was 0 . 01 across the nine subnetwork sizes and four derived delta values , in comparison to the control-trio subnetworks where the median p-value was 1 . 0; this suggested an enrichment in variants occurring in genes with true protein-protein interactions within the AVSD-trios ( rather than randomly occurring simulated protein-protein interactions in the permuted networks ) . A single output run utilizing the MultiNet protein interaction network applying a delta value of 0 . 00126036397514 yielded 86 subnetworks containing 231 genes in the AVSD trios , and the 86 subnetworks were subject to burden testing in the singleton cohort ( see above ) . Protein interaction data were processed with custom python and shell scripts for visualization using the Gephi software tool [77] . Live breeding pairs of the B6;129P2-Nr1d2tm1Dgen/H mouse line ( hereafter referred to as Nr1d2tm1-Dgen ) were obtained from the European Mouse Mutant Archive ( Munich , Germany ) . Animals were in housed and cared for in AAALAC accredited facilities under standard conditions with oversight and approval from the Stanford University APLAC committee ( protocol APLAC-11334 ) . Euthanasia was carried out under anesthesia with isofluorane following APLAC and AAALAC guidelines using carbon dioxide followed by cervical dislocation . Genotyping was performed by PCR of toe or tail clippings with gel electrophoresis using standard methodology with two primer pairs ( CAAGTAACAAGCCTGGGACATAAAG and CTTCGTAGAGGGAGTAATATGACAC yield a 517 bp PCR product from the WT allele; CAAGTAACAAGCCTGGGACATAAAG and GACGAGTTCTTCTGAGGGGATCGATC yield a 757 bp product from the knockout allele ) . Two pairs of heterozygous founder animals were bred , yielding 17 pups ( 2 wild type , 7 Nr1d2tm1-Dgen +/- , and 8 Nr1d2tm1-Dgen -/- ) which did not deviate obviously from expected Mendelian allelic ratios . Spontaneous death within hours after birth occurred in a single +/- and two -/- animals . The thoracic and abdominal cavities of spontaneously deceased homozygous knockout animals were visually inspected to examine the great vessels and visceral situs which were normal , and hearts dissected out and subject to sectioning and H & E staining by standard techniques . Sectioning of tissue samples was performed following either embedding of frozen sections followed by dehydration and fixation in 10% neutral buffered formalin or alternately fixation in 3% paraformaldehyde followed by alcohol dehydration and paraffin embedding . We performed additional matings of +/- and -/- animals , and sacrificed mothers to obtain embryos at e16 . 5 and e17 . 5 . Visualization was performed by brightfield microscopy on a Nikon 90i Eclipse upright with a DS Fi1 camera with a 20x objective . Both P0 Nr1d2tm1-Dgen -/- animals displayed atrioventricular septal defects , a single -/- animal at e17 . 5 displayed an inlet ventricular septal defect , and single -/- animal at e16 . 5 displayed an AVSD . In total , 4 out of 15 -/- hearts assayed displayed a cardiac defect . Spontaneously deceased animals were analyzed , and additionally euthanasia of pregnant female mice was performed to obtain embryonic animals . The crystal structure of the DNA binding domain of NR1D1 ( RCSB 1A6Y ) was visualized with the PyMOL Molecular Graphics System , Version 1 . 7 . 4 Schrödinger , LLC ( New York , New York , USA ) . A construct containing wild-type murine Nr1d2 cDNA construct under control of a CMV promoter in the pCS6 expression vector was obtained from Transomic Technologies Inc . ( Huntsville , Alabama , USA ) , and subject to a site-directed mutagenesis yielding a codon switch at p . R175W which corresponds to the de novo mutation observed in the subject with AVSD ( S4a Fig ) . An Nr1d2 response element vector was constructed cloning 5 tandem repeats of an evolutionarily conserved NR1D2 binding site REV-DR2 [80 , 81] upstream of a minimal CMV promoter driving GFP expression using the pSF-MinCMV-daGFP vector ( Sigma-Aldrich Inc , St . Louis , Missouri , USA ) ( S4b Fig ) . Site-directed mutagenesis and cloning were performed by a commercial provider GENEWIZ Inc . ( South Plainfield , New Jersey , USA ) , and sequence verified in our own laboratory . The three vectors were subject to routine endotoxin free preparation . Commercially available primary HUVEC cells obtained from Cell Applications ( San Diego , California , USA ) were seeded at 80% confluency in black transparent-flat bottom 96 well plates ( Greiner , North Carolina , USA ) and transfected the following day with 100ng of each vector ( response-element , wild-type or p . R175W ) using Lipofectamine 3000 ( Life Technologies , Grand Island , New York , USA ) . Twenty-four hours after transfection , GFP fluorescence was measured on a Tecan Infinite M1000-multimode plate reader ( Tecan Group Ltd , Mannendorf , Switzerland ) . We performed three transfection conditions including response-element+Nr1d2-WT , response-element+Nr1d2-P403W and response-element alone in 24 technical replicates . Autofluoresence of untransfected wells were averaged , and subtracted from the response-element alone and the two experimental conditions . The highest and lowest value from each condition were excluded from analysis yielding 22 technical replicates per experimental condition , and 4 technical replicates for the response element alone . Statistical analysis and graphing of the transfection experiments was performed in Prism 6 Graphpad Software ( La Jolla , California , USA ) .
|
Congenital heart disease ( CHD ) is a leading cause of childhood morbidity in the developed world . There are few prevalent clinical risk factors and though it is possible that up to 90% of risk for CHD may be genetic , the number of genes clinically associated with disease is small . Rather than grouping disparate CHD phenotypes as other studies have done , we studied a single specific malformation- the atrioventricular septal defect ( AVSD ) . Instead of recurrent variation in a handful of genes , we observed de novo and inherited variation in 19 genes associated with human disease , syndromic loci , and genes implicated in cardiac development by mouse knockout . The number of loci identified support the longstanding hypothesis of a complex oligogenic inheritance for a single malformation and suggest that analyses of CHD data to include inherited variation may uncover additional loci contributing risk for cardiac malformations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"genetic",
"networks",
"cardiovascular",
"anatomy",
"protein",
"interaction",
"networks",
"alleles",
"developmental",
"biology",
"network",
"analysis",
"genome",
"analysis",
"morphogenesis",
"cardiology",
"computer",
"and",
"information",
"sciences",
"genomics",
"birth",
"defects",
"gene",
"expression",
"proteomics",
"congenital",
"disorders",
"genetic",
"loci",
"biochemistry",
"anatomy",
"congenital",
"heart",
"defects",
"genetics",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"genetics",
"of",
"disease",
"heart"
] |
2016
|
De Novo and Rare Variants at Multiple Loci Support the Oligogenic Origins of Atrioventricular Septal Heart Defects
|
Protein phosphorylation is a key mechanism to regulate protein functions . However , the contribution of this protein modification to species divergence is still largely unknown . Here , we studied the evolution of mammalian phosphoregulation by comparing the human and mouse phosphoproteomes . We found that 84% of the positions that are phosphorylated in one species or the other are conserved at the residue level . Twenty percent of these conserved sites are phosphorylated in both species . This proportion is 2 . 5 times more than expected by chance alone , suggesting that purifying selection is preserving phosphoregulation . However , we show that the majority of the sites that are conserved at the residue level are differentially phosphorylated between species . These sites likely result from false-negative identifications due to incomplete experimental coverage , false-positive identifications and non-functional sites . In addition , our results suggest that at least 5% of them are likely to be true differentially phosphorylated sites and may thus contribute to the divergence in phosphorylation networks between mouse and humans and this , despite residue conservation between orthologous proteins . We also showed that evolutionary turnover of phosphosites at adjacent positions ( in a distance range of up to 40 amino acids ) in human or mouse leads to an over estimation of the divergence in phosphoregulation between these two species . These sites tend to be phosphorylated by the same kinases , supporting the hypothesis that they are functionally redundant . Our results support the hypothesis that the evolutionary turnover of phosphorylation sites contributes to the divergence in phosphorylation profiles while preserving phosphoregulation . Overall , our study provides advanced analyses of mammalian phosphoproteomes and a framework for the study of their contribution to phenotypic evolution .
Most proteins undergo chemical modifications after their synthesis ( post-translational modifications , PTMs ) . These modifications allow a fine-tuning of protein functions and represent a mechanism to expand the coding capacity of genes [1] . Over the past decade , methods based on mass spectrometry have accelerated the discovery of PTMs [2]–[7] . Each experiment can now detect thousands of modified residues , allowing to probe the functional state of entire proteomes . The PTM that has been studied the most is protein phosphorylation: the addition of a phosphate group to specific amino acids ( serine ( S ) , threonine ( T ) and tyrosine ( Y ) in eukaryotes ) . Phosphorylation has been shown to affect protein functions , interactions , stability and localization [8]–[11] . It is thus of fundamental importance to understand how protein phosphorylation evolves within and between species because changes in phosphorylation profiles may cause changes in protein function and regulation and in organismal phenotypes , including disease development ( e . g . [12] ) . There have been several reports recently on the evolution of phosphoproteomes . For instance , Kim and Hahn [13] identified phosphorylation sites that emerged after the split between humans and chimpanzees and found that these sites are located in proteins involved in crucial biological processes such as cell division and chromatin remodelling . Other studies have looked at the evolution of a subset of phosphoproteomes on a broader evolutionary scale [14] , [15] . For example , Boulais and collaborators [14] performed a phosphoproteomics analysis of mouse phagosomal proteins and then compared these proteins to their orthologs from 10 model organisms , from Drosophila to mouse [14] . They observed that the phagosomal phosphoproteome was extensively rewired during evolution , but that some phosphorylation sites have been maintained for more than a billion years , suggesting their importance for phagosomal functions . Finally , other studies looked at the conservation and divergence of entire phosphoproteomes over a broad evolutionary scale [16]–[19] ( and reviewed in [20] ) in order to understand the evolutionary mechanisms and the constraints acting on phosphorylation sites . These studies found that phosphorylated residues tend to be on average more conserved than their non-phosphorylated counterparts [16] , [17] and that this is particularly true for those that were experimentally shown to play functional roles [17] . Most studies that aimed at studying the evolution of phosphoproteomes so far have looked at the evolutionary conservation of phosphorylation sites in several species without knowing if these sites are actually phosphorylated in species other than the reference . In other words , if a phosphorylation site in one species corresponds to a phosphorylatable amino acid in another species , both residues were considered as conserved phosphorylation events . This assumption was necessary because of the lack of phosphorylation data available for more than one species . However , we can hypothesize that residue conservation does not always imply phosphoregulatory conservation . Indeed , sites could be conserved at the residue level but differ in their phosphoregulation due to changes elsewhere in the protein , for instance , the recognition motifs of the protein by kinases and phosphatases [21] or upstream ( in trans ) in the signalling cascade . This aspect has not been addressed by previous studies , except in a few cases [14] , [22] . However , identifying such sites is of great interest since sites that differ in their phosphoregulation despite being conserved at the residue level could lead to changes in the architecture of phosphorylation networks and , ultimately , contribute to phenotypic evolution . We examine this issue here . Another aspect of phosphoproteomes that can be studied using evolutionary analysis is how phosphorylation sites alone or in combination may affect the function of a protein [1] . Many models of phosphorylation site function stress the importance of conformational changes by protein phosphorylation [1] , [23] , [24] . In other models , phosphorylation sites regulate protein functions without the need for conformational changes but rather through changes in the local charge of the protein [25] , i . e . simply through bulk electrostatics . A corollary of this last model is that the protein phosphorylation code is redundant , i . e . that phosphorylation sites can change their position over time and still maintain their biological function as long as the number of sites in a given protein region is preserved , without affecting organismal phenotypes . By looking at the patterns of evolution of phosphorylation sites , one could find traces of this redundancy by studying rapid phosphorylation site evolutionary turnover ( phosphorylation site gains and losses ) . This evolutionary turnover has been invoked for interpreting the global rapid pattern of evolution in different species [15]–[17] , [26]–[28] . However , evidence for positional redundancy of phosphorylation sites is relatively limited . Two independent pieces of evidence come from the cell cycle phosphorylation networks . Moses and collaborators [29] studied the evolution of cyclin-dependent kinase ( CDK ) consensus phosphorylation sites of the yeast pre-replicative complex [30] . They found that although orthologous proteins contained clusters of CDK consensus sites , the position and the number of phosphorylatable sites were not conserved , suggesting that phosphorylation sites tend to shift their positions during evolution . In a more recent investigation , Holt and collaborators [31] compared the positions of 547 phosphorylation sites on 308 Cdk1 substrates in vivo in the budding yeast and their orthologous sites in other fungi . They found that the precise positioning is conserved only in the very closely related species . However , in both cases the phosphorylation status of the sites in other species was not investigated so it is not clear whether the phosphorylation sites were absent from the orthologous proteins or if they actually shifted during evolution through gains or losses to another position . The extent to which phosphorylation site positional redundancy plays a role in overall phosphoproteome turnover therefore awaits comprehensive phosphorylation data from closely related species , which we have assembled here . We performed an integrated analysis of phosphorylation site evolution between the human and mouse proteomes using a large dataset of phosphorylation sites [6] , [32]–[37] . These two phosphoproteomes are the ones for which we have the greatest amount of phosphoproteomics data between closely related species . We estimated the extent of divergence and conservation between the two phosphoproteomes and we investigated whether phosphorylation site evolutionary turnover could contribute to this divergence .
We assembled a dataset of human ( n = 106 , 877 ) and mouse ( n = 54 , 400 ) phosphorylation sites by collecting data from 7 different databases and experimental studies [6] , [32]–[37] ( Table S1 ) . We successfully mapped 128 , 705 sites onto 11 , 150 human and mouse orthologous proteins: 86 , 065 in humans and 42 , 640 in mouse ( Figure S1 ) . As previously observed [17] , [38] , phosphorylation sites are preferentially located in disordered regions of proteins ( observed vs . expected proportions: 0 . 69 vs . 0 . 62 , p-value = 2 . 2×10−16 ) . Given this asymmetry in the localization of phosphorylation sites , we generated all the null models of our analyses by respecting the proportion of sites in these two structural categories . Our dataset allows comparing the human and mouse phosphoproteomes using both sequence information and the phosphorylation status of each site . Accordingly , we classified orthologous sites into three classes following Freschi et al . [28] ( Figure 1A ) : i ) Site-diverged ( SiD ) : sites phosphorylated in one species and non-phosphorylatable in the other; ii ) State-conserved ( StC ) : sites phosphorylated in both species; iii ) State-diverged ( StD ) : sites that are conserved at the residue level but that have been reported to be phosphorylated in only one of the two species . In order to examine the extent of conservation of phosphorylation between human and mouse , we estimated the fraction of sites belonging to each of these three categories compared to the total number of sites that are phosphorylated in human , mouse or in both species . We first looked at phosphorylation site divergence . We found that 16 , 863 sites ( 16% of the sites that are phosphorylated in human or mouse or both species ) are SiD ( Figure 1B ) . These sites are about 1% less abundant than random expectations obtained by shuffling the phosphorylation statuses of S/T/Y residues ( Figure 1B ) , suggesting that purifying selection is acting on phosphorylation sites to maintain their function but to a limited extent , as previously observed with different approaches ( e . g . [17] ) . These sites , if functional , are expected to reflect differences in phosphoregulation between human and mouse . However , a fraction of these SiD sites might be positionally redundant site pairs such that the functional divergence may be overestimated ( see below ) . We examined other types of conservation and divergence . We first found that 20 , 146 phosphorylation sites ( 18% of the sites that are phosphorylated in human or mouse or both species , Figure 1B ) are StC . This proportion is 2 . 5 times greater than what is expected by chance alone ( Figure 1B ) . We observed this strong signal for conservation in both disordered and ordered regions ( Figure S2 ) . These results suggest an overall conservation of the phosphorylation profiles between the two species , most likely as a result of purifying selection acting to maintain the phosphoregulation of these sites . We performed a similar analysis on clusters of poly-S/T/Y ( stretches of two or more consecutive S/T/Y residues ) rather than single residues and found the same patterns of conservation and divergence ( Figure S3 ) . Despite an overall signal of conservation on the phosphorylation status of proteins , the most represented category of sites in our dataset is StD sites ( 71 , 550 sites or 66% of the sites that are phosphorylated in human , mouse or both species ) . Three different non-exclusive scenarios could explain this large number of StD sites ( Figure 1C ) . The first one implies that state divergence results from an incomplete coverage of phosphoproteomic data , which means that the phosphoproteomes of the two species might have been undersampled , for instance sampled at different depths or in different conditions or tissues ( e . g . [6] ) . The second scenario is that a large fraction of the StD sites identified might result from non-functional phosphorylation sites . Non-functional phosphorylation sites evolve rapidly [17] and could therefore lead to the poor conservation on the phosphorylation status we observed . The third scenario is that a fraction of StD phosphorylation sites is actually diverging in its regulation . Finally , state-divergence could also be inflated by false positive identifications in one species or the other . We examined which scenario or scenarios were compatible with our data . According to the first scenario , StD may mostly result from false-negative phosphorylation sites in the data . This is certainly the case for an important part of the data as our dataset contains twice as much phosphorylation data for humans than mouse , and humans are not expected to have more phosphorylation sites than mouse . We reasoned that if state-divergence is caused by false-negatives in the datasets , we would expect to see the fraction of StC to increase as a function of protein abundance , since highly abundant proteins are more likely to be sampled in both species than rare proteins . Indeed , we found that the proportion of state conserved sites almost doubles between the two extreme classes of abundance ( Figure 1D , see also Figure 2A ) . Admittedly , this effect could also be caused by the fact that phosphoregulation is more conserved on highly-expressed proteins but it is unlikely , as it was recently shown that abundant proteins are enriched in non-functional phosphorylation sites [20] that evolve relatively rapidly [17] . In addition , only conserved residues are considered in this analysis . We also examined whether StC or StD phosphorylation sites were more likely to be found in housekeeping or tissue-specific proteins . Housekeeping proteins are expressed in all tissues , while tissue-specific ones are expressed in one or a few tissues . Accordingly , if StD sites are affected by false negatives we would expect to find them preferentially in tissue-specific proteins . We examined the dataset of housekeeping genes [39] and tissue-specific genes [40] and found that StC sites are preferentially found in housekeeping proteins compared to StD sites ( proportions: 0 . 027 vs . 0 . 019 , p-value = 0 . 005 , Figure 1E ) , while the trend is reversed if we look at tissue specific proteins ( proportions: 0 . 268 for state diverged vs . 0 . 219 for StC , p-value = 6 . 1×10−5 , Figure 1E ) . This result is in agreement with our hypothesis that StD sites are affected by false negatives , although this effect could be due to the fact that phosphoregulation is more conserved on housekeeping proteins . In order to examine whether non-functional phosphorylation sites could contribute to poor state-conservation between species , we used a manually curated dataset of functional phosphorylation sites compiled by Landry and collaborators [17] . Functional sites were identified as sites for which a phenotype was observed when phosphorylatable residues were mutated . If non-functional sites contribute to state-divergence , we would expect functional sites to be overrepresented in StC sites . We found that StC sites are enriched in functional phosphorylation sites compared to StD sites ( proportions: 0 . 0025 vs . 0 . 00046 , p-value<1 . 19×10−14 , Figure 1F ) . This observation suggests that a fraction of the StD sites we identified might be non-functional phosphorylation sites , which would explain their poor conservation status between species . It is important to consider that in both cases these observations are not biased by residue conservation as both StC and StD categories are composed of only phosphorylatable residues . Our observation that the majority of StD sites might result from false-negative phosphorylation site identifications or might be non-functional does not rule out the possibility that at least some of these sites could be actual StD sites that diverge in regulation , for instance due to the sequences surrounding the phosphorylated residues . Kinase recognition motifs on substrates are difficult to compare directly due to their degeneracy [21] . We therefore relied on kinase prediction tools for our analyses . We assigned each site to a protein kinase using the NetPhorest classifier [41] to associate protein kinases with all phosphorylation sites based on the site flanking sequences . NetPhorest classification is based on an atlas of consensus sequence motifs that covers 179 kinases and 104 phosphorylation-dependent binding domains and was built using in vivo and in vitro experimental data [41] . If a site is phosphorylated in one species but not in the other , the sequences surrounding the phosphorylatable residue should match a kinase consensus motif better for the phosphorylated site than for the orthologous non-phosphorylated one . Given that NetPhorest provides a score ( from 0 to 1 ) for many possible kinase-substrate associations , we selected the kinase having the best NetPhorest score and we used this score as a proxy to assess the probability of a given site to be phosphorylated . We relaxed this assumption in some of our analyses . In addition , we performed the same analyses directly using a collection of position weight matrices derived from mammalian kinases and the results are in agreement with what we find with the NetPhorest predictions ( Figure S4 ) . We first examined whether there was an association between S/T/Y phosphorylation and NetPhorest scores and found that the probability for a site to be phosphorylated strongly increases with increasing NetPhorest scores in both mouse and human data ( Figure S5 ) . Another result in support of this observation is that the fraction of state conserved sites increases as a function of NetPhorest scores ( Figure 2A ) and this relationship is independent from protein abundance . We also found that prediction scores are very similar for StC sites ( median scores: 0 . 32 for the human phosphorylation sites vs . 0 . 32 in mouse ones , p-value = 0 . 54 ) and higher than those of sites conserved at residue level but non-phosphorylated in both species ( median scores: 0 . 32 for StC vs . 0 . 20 for non-phosphorylated residues , p-value = 2 . 2×10−16; Figure 2B and Figure S6A–B ) . This confirms again a strong association between NetPhorest scores and the probability that a site is phosphorylated . Surprisingly , we found that scores of StC sites were also higher than the scores of the phosphorylated residues in the StD class ( median scores: 0 . 32 vs . 0 . 22 for humans , p-value = 2 . 2×10−16; 0 . 32 vs . 0 . 26 for mouse , p-value = 2 . 2×10−16; Figure 2B–C and Figure S6A–B ) . This means that sites that are conserved and phosphorylated in both species have a significantly better match to consensus kinase motifs than those that are conserved at the residue level but phosphorylated in one species only . There are several possible explanations for these differences . First , this result could derive from how predictive tools have been developed . For instance , phosphorylation sites may be more often studied on abundant proteins , which would imply that kinase prediction tools are better trained at recognizing phosphorylation sites present on abundant proteins . We tested this hypothesis and found that there is no increase in the average NetPhorest scores as a function of protein abundance ( Figure S7 ) , showing that the NetPhorest classification is not biased towards sites present in highly abundant proteins . Another possibility is that StD sites contain a significantly higher proportion of false-positive phosphorylation sites compared to StC sites , as the latter have been found to be phosphorylated in the two species in completely independent experiments and thus have much stronger experimental support . Indeed , false positive sites would have low NetPhorest scores , similar to non-phosphorylated ones and would therefore contribute lowering the average NetPhorest score for the residues that are phosphorylated in StD sites compared to StC sites . A third possibility is that StD sites could contain a proportion of non-functional phosphorylation sites with non-consensus motifs as shown before by Landry and collaborators [17] who found that phosphorylation sites matching kinase motifs have a higher degree of evolutionary conservation and are thus more likely to be functional . Altogether , these results suggest that the match to a consensus sequence motif could be used to the prioritization of phosphorylate sites for downstream functional analysis in phosphoproteomics experiments . Despite these potentially confounding factors , we found evidence that StD is at least partly caused by divergence in regulatory motifs . We found that scores of phosphorylated StD sites are significantly higher than those of their non-phosphorylated orthologous counterparts in both pairwise comparisons ( phosphorylated in human vs . non-phosphorylated in mouse , median scores: 0 . 216 vs . 0 . 214 , p-value = 3 . 93×10−5; phosphorylated in mouse vs . non-phosphorylated in humans , median scores: 0 . 255 vs . 0 . 245 , p-value = 6 . 38×10−5; Figure 2C ) . The fact that we see the effects in both directions rules out the possibility that NetPhorest scores are systematically higher in humans . In order to identify among the set of StD sites the ones that have the potential to be true StD sites , we directly compared matching orthologous NetPhorest scores of StC and StD sites . We found a strong correlation between the NetPhorest scores for StC sites ( rho = 0 . 95 , p-value<2 . 2×10−16 ) and a weaker correlation between the scores of the StD sites , and this both for those phosphorylated in humans but not in mouse ( rho = 0 . 89 , p-value<2 . 2×10−16 , Figure 2D ) and for those phosphorylated in mouse but not in humans ( rho = 0 . 88 , p-value<2 . 2×10−16 , Figure 2E ) . This result is confirmed when comparing the proportion of StD sites having higher scores in humans than in mouse to the same proportion calculated for StC . We found a slight but significant excess of StD sites having higher scores in human than in mouse compared to StC sites ( proportions: 0 . 284 vs . 0 . 258 , p-value = 8 . 69×10−13 , Figure 2F ) . We found similar results for the StD sites having higher scores in mouse compared to humans ( proportions: 0 . 291 vs . 0 . 261 , p-value = 8 . 69×10−11 , Figure 2F ) . By summing up all these excess StD sites that show high NetPhorest scores in one organism but low scores in the other we concluded that that at least 5% of the StD sites ( either phosphorylated in human or mouse ) present in our dataset have the potential to be sites that are differentially regulated between species , despite a conservation of the actual phosphorylatable residues . Our results do not depend on the NetPhorest algorithm as we performed the same analyses using position weight matrices available from the literature [42]–[53] and all of our conclusions about StC and StD sites were mirrored in these tests , as shown in Figure S4 . Overall , our results show that in addition to the actual divergence in phosphorylated sites ( SiD ) , a significant fraction of the mouse and human phosphoproteomes have diverged through changes in the kinase recognition motifs . These changes in the phosphoregulatory status of proteins represent changes in the protein regulatory network , as illustrated for a particular subnetwork in Figure 3 . Potential StD sites are located in proteins that have fundamental cellular functions , making them good candidates for the investigation of species-specific mechanisms of regulation . Further examples are available in Table S2 . We next examined whether the positional turnover of phosphorylation sites could contribute to SiD between mouse and humans . One prediction of this model is that sites that are lost in one lineage could be compensated for by the gain of other sites in the proximity [28] . Similarly , sites could change their positions as a result of insertions and deletions in the surrounding regions . In order to test this prediction , we developed an algorithm to identify evolutionary clustered sites [28] , i . e . pairs of sites that are SiD between mouse and humans and that are closer to each other in the linear protein space than expected by chance alone ( Figure S8 ) . We found that 123 site pairs belonging to 68 proteins show significant evolutionary clustering of SiD phosphorylation sites ( Table S3; alignments are available in Dataset S1 ) . Ninety percent of the proteins that contain evolutionary clustered site pairs have only one or two of them ( Figure S9 ) with few exceptions ( Table S4 ) . This number also excludes proteins for which we found a high number of evolutionary clustered site pairs due to large clusters of sites that we did not consider ( NOL8 , 10; KI67 , 27; MDC1 , 180 site pairs ) . The median NetPhorest score for these sites is 0 . 29 , suggesting that they are likely to be phosphorylated and not false-positives ( 0 . 20 is the median score for non-phosphorylated residues while 0 . 32 is the median score for phosphorylated residues ) . The typical window within which we found significant clustering between SiD sites is 10 amino acids ( Figure S10 ) and approximately 80% of the sites are less than 40 amino acids distant in the alignment . The observed number of site pairs ( n = 123 ) is likely an underestimate of the contribution of evolutionary site turnover because we need many possible configurations in the neutral model to identify them and phosphorproteomes have likely been under sampled . We found that the proportion of proteins that show significant evolutionary clustering increases with the proportion of available sites ( Figure S11 ) . Furthermore , we found that the number of evolutionary clustered sites is correlated with protein size ( rho = 0 . 26 , p-value = 0 . 03 ) and may thus be biased towards large proteins . If these clustered SiD sites were functionally equivalent at the network level between the two species , we would expect them to be phosphorylated by the same kinases or group of kinases . We used again NetPhorest to test this hypothesis . We determined the proportion of StC , StD and evolutionary clustered sites that were likely to be phosphorylated by the same kinases or group of kinases ( overlap of one or more kinases among the three best kinases predicted by NetPhorest ) [19] and we compared these observations to the random expectations obtained by shuffling the mouse kinase-substrate associations . We found that the proportion of StC and StD sites predicted to be phosphorylated by the same kinases or group of kinases was more than 7 times greater than expected by chance alone , suggesting that , globally , these sites tend to be phosphorylated by the same kinases or group of kinases ( Figure 4A–B ) . We found a slightly significant tendency ( p-value = 0 . 03 ) for the evolutionary clustered sites to be phosphorylated by the same kinase ( Figure 4A ) . We then performed the same analysis , but considering the three best kinases found by NetPhorest assuming that phosphorylation sites could be functionally conserved if they are phosphorylated by closely related kinases as well , as in Tan et al . [19] . We found that evolutionary clustered sites were 1 . 4 times more likely to be phosphorylated by the same group of kinases than expected by chance alone ( p-value<0 . 01; Figure 4B ) . This result suggests that , in general , many evolutionary clustered sites may actually be functionally equivalent . Finally , we performed this analysis using position weight matrices available from the literature [42]–[53] and found qualitatively similar results ( Figure S4F ) . Evolutionary clustered sites could arise through losses and gains of phosphorylation sites in the two lineages . Our algorithm identifies evolutionary clustered sites , but it cannot tell whether these represent gains of phosphorylation sites that compensated for deleterious losses in the same lineage or whether they were simply the result of indels that affected the position of the sites in the human and mouse protein alignments . We therefore aligned the mouse and human proteins with several orthologs belonging to species that diverged after the human-mouse divergence ( Figure 5A ) and manually curated the data in order to identify the possible evolutionary steps that led to these configurations of phosphorylation sites . We manually identified many cases ( n = 17 , 14% ) of evolutionary clustered sites that were most likely caused by indels changing protein length and thus alignment . An example is in the Fanconi anemia group M protein , an ATPase implicated in DNA repair [54] in which S1673 and S1674 are shifted towards the C-terminal in the mouse lineage ( Figure 5B ) . The remaining 86% ( n = 106 ) of the cases of evolutionarily clustered sites could not be simply explained by indels and may thus represent compensatory evolutionary events . We observed such a case in the protein DAB2 ( human site: S723; mouse site: S731 ) , which plays a potential role in ovarian carcinogenesis [55] ( Figure 5C ) . The human S723 has been gained after the split of the Haplorrhini from the other primates , while the second one ( S731 ) has been lost after the split between the rodents and the primates . Another example involves the human T4634 and the mouse site S4632 on LRP2 ( Figure 5D ) . This protein is a membrane receptor of absorptive epithelial cells . Mutations in this protein are associated with Donnai-Barrow syndrome , a genetic syndrome that leads to defects in vision , hearing , craniofacial features and structural abnormalities in brain [56] . In this case the human T4634 site appeared in primates after the split from rodents , while the mouse S4632 site was lost after the split of the Strepsirrhini from the other primates . The biological function of these phosphorylation sites has not been determined but they represent prime candidates for exploring , at the molecular level , the positional redundancy of phosphorylation sites . Here we compared the human and mouse phosphoproteomes in order to gain a detailed picture of phosphoregulatory conservation and divergence between these two species . We found that , globally , phosphorylation sites tend to be conserved between human and mouse . By using phosphorylation data from both species , we showed that the number of the sites that are phosphorylated in both human or mouse is 2 . 5 times higher than expected by chance alone . In addition , we estimated phosphorylation status divergence . We found that the majority of phosphorylation sites that are conserved at the residue level between human and mouse are actually divergent with respect of their phosphorylation status ( StD sites ) . While this is most likely largely due to incomplete coverage between the two species , we showed that at least 5% of the StD sites are actually diverging at the kinase-substrate interaction level . We also found that phosphorylation sites that are phosphorylated in both species are more likely to be functional and have higher kinase assignment scores , suggesting that this conservation criterion could be used to prioritize phosphorylation sites for further characterization [17] , [32] . Taken together , these results suggest that more data is needed in these two species to be able to completely assess the conservation and divergence of their phosphoproteomes . Furthermore , the candidate StD sites might have specific regulatory properties that still have to be characterized and understood . A better understanding of these properties will allow us to make an important step towards in our attempt to describe and explain how small regulatory differences map to the important phenotypic differences among species . Mouse is the best model system to study human biology and diseases . It is therefore important to understand how these two species diverge and phosphoregulatory evolution may play an important role . We identified sites that are phosphorylated in one species but that have diverged in the other so that the site is not phosphorylatable ( SiD sites ) . While the biological meaning of the majority of these sites still remains to be assessed , our analysis suggests that many of them could be functionally redundant . This result supports the finding by Moses and collaborators that phosphorylation site evolutionary turnover has a role in shaping phosphoregulation [29] . If the redundancy hypothesis holds true , we might need to revisit estimations of phosphorylation conservation , since omitting positional redundancy may lead to an underestimation of phosphorylation site functional conservation . Moreover , this implies that we should consider different categories of phosphorylation sites: the ones for which the position along the protein is a determinant for their function ( positionally-dependent phosphorylation sites ) and those for which the global charge rather than the exact position is responsible for their function ( positionally-flexible phosphorylation sites ) .
An extensive dataset of human and mouse phosphorylation sites was built by combining data from 7 different databases and experimental studies [6] , [32]–[37] . All protein sequences and orthology relationships were retrieved from ENSEMBL ( version 69 ) . In this study , only protein sequences for which we could find orthology relationships between a human protein and at least a mouse , dog and opossum protein were considered . This step allowed us to study the evolutionary history of phosphorylation sites . For humans and mouse , orthology relationships were determined for the longest isoforms of each protein . Each group of orthologous sequences was aligned using MUSCLE [57] . Disordered and ordered regions of proteins were predicted using DISOPRED [58] . In order to map phosphorylation sites to our sequences , the following procedure was applied . The sites that were already mapped onto proteins associated with ENSEMBL IDs in the original datasets were directly mapped to our sequences . For all other cases , phosphopeptides were mapped onto proteins using BLAT [59] . All peptides that mapped to more than one protein were removed at this step . Mapped phosphorylation sites and information about protein disorder are available in Dataset S2 . In order to calculate the random expectation for the number of sites belonging to each one of the different categories ( StC , StD and SiD ) , statuses ( 0: non-phosphorylated , 1: phosphorylated ) of phosphorylatable amino acid were shuffled in each protein by preserving the overall proportion of sites for each residue ( S , T or Y ) and the localization in disordered/ordered regions . The null distributions were estimated by iterating this procedure 100 times , calculating each time the number of sites belonging to each category . We calculated random expectations by shuffling the mouse sites only . We also performed the calculations by independently shuffling both human and mouse sites and found similar results . Data on protein abundance were taken from PaxDb [60] ( H . sapiens whole organism integrated dataset ) . In the analysis presented on Figure 1D , proteins were ordered by their abundance and divided in four equal bins . Data on housekeeping genes were retrieved from Eisenberg and Levanon [39] who identified 575 human genes that are expressed in 47 different tissues and cell lines based on microarray data . Data on tissue-specific genes derive from an independent dataset and were retrieved from the TiGER database [40] . About 5 . 3 millions human ESTs were mapped to UniGene clusters and the expression pattern of the all UniGenes in 30 human tissues was determined using the NCBI EST database . 7261 tissue-specific genes were identified . Manually curated data on functional phosphorylation sites ( n = 156 ) were retrieved from Landry et al . [17] . These sites were derived from the manual curation of the primary literature . NetPhorest [41] was downloaded from ( http://netphorest . info ) and was run locally using default options . In order to calculate position weight matrices scores , 29 position weight matrices which scores are based on the same metric were obtained from Benjamin Turk [42]–[53] . These matrices were used to score all 10-mer amino acids in the mouse and human proteomes that have a phosphorylatable amino acid on the sixth position . The score reflects the probability of each 10-mer to be phosphorylated by a specific kinase . Proportions were compared with 2-sample tests for equality of proportions with continuity correction . Distributions were compared with non-parametric Wilcoxon Rank Sum tests . Correlations were calculated with the Spearman method . All these statistical analyses were performed as implemented in R [61] . Site colocalization in orthologous proteins was estimated using a window of positions ( centered on each human phosphorylation site ) . The fraction of colocalized sites over the total number of sites was calculated for a range of window sizes . In order to determine which sites were closer in sequence linear space than expected by chance alone , the mouse phosphorylation sites were shuffled in each protein by preserving the overall proportion of sites for each residue ( S , T or Y ) and disordered/ordered regions , and the fraction of colocalized sites was calculated for each window length . One thousand iterations were performed in order to generate the null model . Also , we masked all the positions in which a phosphorylatable amino acid was present at a given position in both human and mouse . Evolutionary clustered sites were defined as sites that were more likely to be colocalized than expected by chance alone ( null model ) . The closest pair of phosphorylation sites present in these windows was then selected ( see also Figure S8 ) . The phosphorylatable amino acids serine ( S ) and threonine ( T ) differ in biochemical properties compared to tyrosine ( Y ) , another phosphorylatable amino acid [62] . Therefore , S/T and Y sites were considered as belonging to separate classes and not considered to be able to compensate each other . Only 1529 pairs of orthologous proteins that had at least two phosphorylation sites that diverged ( site-divergence ) in human and mouse respectively were considered . Among these pairs , 563 had at least one SiD site that involves a phospho-serine or phospho-threonine in both humans and mouse . Only one single pair had a SiD site that involves a phospho-tyrosine in both humans and mouse . The kinase that was the most likely to phosphorylate each one of the evolutionary clustered sites was inferred using NetPhorest [41] and proportion of evolutionary clustered site pairs phosphorylated by the same kinase was determined . This number was compared to a null distribution obtained by randomly shuffling ( 10 , 000 iterations ) the kinase-phosphorylation site associations between different evolutionary clustered sites . Analogous analyses were performed for StC and StD sites . We then performed the same analysis but this time using the three best kinases predicted by NetPhorest , as proposed by Tan et al . [19] . We therefore considered two evolutionary clustered sites as being phosphorylated by the same group of kinases if they shared one or more kinases ( kinase group ) among the three best kinases predicted to be associated to each site according to NetPhorest . This number was compared to a null distribution obtained by randomly shuffling ( 100 iterations ) the kinases-phosphorylation site associations between different evolutionary clustered sites . Analogous analyses were performed for StC and StD sites . Finally , we performed again all the analyses described above but this time using position weight matrices from the literature ( see section NetPhorest and position weight matrices scores for further details ) instead of NetPhorest to infer the kinase that was the most likely to phosphorylate each one of the StD , StC and evolutionary clustered sites .
|
Understanding how differences in cellular regulation lead to phenotypic differences between species remains an open challenge in evolutionary genetics . The extensive phosphorylation data currently available allows to compare the human and mouse phosphoproteomes and to measure changes in their phosphoregulation . We found a general conservation of phosphorylation sites between these two species . However , a fraction of sites are conserved at the sequence level ( the same amino acid is present in both species ) but differ in their phosphorylation status . These sites represent candidate sites that have the potential to explain differences between human and mouse signalling networks that do not depend on the divergence of orthologous residues . Furthermore , we identified several sites where to a phosphorylation site in one species corresponds a non-phosphorylatable residue in the other one . These cases represent clear differences in protein regulation . Recent studies suggest that phosphorylation sites can shift position during evolution , leading to configurations in which pairs of divergent phosphorylation sites are functionally redundant . We identified more than 100 putative such cases , suggesting that divergence in amino acid does not necessarily imply functional divergence when comparing phosphoproteomes . Overall , our study provides new key concepts and data for the study of how regulatory differences may be linked to phenotypic ones at the network level .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Methods"
] |
[
"sequence",
"analysis",
"systems",
"biology",
"evolutionary",
"biology",
"genomic",
"evolution",
"natural",
"selection",
"gene",
"regulation",
"regulatory",
"networks",
"molecular",
"genetics",
"comparative",
"genomics",
"biology",
"genomics",
"population",
"genetics",
"signaling",
"networks",
"computational",
"biology",
"evolutionary",
"genetics"
] |
2014
|
Functional Divergence and Evolutionary Turnover in Mammalian Phosphoproteomes
|
For nearly 20 years , the principal biological function of the HIV-2/SIV Vpx gene has been thought to be required for optimal virus replication in myeloid cells . Mechanistically , this Vpx activity was recently reported to involve the degradation of Sterile Alpha Motif and HD domain-containing protein 1 ( SAMHD1 ) in this cell lineage . Here we show that when macaques were inoculated with either the T cell tropic SIVmac239 or the macrophage tropic SIVmac316 carrying a Vpx point mutation that abrogates the recruitment of DCAF1 and the ensuing degradation of endogenous SAMHD1 in cultured CD4+ T cells , virus acquisition , progeny virion production in memory CD4+ T cells during acute infection , and the maintenance of set-point viremia were greatly attenuated . Revertant viruses emerging in two animals exhibited an augmented replication phenotype in memory CD4+ T lymphocytes both in vitro and in vivo , which was associated with reduced levels of endogenous SAMHD1 . These results indicate that a critical role of Vpx in vivo is to promote the degradation of SAMHD1 in memory CD4+ T lymphocytes , thereby generating high levels of plasma viremia and the induction of immunodeficiency .
The Vpx accessory protein is encoded by HIV-2 , related SIVsm strains , SIVmnd , and SIVrcm [1–4] . Vpx has been reported to antagonize restriction imposed by SAMHD1 in cultured myeloid lineage ( dendritic cells , monocytes , and macrophages ) and quiescent CD4+ T cells [5–8] . Early studies also showed that SIVmac239 , carrying vpx gene deletions , exhibited an attenuated replication phenotype in inoculated macaques [9 , 10] . It is presently unclear whether compromised infection of myeloid lineage cells in vivo is responsible for this phenotype or if endogenous SAMHD1 must also be suppressed in memory CD4+ T lymphocytes , the cell lineage that sustains high levels of set-point viremia attending pathogenic infection . Although the HIV-1 genome does not encode Vpx , most studies assessing Vpx degradation of SAMHD1 during virus infections have utilized pseudotyped HIV-1 constructs , in combination with SIV VLPs expressing Vpx , in single-cycle replication assays . Only a single study has utilized replication-competent HIV-1 to monitor Vpx-mediated suppression of SAMHD1 during an in vitro infection . In that experiment , SAMHD1 was reported to block virus infection in resting human CD4+ T lymphocytes unless SIVmac239 Vpx was co-packaged into an HIV-1 expressing GFP construct [5] . However , even though SAMHD1 levels had been markedly depleted and HIV-1 directed GFP expression became detectable intracellularly in the presence of Vpx , no progeny virions were produced . The relevance of these in vitro functional studies of Vpx to the induction of immunodeficiency during pathogenic infections of macaques with SIVsm strains , such as SIVmac , in which the vpx gene is an intrinsic and evolutionarily conserved element , is not clear . It has been suggested that the antiviral activity of endogenous SAMHD1 may be limited to non-cycling cell lineages such as terminally differentiated myeloid cell subsets or , more recently , quiescent CD4+ T lymphocytes . Non-cycling memory CD4+ T lymphocytes are , in fact , the principal targets of both HIV and SIV during the initial weeks of the acute in vivo infection . Prodigious numbers of resting memory CD4+ T cells become infected in lymphoid tissues and blood and large amounts of circulating progeny virions are produced during this phase of the infection [11–13] . Furthermore , the relatively low levels of set point viremia and slow disease progression previously reported in rhesus macaques inoculated with SIV Vpx deletion mutants [9 , 10] suggests that Vpx may also be functionally important in counteracting SAMHD1 in virus-producing CD4+ memory T lymphocytes during the later chronic phase of the in vivo infection . Here we examine replication-competent SIV Vpx mutants , disabled in their capacity to degrade SAMHD1 . We show that when macaques were inoculated with SIV carrying the Q76A Vpx point mutation , which specifically affects the interaction of Vpx with DCAF1 and the subsequent recruitment of SAMHD1 for degradation , virus acquisition , progeny virion production in memory CD4+ T lymphocytes during the acute infection , and the maintenance of set point viremia were greatly attenuated . Revertant viruses , which emerged in two infected animals , carried substitutions located in likely contact points of Vpx with the c-terminal domain of DCAF1 . Thus our data indicate that contrary to the commonly held belief that the principal function of SIV Vpx is to facilitate virus replication in myeloid lineage cells , the need to degrade endogenous SAMHD1 during SIV infections of memory CD4+ T cells in vivo is critically important and drives the selection for Vpx revertant viruses , capable of mediating the degradation of SAMHD1 and generating high levels of plasma viremia .
Several earlier studies have reported that replication competent HIV-2 and SIVmac mutants , unable to express the Vpx protein , exhibit delayed infection kinetics and low levels of progeny virus production in cultured rhesus peripheral blood mononuclear cells ( PBMC ) [9 , 14 , 15] . To ascertain whether a Vpx point mutant , specifically defective in recruiting DCAF1 and subsequently degrading SAMHD1[7 , 16–18] , might possess similar properties , the X-Q76A Vpx mutant , carrying a 2-nucleotide substitution , was constructed ( Fig 1A ) . A second Vpx mutant ( X-del ) , containing a TAA stop codon at residue 2 that prevents the synthesis of any SIV Vpx protein was also prepared ( Fig 1A ) . Both Vpx mutations were introduced into molecular clones of wild type ( WT ) T cell tropic ( SIVmac239 ) or WT macrophage tropic ( SIVmac316 ) SIVs . Wild type and Vpx mutant SIV inocula were prepared by transfecting full-length infectious molecular clones into 293T cells as described in Methods . Vpx expression in HeLa cells and its incorporation into progeny virions released into the transfection supernatant medium were confirmed by immunoblotting ( Fig 1B ) . Compared to the WT viruses , both types of SIVmac239 or SIVmac316 Vpx mutants generated reduced amounts of progeny virions during infections of cultured Concanavalin A ( ConA ) stimulated rhesus PBMC ( Fig 2A and 2B ) . The defective replication phenotype of the Vpx mutants was more profound during infections of rhesus monocyte derived macrophage ( MDM ) infections ( Fig 2C and 2D ) . In contrast to the robust replication of the WT macrophage tropic SIVmac316 in rhesus MDM , replication of both the SIVmac316 Vpx deletion mutant and the SIVmac316 Vpx point mutant was blocked in these cells . As expected , neither the WT nor the Vpx mutants of the T cell tropic SIVmac239 were able to infect rhesus MDM ( Fig 2C ) . The SIVmac239 and SIVma316 X-Q76A Vpx point mutants both exhibited replication kinetics in human SupT1-R5 cells indistinguishable from corresponding WT viruses ( Fig 2E and 2F ) , consistent with results reporting that SupT1-R5 cells do not express the SAMHD1 protein [8] , and confirming the absence of any inherent replication defects in these mutant viruses . As noted earlier , a majority of previous studies evaluating virion-associated Vpx-mediated degradation of SAMHD1 have been conducted in non-lymphoid cells using pseudotyped virus preparations capable of only single cycles of replication . To determine whether Vpx-mediated degradation of endogenous SAMHD1 occurred during the course of spreading infections of replication-competent virus in CD4+ T lymphocytes , freshly collected , and negatively selected , Con-A stimulated rhesus macaque CD4+ T cells were infected with WT SIVmac239 at a multiplicity of infection ( MOI ) = 0 . 2 . Cells and supernatant samples were collected daily and examined for levels of progeny virions released into the medium ( 32P-reverse transcriptase [RT] activity ) and endogenous SAMHD1 ( immunoblotting ) . Newly produced virus first became detectable on day 2 post infection ( PI ) and steadily increased on days 3 and 4 ( Fig 3A ) . The levels of endogenous SAMHD1 , present in the CD4+ T cells on days 1 and 2 , markedly declined on days 3 and 4 PI when compared to an unrelated cellular protein ( GAPDH ) ( Fig 3B ) . In an independent experiment , also assessing the status of endogenous SAMHD1 , ConA-activated rhesus CD4+ T lymphocyte cultures were infected with WT SIVmac239 , WT SIVmac316 , or the two different and corresponding Vpx defective mutants , all at a MOI = 0 . 2 . Based on the results of the experiment shown in Fig 3A and 3B , cells were collected on day 3 PI and lysates were examined by immunoblotting for levels of endogenous SAMHD1 . As shown in Fig 3C , Con A-stimulated rhesus CD4+ T cells , infected with both of the WT SIVmac viruses , contained reduced levels of endogenous SAMHD1 compared to levels in mock infected cells or in cells infected with the corresponding Vpx mutant viruses . Together , these results demonstrate that SIV Vpx mutants bearing the Q76A point mutation , which specifically blocks the recruitment of DCAF1 , are defective in degrading endogenous SAMHD1 ( Fig 3 ) and are attenuated during spreading infections in cultured activated PBMC ( Fig 2 ) compared to their WT counterparts . It has recently been reported that phosphorylation of human SAMHD1 at the threonine 592 residue ( Thr592 ) greatly reduces its capacity to restrict HIV-1 replication , but does not affect its dNTPase activity [19–21] . To first verify that the Thr592 residue was actually present in rhesus SAMHD1 , mRNAs were prepared from a mixture of PBMC samples collected from six monkeys , amplified by RT-PCR , and individual amplicons were analyzed by nucleotide sequencing . Although rhesus and human SAMHD1 proteins differed at 38 of 626 amino acid positions , the Thr592 residue was present in both proteins ( S1 Fig ) . It was therefore of interest to ascertain the phosphorylation status of SAMHD1 in primary rhesus mononuclear cells , particularly in the Con A-activated rhesus PBMC in which SIV production was shown modestly reduced in the absence of Vpx ( Fig 2A and 2B ) . Cell lysates from either non-activated or ConA activated CD4+ T cells , from two different macaques , were subjected to electrophoresis in Phos-tag Acrylamide and immunoblotting using the anti-SAMHD1 antibody . As shown in Fig 3D , a significant fraction of SAMHD1 was phosphorylated in ConA activated rhesus CD4+ T cells , whereas in unstimulated CD4+ T cells , the majority of SAMHD1 was unphosphorylated . Taken together , the increased levels of phosphorylated nonrestrictive SAMHD1 in the stimulated PBMC cultures may very well have minimized the differences between the infectivities of WTand the Vpx mutants measured in these cells . In vivo , the principal target of both HIV-1 and SIVmac are memory CD4+ T lymphocytes [11 , 22 , 23] . The levels of SAMHD1 in unstimulated rhesus macaque memory CD4+ T cells , freshly collected from either blood or spleen , were examined by immunoblotting using anti-human SAMHD1 antibody . As shown in Fig 4A , higher levels of SAMHD1 were detected in memory CD4+ T lymphocytes than those measured in naïve CD4+ T cells from the same two sources . Among rhesus myeloid lineages , circulating CD14+ monocytes expressed levels of SAMHD1 similar to those present in memory CD4+ T cells , whereas much higher concentrations of SAMHD1 were detected in macaque alveolar macrophage , collected by bronchoalveolar lavage , and rhesus MDM derived in vitro . Cell lysates from the same preparations of freshly collected unstimulated rhesus macaque memory CD4+ T cells from blood or spleen were subjected to electrophoresis in Phos-tag Acrylamide and immunoblotting using the anti-SAMHD1 antibody . As shown in Fig 4B , most of the SAMHD1 in memory CD4+ T cells was unphosphorylated . The relatively high levels of expression of unphosphorylated endogenous SAMHD1 measured in memory CD4+ T lymphocytes suggested that SAMHD1 might significantly restrict viral infections in vivo , reduce virus production systemically , and affect disease progression if not adequately suppressed by Vpx . To specifically examine whether Vpx can degrade endogenous SAMHD1 in rhesus memory CD4+ T cells , a large preparation of this subset was purified by FACS and infected with VSV-G pseudotyped WT SIVmac239 or its X-del or X-Q76Aderivatives; SAMHD1 expression was evaluated 24 h later by immunoblotting . As shown in Fig 4C , the levels of SAMHD1 in WT SIVmac239 infected memory CD4 T cells were markedly reduced compared to that in cells infected with the Vpx mutants . More importantly , to directly ascertain whether Vpx also degrades endogenous SAMHD1 in memory CD4+ T cells in vivo , we next determined if SAMHD1 degradation could be detected in memory CD4+ T lymphocytes , collected from SIVmac239 infected macaques . During the first weeks of the SIV acute infection , CD4+ memory T lymphocytes are massively infected systemically [11] . For example , on day 10 post inoculation , levels of cell-associated SIV DNA were reported to be in the range of 1 x 105 copies/105 memory CD4+ T cells in PBMC , inguinal and mesenteric lymph nodes , and jejunum mucosa . The establishment of such a prodigious in vivo infection provided a window of opportunity to directly examine the status of endogenous SAMHD1 in memory CD4+ T cells during the acute infection . The infection kinetics of SIVmac239 in two previously described rhesus monkeys ( 95D132 and H589 ) [24] , inoculated intravenously ( IV ) with 1 × 104 TCID50 of SIVmac239 is shown in Fig 4D . Based on this result , a macaque was inoculated IV with 1 × 104 TCID50 of WT SIVmac239 , sacrificed on day 9 PI , and memory CD4+ T cells from PBMC and spleen were collected by flow cytometric sorting . Memory CD4+ T lymphocytes were similarly prepared from the PBMC and spleen of an uninfected monkey . Immunoblotting revealed markedly reduced levels of SAMHD1 in memory CD4+ T cells purified from PBMC and spleen in the day 9 infected animal compared to those present in similar cells from the uninfected monkey ( Fig 4E ) . This result demonstrates that levels of endogenous SAMHD1 in memory CD4+ T lymphocytes are greatly diminished at the time of peak virus production in vivo , presumably reflecting SIVmac239 LTR-directed expression of Vpx . An extensive literature exists reporting that SIVmac and HIV-2 Vpx is a major facilitator of virus replication in cultures of terminally differentiated myeloid cells [10 , 25–27] . In an attempt to translate this particular cell-dependent Vpx function to the organismal level , an experiment was designed to direct WT and Vpx-deficient viruses to at least one myeloid lineage cell type ( viz . dendritic cells ) by using a mucosal rather than an IV route of virus inoculation . Accordingly , two macaques ( DCXX and J7L ) were challenged intrarectally ( IR ) with 1 × 103 TCID50 of WT SIVmac239 and two macaques ( J3L and J5R ) were inoculated by the same route with 1 × 103 TCID50 of WT SIVmac316 . Virus acquisition only occurred in the two WT SIVmac316 challenged monkeys ( J3L and J5R ) ( Fig 5A and 5B ) . After waiting 6 additional weeks , the same two macaques ( DCXX and J7L ) , which had previously been inoculated with 1 x 103 TCID50 of WT SIVmac239 , were re-inoculated intrarectally with 1 x 104 TCID50 of SIVmac239 . Both monkeys rapidly became infected and generated high levels of peak plasma viremia ( Fig 5A ) . Based on these results , we elected to inoculate four additional macaques intrarectally with 1 x 104 TCID50 of the SIVmac239 X-Q76A Vpx mutants and four other animals with 1 x 103 TCID50 of the SIVmac316 X-Q76A Vpx mutant to assess the infectivities of the Vpx point mutant viruses in vivo . Both macaques inoculated with WT SIVmac239 generated peak plasma viremia levels of 1 . 3 × 108 RNA copies/ml between days 10 and 14 PI , and developed viral set points of 2 . 9 x 106 and 6 . 0 x 106 RNA copies/ml , respectively ( Table 1 ) . The four monkeys ( JWR , JHL , K2M , and K42 ) inoculated by the IR route with the SIVmac239 Vpx Q76A point mutant all exhibited attenuated replication phenotypes ( Fig 5A and Table 1 ) . In three of these animals ( JWR , K2M , and K42 ) , peak plasma viral loads ranged from 3 . 40 × 105 to 2 . 66 × 106 RNA copies/ml , 50 to 400-fold lower levels than those measured in WT SIVmac239 infected macaques . In two of these monkeys , the time of peak virus production was delayed to day 17 or day 28 PI . One of the animals ( JHL ) required 3 successive IR inoculations of the SIVmac239 Vpx Q76A mutant , spaced 6 weeks apart , to establish the SIV infection; a relatively low peak plasma viremia ( 2 . 25 x 105 RNA copies/ml ) in this macaque was delayed until day 21 post challenge . It should also be noted that the viral set points in three of the animals inoculated with the SIVmac239 Vpx point mutant were markedly reduced compared to monkeys infected with WT virus ( Fig 5A ) , falling below the level of detection in one macaque ( JHL ) by week 12 PI . As shown in Fig 5C , memory CD4+ T cells rapidly declined in the two monkeys inoculated with WT SIVmac239 but did not change appreciably in animals infected with the Vpx mutants . The memory CD4+ T cell subset , as a percentage of total CD4+ T cells in these two infected animals , rapidly declined compared to levels in the 4 macaques inoculated with the Vpx mutants ( S2A Fig ) . A similar result was obtained with the WT macrophage tropic SIVmac316 and its Vpx Q76A mutant . As noted above , the two animals ( J3L and J5R ) inoculated with 1 × 103 TCID50 of WT SIVmac316 became infected following a single challenge and developed levels of peak plasma viremia of 1 . 62 × 107 and 9 . 00 × 106 RNA copies/ml on days 21 and 24 PI , respectively ( Fig 5B and Table 1 ) . The set point levels of viremia ( 3 . 0 × 104 and 1 . 0 × 105 copies/ml ) in these two monkeys were somewhat lower than those measured in the macaques inoculated with WT SIVmac239 . When the SIVmac316 Vpx Q76A mutant was similarly evaluated in four animals , its replication properties were even more severely debilitated than the analogous SIVmac239 Vpx mutants ( Fig 5B and Table 1 ) . An infection was established in two of these four monkeys ( JLP and DX39 ) following a single IR challenge , but the peak viral loads ( 7 . 05 ×104 and 9 . 08 × 104 RNA copies/ml ) were 2 logs lower than that measured with WT SIVmac316 and were markedly delayed ( until day 35 and day 52 post inoculation , respectively ) . In addition , the viral set points at weeks 22 to 24 in these two macaques were quite low ( 4 . 33 × 102 and 2 . 00 × 103 RNA copies/ml ) . Establishment of an SIVmac316 infection in the two remaining animals receiving the SIVmac316 Q76A Vpx mutant virus ( JA4X and K31 ) required 2 and 3 successive inoculations , respectively , with a peak viral load reaching only 3 . 7 × 103 RNA copes/ml at day 77 PI in the latter macaque ( Fig 5B and Table 1 ) . Taken together , these results indicate that the Q76A Vpx mutation is profoundly disabling in vivo , affecting SAMHD1 degradation , SIVmac acquisition , the production of progeny virions during the acute infection , and the maintenance of set-point viremia , all of which are replication functions that occur in memory CD4+ T cells . A modest reduction of memory CD4+ T lymphocytes occurred in the two monkeys inoculated with WT SIVmac316 but not in the animals infected with the Vpx mutants ( Fig 5D ) . The memory CD4+ T cell subset in these two macaques , as a percentage of total CD4+ T cells , declined somewhat compared to levels measured in the monkeys inoculated with the Vpx mutants ( S2B Fig ) . During the chronic phase of their infections , one of the four recipients of the SIVmac239 Vpx mutant ( macaque K42 ) and one of the four recipients of the SIVmac316 Vpx mutant ( macaque JA4X ) developed elevated set-point viremia levels that distinguished them from the six other monkeys inoculated with the Q76AVpx mutant ( indicated by the blue curves in Fig 5A and 5B ) . In fact , the set-point virus load in animal JA4X was similar to those generated by the two recipients of WT SIVmac316 ( Fig 5B ) . It should be noted that to minimize the emergence of revertant viruses during infections in vivo , a Q76A Vpx point mutant , containing a two nucleotide substitution , was purposely constructed . The possible emergence of Vpx revertant viruses was initially investigated by performing single genome amplification ( SGA ) analyses of plasma samples collected at week 35 PI from all 8 monkeys inoculated with the SIVmac239 or the SIVmac316 Q76A Vpx point mutants ( Fig 6 ) . Nucleotide sequence analyses revealed that 13 of 14 vpx gene amplicons from the putative revertant virus circulating in macaque K42 had acquired an A76S ( GCA to TCA ) substitution at the site of the original Q76A mutation in the SIVmac239 Vpx mutant . In the case of the putative revertant virus recovered from monkey JA4X , all of the week 35 amplicons had retained the original Q76A mutation; however , 25 of 27 carried a “second site” I32T ( ATT to ACA ) change . Both putative Vpx revertant viruses also had acquired a G19E substitution ( Fig 6 ) . However , the latter was the only change present in some vpx gene amplicons from macaque K2M ( Fig 6 ) and this single amino acid substitution failed to restore high levels of set-point viremia in the animal ( see Fig 5A ) . The “second site” I32T Vpx substitution identified in virus circulating in macaque JA4X would be consistent with and support a recent genetic study reporting that Ile32 is a critical residue mediating Vpx and DCAF1 interactions [28] . The Vpx amino acid changes associated with revertant viruses isolated from macaques JA4X and K42 are located in helix 1 and helix 3 , respectively , and are shown diagrammatically in Fig 5E . To ascertain whether revertants had , in fact , emerged in monkeys JA4X or K42 , virus stocks were prepared from each animal by co-cultivating their PBMC , collected at week 40 PI , with SupT1-R5 cells . We then confirmed that these recovered swarm virus stocks ( K42 and JA4X ) had each retained their respective putative revertant mutation by SGA analysis: 10 of 10 amplicons from the K42 stock carried the Vpx A76S revertant change; 15 of 15 amplicons from the JA4X stock contained the original Vpx Q76A mutation as well as the I32T substitution ( S3 Fig ) . The recovered viruses were then assayed for infectivity in rhesus macaque PBMC , side-by-side , with the corresponding starting Q76A Vpx mutant viruses . In contrast to the original Q76A Vpx mutants , which exhibited attenuated replication phenotypes , the viruses isolated from animals K42 or JA4X had robust infection kinetics in rhesus PBMC and released high levels of progeny virions ( Fig 7A and 7B ) . We next determined whether the capacity of the putative revertant viruses to degrade endogenous SAMHD1 was restored during productive in vitro infections . Rhesus CD4+ T lymphocyte cultures were prepared as described for Fig 3C and infected with the viruses recovered from monkeys K42 or JA4X , the parental WT SIVmac239 , WT SIVmac316 , and the two corresponding Q76A Vpx mutants . Infected cells were collected at day 3 PI and lysates were examined by immunoblotting for endogenous SAMHD1 . As shown in Fig 7C , macaque CD4+ T lymphocytes infected with the two putative revertant viruses and both WT viruses contained lower levels of endogenous SAMHD1 compared to that present in cells infected with the starting Q76A Vpx mutant viruses . These results indicate that viruses recovered from macaques K42 and JA4X had re-acquired the capacity to degrade endogenous SAMHD1 as well as to release large amounts of progeny virions during spreading infections in Con-A activated rhesus CD4+ T cells . The functional significance of the Vpx substitutions in revertant virus populations that had emerged was directly assessed by inserting the changes shown in Fig 5C into the genetic background of SIVmac316 and evaluating the replication phenotypes of the resultant constructs during infections of rhesus PBMC . As shown in Fig 7D , the X-Q76S revertant virus released nearly three-fold more progeny virions than the starting X-Q76A Vpx mutant in rhesus PBMC . The Q76A/I32T second-site revertant virus was less robust and generated intermediate amounts of virus . These results indicate that both revertant changes augmented virus replication in rhesus PBMC compared to the initial attenuated Q76A Vpx mutant . A recent structural study of Vpx and DCAF1 reported that residues Ile32 and Gln76 of Vpxsm both make contact with Trp1156 of the DCAF1 cytoplasmic domain [17] . Vpx residues Ile32 and Gln76 are in close physical proximity within the SAMHD1- DCAF1-Vpx ternary complex with Ile32 making hydrophobic interactions with the aliphatic base of the Gln76 side chain . The carbonyl side chain of Vpx residue Gln76 makes H-bond interactions with DCAF1 Trp1156 and the NH2 side chain of Vpx Gln76 bonds with the main chain carbonyl of DCAF1 Asn1135 ( Fig 8A and 8B ) . In addition , the main chain carbonyl of Vpx Gln76 hydrogen bonds with a water molecule ( W1 ) that mediates extensive contacts between Vpx and DCAF1 . Thus Gln76 appears to play a critical role at the Vpx-DCAF1 interface . Modeling the Gln76Ala disabling mutation in Vpx used in this study on the Vpx-DCAF1-SAMHD1 crystal structure places DCAF1 Trp1156 in a predominantly unfavorable hydrophobic environment and creates a potentially destabilizing void at the interface , and also results in loss of interactions mediated by Gln76 in WT ( Fig 8C ) . Substituting the Vpx Ala mutation at position 76 with Ser , as is present in the K42 revertant virus , is predicted to partially fill the cavity created by the original Gln76Ala change and to restore some of the interactions with DCAF1 ( Fig 8D ) . Similarly , the Ile32Thr change , in the presence of the original Vpx Ala mutation , is predicted to restore a hydrogen bond interaction between Vpx and DCAF1 and also stabilizes the interface providing a complementary polar site for DCAF1 Trp1156 to interact with ( Fig 8E ) . Thus by recreating a shared surface to which SAMHD1 binds , Vpx substitutions that are predicted to facilitate interaction with DCAF1 augment SIV replicative capacity in vivo .
Our results demonstrate that Vpx is critically important for countering SAMHD1 during SIVmac infections of memory CD4+ T lymphocytes both in vitro and in vivo . SIVs carrying the Q76A Vpx point mutant , which specifically blocks the interaction of Vpx with DCAF1 and the subsequent recruitment of SAMHD1 to the Cullen 4A-ring ubiquitin ligase complex , exhibited an attenuated replication phenotype during infection of ConA activated rhesus macaque PBMC and were deficient in degrading endogenous SAMHD1 in cultured rhesus CD4+ T lymphocytes . In a monkey inoculated with WT SIVmac239 , levels of endogenous SAMHD1 were markedly reduced in circulating and tissue-associated memory CD4+ T cells on day 9 PI of the acute infection , compared to the levels measured in lymphocytes from an uninfected animal . When macaques were inoculated with the T cell tropic SIVmac239 or the macrophage tropic SIVmac316 carrying the Q76A Vpx point mutation , virus acquisition was markedly impaired: multiple IR inoculations were required to establish infections and peak levels of virus production were both delayed and reduced . The maintenance of set point viremia was also severely attenuated in monkeys inoculated with the SIV Vpx point mutation unable to recruit DCAF1 . Revertant viruses , which emerged in two recipients of the Vpx point mutant , carried either an A76S change at the original mutation site or an I32T “second site” change , two substitutions located in likely contact points of Vpx with the C-terminal domain of DCAF1 . Both SIVmac Vpx revertants exhibited an augmented replication phenotype in vitro and in vivo and were able to degrade SAMHD1 . These results point to a requirement for Vpx to maintain robust SIVmac replication in memory CD4+ T cells during all phases of the in vivo infection and provide evidence that selective pressure will be exerted in these cells to restore activity occurs if Vpx function is compromised . Although it has been previously reported that an SIVmne mutant , carrying Vpx changes that prevent binding to DCAF1 , exhibited attenuated infectivity in pig-tailed macaques [29] , our study is the first to show that expression of functional Vpx during the acute SIV infection causes near complete SAMHD1 depletion in memory CD4+ T cells in vivo . It must be noted that the Q76A Vpx mutation does not specifically block the interaction of Vpx with SAMHD1 , but rather , the binding of Vpx to DCAF1 . This may lead to the subsequent recruitment of SAMHD1 to the Cullen 4A-ring ubiquitin ligase complex . On the other hand , the possibility exists that abrogating the binding of Vpx with DCAF1 may allow DCAF interacting cellular partners , other than SAMHD1 , to disable SIV replication [30–32] . It is important to point out that 90 to 95% of sorted CD3+ , CD4+ , CD8- , CD95+ memory cells collected from uninfected rhesus macaques are not activated , as measured by Ki67 staining [12] . The median % activation of CD4+ T lymphocytes on day 14 post SIVmac239 infection has been reported to be 20% [33] . Thus , a majority of T cells targeted by SIV in vivo are not dividing . In this regard , SAMHD1 has been previously reported to suppress virus infection in resting but not in cycling CD4+ T cells [5] . The dependence of Vpx restriction on cell activation status was subsequently shown to be due to the phosphorylation state of SAMHD1 [19–21] . This is shown in Fig 3D for SAMHD1 , which becomes hyperphosphylated as a consequence of activating rhesus CD4+ T cells with the plant lectin ConA . Since infectivity assays required the use of rhesus PBMC , activated with Con A to achieve demonstrable virus replication in vitro , the relatively modest attenuated replication phenotype of the SIVmac Q76A Vpx mutants compared to WT SIVmac repeatedly observed in rhesus PBMC ( Figs 2 and 7 ) might simply reflect the increased levels of phosphorylated nonrestrictive SAMHD1 in the stimulated PBMC cultures . It is now recognized that the gastrointestinal ( GI ) tract is a major site of HIV-1 and SIV replication and the number of CD4+ T cells is markedly reduced during all phases of the virus infection [34 , 35] . Despite antiretroviral treatment , immune reconstitution in the GI mucosa is variable and incomplete [36 , 37] . A recent study of 5 monkeys , chronically infected for 2 to 4 years with an SIVmac239 mutant carrying a 101 base pair deletion of the vpx gene , reported that virtually no virus infected macrophages were present in lymph node , spleen and colon specimens , collected at the time of their death from immunodeficiency [38] . This result would be consistent with numerous reports showing that Vpx expression is required for efficient replication of SIVmac in myeloid lineage cells . This study had an even more interesting finding . In contrast to animals infected with wt SIVmac239 , which experience robust virus replication in the gut-associated lymphoid tissue ( GALT ) and associated severe sustained depletion of CD4+ T cells at this site , the GALT of macaques infected with the SIV Vpx mutant contained very few virus infected CD4+ T lymphocytes and had sustained only a minimal loss of this T cell subset . The latter finding suggests that the inability to optimally infect macrophages in the intestinal mucosa may mitigate vigorous virus replication throughout the GALT , thereby significantly altering the typical course of pathogenic lentivirus infections in vivo . At present , it is not clear why HIV-1 neither encodes nor requires an antagonist of SAMHD1 to ensure robust replication in memory CD4+ T cells . Perhaps it is due to the reported high activity of its reverse transcriptase compared to that of SIVmac even when intracellular dNTP concentrations are relatively low [39] . Although the mitigating effects of Vpx may be more apparent in cultured cells of the myeloid lineage , which have very high levels of endogenous SAMHD1 ( Fig 1 ) and low levels of dNTPs , our results indicate that endogenous SAMHD1 can potently restrict SIVmac replication in infected rhesus monkeys unless counteracted by Vpx . This need to degrade SAMHD1 during SIV infections of memory CD4+ T cells in vivo selects for Vpx revertant viruses capable of generating high levels of plasma viremia and inducing immunodeficiency .
This study was carried out in strict accordance with the recommendations of the Public Health Services ( PHS ) Policy of Humane Care and Use of Laboratory Animals . Rhesus macaques ( Macaca mulatta ) were housed in a biosafety level 2 NIAID facility and conducted in accordance with protocols LMM32 , approved by the Institutional Animal Care and Use Committees of NIAID/NIH . Appropriate sedatives , anesthetics and analgesics were used during handling and surgical manipulations to ensure minimal pain , suffering , and distress to animals . Furthermore , housing , feeding and environmental enrichment were in accord with recommendations of the Weatherall report . Animals were euthanized in accordance with the recommendations of the panel on Euthanasia of the American Veterinary Medical Association ( AVMA ) Guidelines for the Euthanasia of Animals ( Section 2 . 3 ) . The macaques used in this study were negative for the MHC class I Mamu-A*01 allele . Phlebotomies , euthanasia and sample collection were performed as previously described [40] . Viral RNA levels in plasma were determined by real-time RT-PCR ( ABI Prism 7900HT sequence detection system; Applied Biosystems ) as previously reported [40] . EDTA-treated blood samples were stained for flow cytometric analysis as described previously [12 , 41] , using combinations of the following fluorochrome-conjugated MAbs: CD3 ( fluorescein isothiocyanate [FITC] or phycoerythrin [PE] ) , CD4 ( PE , peridinin chlorophyll protein-Cy5 . 5 [PerCP-Cy5 . 5] , or allophycocyanin [APC] ) , CD8 ( PerCP or APC ) , CD28 ( FITC or PE ) , and CD95 ( APC ) . All antibodies were obtained from BD Biosciences ( San Diego , CA ) , and samples were analyzed by four-color flow cytometry ( FACSCalibur; BD Biosciences Immunocytometry Systems ) . Data analysis was performed using CellQuest Pro ( BD Biosciences ) and FlowJo ( TreeStar , Inc . , San Carlos , CA ) . In this study , naïve CD4+ T cells were identified by their CD95low CD28high phenotype , whereas memory CD4+ T cells were CD95high CD28high or CD95high CD28low in the CD4+ small lymphocyte gate [41] . Viral RNA was purified from macaque plasma employing the QIAamp Viral RNA Mini kit ( QIAGEN ) and immediately converted to cDNA using SuperScript III reverse transcriptase ( LifeTechnologies ) and a random primer . The newly synthesized single-stranded cDNA was serially diluted and SGA PCR amplification [42] was performed using Platinum Taq High Fidelity polymerase ( LifeTechnologies ) . First round PCR was performed using primers ( forward primer: GAAGGGGAGGAATAGGGGATATGAC and reverse primer: CAAAACTGGCAATGGTAGCAACAC ) with the following parameters: 1 cycle of 94 C for 2 min , 35 cycles of a denaturing step of 94°C for 15 s , an annealing step of 55°C for 30 s , and an extension step of 68°C for 2 min , followed by a final extension of 68 C for 7 min . Second round PCR was performed using primers ( forward primer: CCACTACAGGAAGGAAGCCATTTAG and reverse primer: GCTCCCTCAAGGGTGTCTCCATGTCTATTATA ) with the following PCR parameters: 1 cycle of 94 C for 2 min , 45 cycles of a denaturing step of 94°C for 15 s , an annealing step of 55°C for 30 s , and an extension step of 68°C for 2 min , followed by a final extension of 68°C for 7 min . A human embryonic kidney cell line , 293T ( HEK-293T , CRL-11268 , ATCC , Manassas , VA ) , was cultured in Dulbecco’s modified minimal essential medium supplemented with 10% heat-inactivated FBS . PM1 , and SupT1-R5 cells were cultured in RPMI-1640 supplemented with 10% heat-inactivated FBS . Rhesus monkey PBMCs were prepared , CD8+ T cell depleted , and cultured as described previously [43 , 44] . Rhesus CD4+ T lymphocytes were purified by negative selection using a MACS kit ( [Miltenyibiotec] for non-human primates ) . Macaque CD14+ cells were purified by CD14 microBeads ( Miltenyibiotec ) employing a MACS cell isolation system . MDMs were induced from CD14+ cells by culturing with 100ng/ml of M-CSF ( Peprotech ) for 1 week . Adherent alveolar macrophages were isolated from bronchoalveolar lavage ( BAL ) samples . Fresh whole blood was diluted 1:1 with phosphate buffered saline ( PBS ) and layered over Ficoll-Paque Plus ( GE Healthcare ) and then centrifuged at 2000 rpm for 25 minutes . The PMBC were then removed and washed twice with media , RPMI-1640 supplemented with 10% heat inactivated fetal bovine serum , penicillin , streptomycin and L-glutamine . Whole spleens were digested into single cell suspensions by grinding tissue through a 0 . 22 μm cell strainer followed by lysis of red blood cells with ACK lysis buffer ( Life technologies ) and two washes with complete RPMI media . Splenocytes and PBMCs were stained with the live dead exclusion dye Aqua blue ( Invitrogen ) then with the following mAbs: αCD3-Alexa700 ( clone SP34-2 , BD Pharmingen ) , αCD8-Pacific Blue ( clone RPA-T8 , BD Pharmingen ) , αCD4-PECy5 . 5 ( clone OKT4 , eBioscience ) , αCD28-ECD ( clone 28 . 2 , Beckman Coulter ) , and αCD95-PECy5 ( clone DX2 , BD Pharmingen ) . Memory CD4 T cells were sorted as live , single , lymphocytes expressing CD3 , CD4 , CD95 without expression of CD8 using a modified FACSAria ( BD Immunocytometry Systems ) . Compensation was performed electronically using capture beads stained singly with the individual mAbs . Virus stocks were prepared by transfecting 293T cells with SIV molecular clones using Lipofectamine 2000 ( LifeTechnologies ) ; culture supernatants were collected 48 h later and stored at −80°C until use . Infectious virus titers were determined by end-point dilution using SupT1-R5 cells . Virion-associated 32P-RT activity was measured as described previously [45] . SupT1-R5 , and ConA-stimulated rhesus PBMCs ( 2 × 106 cells in 500 μl ) were infected with transfected cell supernatants of indicated viruses ( normalized by particle-associated 32P-RT activity ) by spinoculation [46] for 1 h , and maintained for 12 days . Tissue culture medium was replaced daily and monitored for RT activity . VSV-G [47] pseudotyped SIVs ( WT SIVmac239 , SIVmac239 X-del , SIVmac239 X-Q76A ) were prepared by transfecting 293-T cells ( VSV:SIV plasmid ratio of 1:5 ) as described earlier . Sorted rhesus memory CD4+ T cells , maintained in RPMI-1640 supplemented with 10% heat-inactivated FBS and 20U/ml of IL-2 , were spinoculated ( 1 ml ) with pseudovirions present in 293-T cell transfection supernatants . Infected CD4+ memory T cells , collected at 24 h post infection , were analyzed for levels of SAMHD1 by immunoblotting . Vpx deletion and point mutants were constructed by PCR mutagenesis of SIVmac239 and SIVmac316 genomes using the following primer pairs: SIVmac239 and SIVmac316 X-del , forward primer: GATCCCAGGGAGAGAATCCCACCTGGAAACAG and reverse primer: TTACATCGCTTACTACTTTCAGTGCTAAGTACTGTAGGCTTGG; SIVmac239 and SIVmac316 X-Q76A , forward primer: AGGCTTTATTTATGCATTGCAAGAAAGGCTGTAGATGTCTAGGGGAAG and reverse primer: TTGCTATTAAACACAAGTATCTGTATTTTACATAGCTTGGTGACATCCC SIVmac316 X-Q76S , forward primer: TCAAAGGCTTTATTTATGCATTGCAAGAAAG and reverse primer: TATTAAACACAAGTATCTGTATTTTACATAGCTTGG . The X-Q76A/I32T double mutant was constructed by PCR mutagenesis of SIVmac316 X-Q76A using forward primer: CAAACAGAGAGGCGGTAAACCACCTAC and reverse primer: TCTCCTCTACTGTTCTGTTTAGCCATTCG . PCR amplifications were performed using Platinum PFX DNA polymerase ( LifeTechnologies ) , with the following PCR parameters: 1 cycle of 94 C for 2 min , 32 cycles of a denaturing step of 94°C for 15 s , an annealing step of 58°C for 30 s , and an extension step of 68°C for 20 min , followed by a final extension of 68 C for 7 min . Following amplification , the PCR product was gel purified , treated with T4 polynucleotide kinase ( LifeTechnologies ) , and then blunt-end ligated to created circular full-length infectious clones . Cells ( 2 x 106 ) were washed with phosphate-buffered saline ( PBS ) and lysed in 100μl of SDS sample buffer ( LifeTechnologies ) . Whole-cell extracts ( 15μl [from 3 x 105 cells] ) were separated on 10% acrylamide gels ( LifeTechnologies ) or Phos-Tag gel ( Wako Chemicals ) . Following electrophoresis , the gel was transferred to a PVDF membrane using an iBlot Gel Transfer system ( LifeTechnologies ) and stained using anti-SAMHD1 ( Proteintech ) , anti-tubulin ( Sigma-Aldrich ) , and anti-GAPDH ( Santa Cruz Biotechnology ) polyclonal antibodies .
|
Primate lentiviruses , such as HIV and its SIV simian relative , encode accessory proteins that suppress cellular restriction factors interfering with efficient replication . One of these , designated Vpx , is produced in infected cells by HIV-2 and some SIV strains , which cause endemic infections in African monkeys . The primary function of Vpx has long been thought to facilitate infectivity in dendritic cells and macrophage by degrading the Sterile Alpha Motif and HD domain-containing protein 1 ( SAMHD1 ) , which restricts virus replication in these cells . Using SIVmac carrying a mutated Vpx gene with a single amino acid change that prevents it from binding to DCAF1 and subsequently mediating the degradation of SAMHD1 , we show that virus infection of CD4+ T lymphocytes is markedly compromised both in vitro and in vivo . The SIV Vpx mutant is severely attenuated in establishing new infections in inoculated rhesus monkeys , in producing high levels of virus progeny , in degrading SAMHD1 in memory CD4+ T cell in infected animals , and in inducing symptomatic disease . Thus , although once considered to be only critical for optimal replication in macrophage based on earlier studies performed with cultured cells , the SIV Vpx protein is functionally important in vivo for establishing the primary infection in rhesus macaques , sustaining high levels of virus replication in CD4+ T lymphocytes , and promoting the onset of symptomatic immunodeficiency .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
The Expression of Functional Vpx during Pathogenic SIVmac Infections of Rhesus Macaques Suppresses SAMHD1 in CD4+ Memory T Cells
|
Fatty liver disease ( FLD ) is characterized by lipid accumulation in hepatocytes and is accompanied by secretory pathway dysfunction , resulting in induction of the unfolded protein response ( UPR ) . Activating transcription factor 6 ( ATF6 ) , one of three main UPR sensors , functions to both promote FLD during acute stress and reduce FLD during chronic stress . There is little mechanistic understanding of how ATF6 , or any other UPR factor , regulates hepatic lipid metabolism to cause disease . We addressed this using zebrafish genetics and biochemical analyses and demonstrate that Atf6 is necessary and sufficient for FLD . atf6 transcription is significantly upregulated in the liver of zebrafish with alcoholic FLD and morpholino-mediated atf6 depletion significantly reduced steatosis incidence caused by alcohol . Moreover , overexpression of active , nuclear Atf6 ( nAtf6 ) in hepatocytes caused FLD in the absence of stress . mRNA-Seq and qPCR analyses of livers from five day old nAtf6 transgenic larvae revealed upregulation of genes promoting glyceroneogenesis and fatty acid elongation , including fatty acid synthase ( fasn ) , and nAtf6 overexpression in both zebrafish larvae and human hepatoma cells increased the incorporation of 14C-acetate into lipids . Srebp transcription factors are key regulators of lipogenic enzymes , but reducing Srebp activation by scap morpholino injection neither prevented FLD in nAtf6 transgenics nor synergized with atf6 knockdown to reduce alcohol-induced FLD . In contrast , fasn morpholino injection reduced FLD in nAtf6 transgenic larvae and synergistically interacted with atf6 to reduce alcoholic FLD . Thus , our data demonstrate that Atf6 is required for alcoholic FLD and epistatically interacts with fasn to cause this disease , suggesting triglyceride biogenesis as the mechanism of UPR induced FLD .
The unfolded protein response ( UPR ) acts in most cells to maintain homeostasis within the protein secretory pathway during physiological conditions . During stress , the UPR becomes further induced to mitigate the accumulation of misfolded or unfolded proteins in the endoplasmic reticulum ( ER ) . If UPR activation cannot relieve the excess protein load , the ER becomes dilated and dysfunctional , and all branches of the UPR remain chronically activated in a condition referred to as ER stress . Many studies have implicated ER stress in a range of pathologies , and there is a clear association between UPR activation and metabolic diseases such as fatty liver disease ( FLD ) [1]–[5] . However , it is not known whether factors that control the UPR can also directly impact lipid metabolism and it remains unclear how UPR activation causes FLD . FLD is the most common hepatic pathology worldwide [6] , and alcohol abuse is a leading cause . Even a single episode of binge drinking causes lipid accumulation in hepatocytes ( steatosis ) in over 90% of drinkers [7] . While acute steatosis can resolve , chronic steatosis can render hepatocytes susceptible to damage and is a prerequisite step in developing more severe liver disease , including steatohepatitis and cirrhosis . Most FLD etiologies are accompanied by impairment of protein secretion by hepatocytes that result in serum protein deficiencies , which are most apparent in chronic alcoholics and contribute to the systemic complications of alcoholic liver disease ( ALD ) . These defects are reflected in many reports demonstrating that alcohol induces some UPR sensors and targets in the liver of mice [1] , [8] , rats [9] , micropigs [10] , and zebrafish [11]–[13] . Moreover , deleting ATF4 or CHOP , two key UPR effectors , reduces ethanol-induced liver injury [8] , [14] , indicating that UPR activation can contribute to ALD . However , it is not clear whether these genes , or other UPR effectors , contribute to steatosis . There are three main UPR sensors that function through shared and independent mechanisms to maintain ER function by enhancing protein processing and folding in the ER , and by promoting degradation of terminally misfolded secretory proteins . Activating transcription factor 6 ( ATF6 ) and inositol-requiring enzyme 1-alpha ( IRE1α or ERN1 ) pathways generate the active transcription factors nATF6 and XBP1 , respectively , that induce hundreds of UPR target genes [15] , [16] . PRKR-like endoplasmic reticulum kinase ( PERK or EIF2AK3 ) phosphorylates EIF2A to repress translation [17] and to promote production of ATF4 [18] , which induces a subset of target genes [14] , [18] . Each of these primary sensors has been evaluated for its contribution to FLD caused by a robust ER stressor [19] , [20] . ATF6 knockout mice fail to resolve steatosis caused by acute stress due to tunicamycin injection [16] , [21] , [22] , suggesting that loss of ATF6 promotes fatty liver . Our previous work using zebrafish confirmed the finding that steatosis caused by acute stress is augmented by atf6 loss , and also demonstrates that steatosis caused by chronic stress is reduced when Atf6 is depleted [23] . Thus , it appears that atf6 loss can alternatively enhance or reduce FLD , depending on the nature and duration of the stress . The relationship between metabolic disease and UPR activation is under intensive investigation , yet several important questions remain unanswered . First , is UPR activation a cause or consequence of FLD ? While some studies indicate that lipotoxicity and fatty acid accumulation can induce the UPR [24] , [25] , there is incontrovertible evidence that robust UPR activation is sufficient to induce steatosis [2] , [26] , [27] . These data are incorporated into a current model proposing that robust UPR activation can cause steatosis and , if the lipid burden is not resolved , this can further augment cellular stress and contribute to chronic UPR induction . The second question is which , if any , of the main UPR factors directly cause fatty liver ? All studies to date that have addressed this question , including our work in zebrafish [23] , utilize loss of function approaches whereby a key UPR gene is deleted and the effect on stress-induced steatosis is evaluated . However , these studies are complicated by the extensive crosstalk between UPR effectors , and by the ability of cells to adapt to changes in UPR capacity . Third , what is the mechanism by which UPR activation causes fatty liver ? Activation of lipogenic genes by the sterol response element binding protein ( SREBP ) transcription factors has been implicated as an essential pathway in both alcoholic [28] , [29] and non-alcoholic [30]–[32] FLD . Since SREBPs and ATF6 are activated by a shared proteolytic mechanism [33] , it is possible that these two proteins are activated simultaneously in response to stress . However , the role of SREBPs in FLD caused by UPR activation has not been addressed . Here , we use genetic and genomic approaches in a well-established zebrafish model of ALD [11]–[13] , [34]–[37] to demonstrate that blocking Atf6 prevents alcohol-induced steatosis and that overexpression of the nuclear , active form of Atf6 ( nAtf6 ) causes steatosis . We found that nAtf6 overexpression increases expression of genes in the lipogenic pathway and alters lipid flux to promote triglyceride synthesis by a mechanism that does not involve Srebps . Instead , we demonstrate an epistatic interaction between atf6 and fasn in driving alcoholic steatosis . These findings demonstrate the first clear and causative link between a central UPR sensor and FLD .
We previously reported that transcription of multiple UPR effector genes and translation of the major ER chaperone , Bip , are induced in zebrafish livers following alcohol exposure [11]–[13] . We have also demonstrated that xbp1 splicing , a direct measure of Ire1a activation , is only detectable within the first few hours of exposure to ethanol [13] yet UPR target gene induction persists long afterwards [12] , suggesting that the Ire1a pathway is not entirely responsible for the UPR in ALD . Our previous finding that Eif2a is phosphorylated in the liver of zebrafish with ALD [13] suggests that Perk is activated by alcohol , but since Eif2a can also be phosphorylated by other kinases [38] , this has yet to be resolved . The lack of reliable antibodies that recognize the active form of Atf6 , Perk and Ire1a in zebrafish hampered direct measurement of the activation status of each pathway ( not shown ) . However , quantitative real-time PCR ( qPCR ) analysis of liver samples from 5 day post fertilization ( dpf ) larvae treated with alcohol demonstrated that atf6 mRNA was significantly induced in the liver ( Figure 1A ) prior to the onset of steatosis ( see Figure 1B and [13] ) . Since Atf6 has been shown to act in a positive feedback loop to induce its own expression [39] , these data suggest that the increase in atf6 expression could reflect Atf6 activation in response to alcohol . To test whether Atf6 was required for alcoholic steatosis , we injected a morpholino ( MO ) targeting either the translation initiation site ( atf6MO-ATG ) [23] or the boundary between coding exon 1 and intron 1 of atf6 ( atf6MO-SPL ) ( Table S1 and Figure S1A ) , which blocked splicing ( Figure S1B ) and introduced an early stop codon ( Figure S1C ) . Morphants and uninjected controls were treated with 350 mM ethanol for 32 hours , stained with the neutral lipid dye , oil red O and scored for the presence or absence of steatosis ( Figure 1C and S2A ) . Consistent with previous findings [11]–[13] , [23] , [34] , the percent of larvae with steatosis ( i . e . steatosis incidence ) was significantly higher in uninjected larvae exposed to 350 mM ethanol for 32 hours ( 13% in untreated larvae vs . 53% in ethanol treated larvae; p<0 . 05 ) . atf6 morphants had reduced steatosis incidence ( p<0 . 05 , Figure 1C-D ) , but still displayed other gross morphological abnormalities caused by ethanol including increased liver circularity , an indication of hepatomegaly ( Figure S2B-C ) , a common feature of liver disease . Thus , Atf6 is required for alcoholic steatosis . To determine whether Atf6 is sufficient to cause FLD , we created a transgenic zebrafish line expressing the predicted nuclear , active fragment of zebrafish Atf6 ( amino acids 1-366; Figure S3A ) fused to mCherry under the hepatocyte-specific promoter fabp10 , along with a cassette that expressed GFP in cardiomyocytes as a marker of transgenesis ( ( Tg ( fabp10:nAtf6-cherry; cmlc2:GFP ) hereafter called nAtf6 TG ) . There is high sequence identity between human and zebrafish Atf6 proteins in the transmembrane , leucine zipper and basic ( bZIP ) domains and the protease recognition sites are highly conserved ( Figure S3A ) . To confirm its nuclear localization , the predicted nAtf6 fragment of the zebrafish protein was fused to GFP and transfected into HepG2 cells ( Figure S3B ) . Transgenics were selected based on GFP expression in the heart , appeared grossly normal throughout development ( Figure S3C ) , and developed into viable , fertile adults . nAtf6-mCherry was not detectable using low-resolution fluorescence microscopy in larvae from all four of the transgenic lines we generated . However , we confirmed transgene expression by PCR and blotting for mCherry ( Figures S3D-E ) . atf6 mRNA expression in the liver was ∼4 fold higher in nAtf6 TG larvae compared to wildtype ( WT ) larvae ( Figure 2A and S4 ) at 5 dpf , and was persistently elevated in older fish ( Figure S4D and not shown ) . We detected mCherry mRNA in 5 dpf nAtf6 TG larvae by PCR , and Western blotting for mCherry detected the transgene in the liver of adult fish , albeit at much lower levels than achieved in transgenic fish expressing nuclear-localized mCherry without fusion to nAtf6 ( i . e . Tg ( fabp10:nls-mCherry ) ; [40] Figures S3D-E ) . We confirmed that nAtf6 overexpression did not induce the other UPR branches by assessing xbp1 splicing ( Figure S4A ) and Eif2a phosphorylation ( Figure S4B ) in the liver of nAtf6 TG 5 dpf larvae . We found no difference in these markers in unstressed nAtf6 TG and WT larvae , however , xbp1 splicing was lower in transgenics than in WTs after exposure to stress ( i . e . 1 µg/ml tunicamycin for 48 hours; Figure S4A ) . This suggests that nAtf6 overexpression adapts hepatocytes to withstand a robust , pharmacologically-induced ER stress . We next examined the expression of UPR target genes in the liver of 5 dpf transgenic and control larvae . We found that bip ( also called hspa5 ) , a transcriptional target of Atf6 , was elevated at the protein ( Figure S4C ) and mRNA levels ( Figure 2A and S4D ) in nAtf6 TG larvae . mRNA-Seq analysis comparing control larvae ( Tg ( fabp10:nls-mCherry ) ; n = 2 pools of ∼40–50 livers each; see [40] and GEO dataset GSE52605 ) , ethanol-treated ( n = 2 pools of ∼40–50 livers each ) , and nAtf6 TG larvae ( n = 1 pool of ∼40–50 livers ) revealed 33 UPR target genes to be highly and significantly upregulated in nAtf6 TG livers ( Tables S2-S4 and Figure 2A; GEO dataset GSE56498 ) . These genes were identified as ATF6 target genes in other studies [16] , [41] or by having the UPR or protein folding as a primary GO descriptor , and included ER chaperones ( bip/hspa5 , dnajc3 , and grp94/hsp90b1 ) , quality control effectors ( canx , calret , calrl , and calrl2 ) , protein disulfide isomerases ( pdia3 , pdia4 ) and ERAD components ( edem1 , derl1 and derl3 ) . A subset of these genes was confirmed by qPCR at 5 dpf and at 14 dpf ( Figure 2A and S4D-E ) . atf4 and derl1 were the only genes detected as significantly upregulated by qPCR but not by mRNA-Seq . Analysis of the mRNA-Seq data from ethanol-treated larvae revealed a striking overlap between the “UPR-ome” induced by ethanol and by nAtf6 overexpression: of the 49 UPR target genes induced by ethanol , 30 were also induced in the liver of nAtf6 TG larvae ( Figure 2B and Table S2 ) . It is possible that the remaining 19 genes unique to the ethanol-induced transcriptome are attributed to the transient , early increase in xbp1 splicing [13] or to other transcription factors that regulate their expression . Collectively , these data indicate that Atf6 is a major regulator of the UPR in ALD . We next asked whether nAtf6 was sufficient to cause FLD ( Figure 2C-E ) . Induction of the fabp10:nAtf6-cherry transgene occurs at ∼2 . 5 dpf [42] and by 4 dpf , 44% of nAtf6 TG larvae developed steatosis which increased to 69% and 76% by 5 and 5 . 5 dpf , respectively ( Figure 2C-D and S5A ) and persisted until at least 14 dpf ( Figure S5A ) , but there were no marked histological abnormalities in the nAtf6 TG liver at 5 dpf ( Figure S5B ) . Hepatic triglyceride ( TAG ) levels were nearly doubled at 5 dpf , and six times higher in 14 dpf nAtf6 TG larvae ( Figure 2E ) , clutch-to-clutch variability non-withstanding . However , TAG levels were the same in WT and nAtf6 TG adults ( Figure 2E ) , despite persistent expression of the transgene ( Figure S3E and not shown ) . We speculate that , over time , animals adapt to nAtf6 overexpression or that other metabolic parameters affect hepatic lipid accumulation in adult fish . Together , our data demonstrate that Atf6 is necessary for alcoholic steatosis and sufficient to cause steatosis in the absence of any other stress . Steatosis is caused by elevated TAGs in hepatocytes resulting from increased lipid synthesis or uptake , or from decreased lipid utilization or export . We therefore analyzed pathways relevant to these processes in the mRNA-Seq dataset from nAtf6 TG livers to determine if any were dysregulated . TAGs are generated from linking free fatty acids ( FFAs ) to a glycerol backbone , which is generated by conversion of dihydroxyacetone phosphate ( DHAP ) to glycerol-3-phosphate ( G3P ) by G3P dehydrogenase ( GPD1 ) ( see schematic in Figure 3A ) . We found genes that participate in TAG synthesis ( Figure 3A-B ) or export ( Figure S6 ) were dysregulated in nAtf6 TG livers at 5 dpf , albeit not to the same degree as observed for the UPR target genes ( Figure 2 ) . DHAP is generated via glycolysis or by glyceroneogenesis , a pathway that has significant overlap with gluconeogenesis ( see [43] and schematic in Figure 3A ) . mRNA-Seq analysis demonstrated that several genes involved in glycolysis were downregulated in nAtf6 TG livers including aldob , glut2/slc2a2 , gck , pgk1 , and pklr ( Figure 3A-B ) . Genes promoting glyceroneogenesis , particularly pck1 and got1 , which generates oxaloacetate for conversion to phosphoenoylpyruvate by pck1 , were upregulated , as was gpd1b , which drives glycerol formation . This suggests that nAtf6 overexpression increases factors regulating glyceroneogenesis preferentially over those that promote glycolysis ( Figure 3A-B ) . In addition , ech1 and tecrb , two genes responsible for fatty acid elongation in the mitochondria and ER , respectively , were induced in nAtf6 TG livers by mRNA-Seq ( Figure 3A-B ) . Some of these genes were confirmed by qPCR analysis of nAtf6 TG livers , with the median fold change from qPCR on 3 to 6 clutches of livers from 5 and 14 dpf larvae depicted in Figure S7 . TAGs are assembled in the hepatocyte ER and secreted in association with lipoproteins that are required for transport in plasma . Several genes that regulate lipoprotein assembly and export , including apom , apoeb , apoea , cetp and zgc:162608 and zgc:194131 , two apolipoprotein-like genes , were downregulated in nAtf6 TG livers ( Figure S7 and Table S3 ) . pdia2 , which interacts with MTP to transport lipids out of hepatocytes [44] , was also downregulated . These data suggest that hepatic lipid transport may be impaired by nAtf6 overexpression . While these expression data do not capture the complex metabolic and post-translational regulation of the proteins encoded by these genes , they do suggest enhanced lipogenesis or suppressed lipoprotein export as potential mechanisms of steatosis in nAtf6 transgenics . Acetate is a precursor to acetyl-CoA , the major building block of fatty acid and , therefore , TAG synthesis . To determine if nAtf6 overexpression increased de novo lipogenesis , we measured the incorporation of 14C-acetate into lipids in nAtf6 TG larvae ( Figure 4A ) and in HepG2 cells overexpressing nATF6 ( Figure 4B-D ) compared to their respective controls . Larvae were incubated with 14C-acetate from 3–5 dpf and we measured the amount of radiolabel present in the lipid fraction compared to that present in the lysate of the whole larvae . In an average of 5 individual pools of larvae , more 14C radiolabel was detected in the lipid fraction of nAtf6 TG larvae than in wildtype larvae ( Figure 4A ) . Since the liver in 5 dpf zebrafish larvae is too small to analyze incorporation in this organ exclusively , analysis of the whole larvae could not determine whether the label was preferentially incorporated into hepatocytes of nAtf6 TG larvae . We thus used human hepatoma ( HepG2 ) cells transfected with human nATF6 to address this question . nATF6 overexpression causes accumulation of oil red O stained cytoplasmic droplets ( Figure 4B ) that are both greater in number and larger in size than in cells transfected with GFP ( Figure 4B ) . These findings were confirmed in 293T cells ( Figure S8 ) and correlate with our findings in nAtf6 TG larvae ( Figure 2 ) . Thus , nAtf6 overexpression causes lipid accumulation across cell types and species . Moreover , in eleven independent cell samples from two separate experiments , HepG2 cells overexpressing nATF6 had slightly more 14C radiolabel in the lipid fraction compared to GFP-transfected cells ( Figure 4D ) . While we found that the incorporation tended to be higher in the nATF6 expressing cells , the average increase was moderate , at best . We attribute this to the transfection efficiency of HepG2 cells combined with experimental conditions used to label lipids , which likely act to select against nATF6 expressing cells . Fatty acid synthase ( fasn ) mediates multiple steps in fatty acid synthesis that culminate in generation of palmitate from acetyl-CoA , a building block for TAG biosynthesis . fasn transcription in hepatocytes is primarily regulated by Srebp1c [45]–[47] , and we previously demonstrated that fasn transcripts and other Srebp target genes are induced in zebrafish with ALD [12] , [13] . Moreover , blocking Srebp activation by injecting a morpholino targeting the Srebp activating protein , Scap , reduced the incidence of alcoholic steatosis [12] , [13] . Since Atf6 and Srebps are activated by similar mechanisms – they are both retained as inactive precursors in the ER and processed by the same Golgi-resident enzymes [33] – Srebp mediated lipogenesis has been proposed as a mechanism by which UPR activation causes FLD [27] , [32] . We tested this using a genetic approach . First , we asked whether blocking Atf6 affected the induction of lipogenic gene expression in the liver of ethanol treated larvae . Both fasn ( Figure 5A ) , and acc1 ( acaca; Figure S9A ) were upregulated by ∼2 fold in the liver of ethanol treated larvae , and this was blocked by atf6 MO injection ( Figures 5A and S9A ) . However , induction of other Srebp target genes was not blocked by Atf6 depletion , including srebp1 ( Figure S9A ) and the Srebp2 targets hmgcra , hmgcs1 and mvk ( Figure S9B ) . Since the entire Srebp target gene set was not uniformly affected by atf6 loss , we conclude that Atf6 does not interact with Srebps . Instead , Atf6 may affect expression of specific lipogenic genes ( i . e . fasn and acc1 ) by a mechanism distinct from Srebp1 . Interestingly , we found mild induction of Srebp2 target genes in untreated atf6 morphants ( Figure S9B ) , supporting conclusions based on experiments in mammalian cultured cells where ATF6 was found to directly suppress SREBP2 activity [48] . Next , we asked whether nAtf6 was sufficient to induce fasn and acc1 by assessing their expression in livers of nAtf6 TG larvae . Both fasn ( p<0 . 05; Figure 5B ) and acc1 ( Figure S9C ) were increased by ∼2 fold at 5 dpf , and fasn expression was also higher at 14 dpf in nAtf6 TG larvae and in HepG2 cells overexpressing nATF6 ( Figure 5B ) . However , other Srebp1 and Srebp2 genes were not significantly induced by nAtf6 overexpression in nAtf6 TG larvae ( Figure S9C-D ) or in HepG2 cells overexpressing human nATF6 ( not shown ) , indicating that nAtf6 does not activate the entire Srebp target gene program . We then tested whether nAtf6 functioned upstream of Srebps by assessing whether scap morpholino injection suppressed steatosis in nAtf6 TG larvae . While scap morphants are resistant to ethanol-induced steatosis [12] we found no difference in steatosis incidence in nAtf6 transgenics ( Figure 5B ) . Finally , to test whether atf6 and scap interacted to cause alcoholic steatosis we injected low concentrations of morpholinos targeting both genes and found that they did not synergize to suppress alcoholic steatosis ( Figure S10 ) . We thus conclude that lipogenic gene induction and steatosis caused by nAtf6 overexpression does not require Srebps . The functional relevance of the finding that nAtf6 overexpression increases lipogenesis was tested by assessing the effect of morpholino-mediated fasn ( Table S1 ) knockdown on steatosis incidence in nAtf6 TG larvae . Injection of high concentrations of fasn MO induced severe morphological defects , so we optimized the fasn MO concentration to have minimal toxicity . This had no effect on steatosis incidence in control larvae , but reduced steatosis incidence in nAtf6 TG larvae from 74% to 42% ( Figure 5C ) . Furthermore , we found that atf6 and fasn interacted epistatically to cause alcoholic steatosis by co-injecting concentrations of fasn and atf6-SPL MOs that did not have any effect on alcoholic steatosis on their own but , together , significantly reduced steatosis incidence in response to ethanol ( Figure 5D ) . Thus , Atf6 and Fasn function in the same genetic pathway . Taken together , our data suggest a model by which Atf6 causes steatosis , in part , by inducing fasn and TAG synthesis by a mechanism that is independent of Srebps ( Figure 6 ) .
There is a wealth of recent literature showing that the UPR is activated in FLD and that this aspect of the disease is conserved across species [4] , [16] , [21] , [23] , [49]–[51] . However , it has not been shown that any single UPR factor can directly cause this disease and the mechanism by which UPR activation causes lipid accumulation is unclear . Here , we use zebrafish genetics and biochemical analyses to show that Atf6 is both necessary and sufficient for steatosis by showing that: ( i ) Atf6 is required for alcoholic steatosis , ( ii ) activation of Atf6 is sufficient to cause steatosis , ( iii ) Atf6 activation induces expression of genes involved in glyceroneogenesis and fatty acid elongation and causes de novo lipogenesis , and ( iv ) Atf6 epistatically interacts with fatty acid synthase ( fasn ) , a key enzyme involved in TAG biosynthesis , to cause FLD . This is among the first data to directly link a main UPR sensor and lipid metabolic pathways . A number of studies have demonstrated that loss of one of the key UPR sensors – PERK [19] , IRE1α [20] or ATF6 [16] , [21] , [23] – enhances steatosis caused by acute stress . Additionally , we previously demonstrated that Atf6 depletion suppresses steatosis caused by chronic stress [23] . Thus , it is clear that UPR activation is a conserved , common feature of FLD , but whether it functions to promote or prevent steatosis appears to depend on the nature and duration of the FLD-causing stress . Importantly , these loss of function studies have described a requirement for the UPR in FLD caused by tunicamycin injection , but not in the context of a stress that mirrors human conditions , such as obesity or alcohol abuse . Here , we show that Atf6 is induced in ALD , that its activation precedes steatosis , and that knocking down Atf6 reduces alcoholic steatosis . Thus , Atf6 is required for ALD . Moreover , since overexpression of nAtf6 is sufficient to cause FLD in the absence of any other stress , we conclude that this branch of the UPR is a pathophysiological mechanism of FLD . Atf6 and Xbp1 are the primary transcription factors that both independently and cooperatively regulate expression of hundreds of UPR target genes . We previously found significant upregulation of many UPR target genes in zebrafish with ALD [11] , [13] and hypothesized that since xbp1 splicing in the liver was only an early , transient response to alcohol [13] , other transcription factors must participate in this robust UPR . Transcriptome analysis of the livers from nAtf6 TG and ethanol treated larvae enabled us to identify a set of target genes that are likely to be directly activated by Atf6 . Many of these genes were also reported to be upregulated in mice overexpressing inducible ATF6 in the heart [41] and they were not induced by tunicamycin treatment of Atf6-/- MEFs [16] . Genes occupying the intersection of these different datasets are likely bona fide Atf6 targets and the transgenic larvae we generated provides a system to test this directly . In contrast , many genes that were downregulated when ATF6 was overexpressed in the mouse heart were unchanged in nAtf6 transgenic zebrafish livers and , conversely , genes such as canx , derl1 , and atf4 were upregulated in the nAtf6 TG zebrafish liver samples but not in the mouse model . The variations between the two datasets may be due to inherent differences in the models or approaches used to detect changes in gene expression , or may reflect the ability of Atf6 to differentially regulate target genes depending on cell type: hepatocytes possess a significantly higher secretory capacity than cardiomyocytes , and thus the Atf6 transcriptome in hepatocytes may be more extensive in order to maintain ER homeostasis . Finally , comparative transcriptome analysis between nAtf6 TG and ethanol treated larvae identified significant overlap between the UPR target genes in these two datasets , indicating that Atf6 is the main transcriptional driver of the ethanol-induced “UPR-ome” . What is the mechanism by which the UPR causes FLD ? Our data showing that nAtf6 overexpression induces the expression of some lipogenic enzymes and increases lipid synthesis , and that blocking fatty acid synthesis reduces FLD , fits a model ( Figure 6 ) whereby increased lipid production is , in part , the mechanism of steatosis in FLD caused by UPR activation . However , despite our finding that Atf6 depletion suppresses fasn expression and nAtf6 overexpression induces fasn , we have no evidence that fasn is a direct transcriptional target of Atf6 . Indeed , the level of fasn induction is far less than the established Atf6 target , bip , and there are no canonical UPREs or ERSEs [15] in the fasn promoter ( not shown ) , thus other pathways are also likely at play . The mechanism by which nAtf6 induces fasn and acc1 require further investigation , and the tools we describe here will facilitate such studies . The SREBP transcription factors are well-characterized regulators of hepatic lipogenesis; SREBP1c functions by increasing the expression of the full panel of genes required for TAG biogenesis [52] . One possibility , suggested by the finding that the UPR and SREBPs are activated in parallel in some systems [26] , [27] , [32] , is that nAtf6 causes Srebp activation , leading to increased expression of fasn and other lipogenic genes and promoting lipogenesis . Our data argues against this: ( i ) the full panel of Srebp targets are not induced by nAtf6 overexpression in zebrafish , ( ii ) Atf6 and Srebps did not epistatically interact to modify alcoholic steatosis and ( iii ) Srebp activation is not required for steatosis in nAtf6 TG larvae . While it does appear that Srebp activation is required for alcoholic steatosis [12] , [28] , [29] , the current study shows that it is not the only important pathway that regulates hepatic lipid metabolism , as steatosis in nAtf6 TG larvae is independent of Srebps . While Srebps are not required for the effects of Atf6 on steatosis , there does appear to be some interaction between Atf6 and Srebp2 , consistent with in vitro data showing that ATF6 suppresses SREBP2 [48] . We found that Atf6 causes upregulation of Srebp2 target genes , and while this modest increase in Srebp2 target genes in atf6 morphants did not appear to have a functional impact on cholesterol biogenesis , since atf6 morphants do not develop steatosis but are protected from it , it does suggest that Atf6 may suppress Srebp2 activity . We used mRNA-Seq analyses to identify potential pathways that are dysregulated in response to Atf6 overexpression , although it is clear that the complex post-translational regulation of these pathways are not captured by gene expression studies . Our data suggested that lipoprotein export may be impaired by Atf6 overexpression , but we did not find that the flux of 14C-acetate into the extracellular lipids was impaired in cells overexpressing nATF6 . While there are some technical caveats to these biochemical studies , we speculate that decreased lipoprotein export is not the only mechanism by which nAtf6 overexpression causes steatosis . Future studies to optimize cell conditions to ensure maximal cell viability and sustained high levels of nAtf6 expression will be required to fully address this . While steatosis is clearly a first step on the continuum to more severe liver disease , we also propose a different possibility: that lipid accumulation is a sign of an active and adaptive stress response . Protein folding in the ER is metabolically demanding , and when secretory cargo increases or the ER becomes full of unfolded proteins , increased lipid accumulation may enhance the ATP supply and thereby sustain the increased protein folding demands . Thus , the UPR may serve to enhance the protein folding capacity of the ER function in two ways: by activating genes required for protein folding , quality and export , and by promoting metabolic flux to TAGs as an energy source to withstand the challenge presented by a high level of secretory cargo . In this scenario , acute steatosis would not be a pathological response but , instead , would provide a protective role to sustain hepatocyte function during stress . However , if the stress response persists and steatosis becomes chronic , lipotoxicity could potentiate liver injury . Based on these data , it is tempting to speculate that enhancing protein folding in the ER by chemical chaperones , which have been used in both mouse models and humans to alleviate UPR activity , attenuate fatty liver disease and increase insulin sensitivity [3] , [53] , [54] , could reducr the demand for energy and thus reduce the need to accumulate lipid . Whether Atf6 functions both as sensor of unfolded proteins and of metabolic demand remains to be elucidated .
Adult wildtype ( WT , TAB14 and AB ) , Tg ( fabp10:dsRed ) [55] and Tg ( fabp10:nls-mCherry ) [40] zebrafish were maintained according to standard conditions . Larvae were exposed to 350 mM ethanol ( Pharmco-AAPER , Brookfield , CT ) in fish water starting at 96–98 hours post fertilization ( hpf ) for up to 32 hours as described [11] , [12] . All zebrafish protocols were approved by Mount Sinai's Institutional Animal Care and Use Committee . Morpholinos targeting the translation initiation ATG of atf6 ( atf6MO-ATG ) [23] , an atf6 intron-exon boundary ( atf6MO-SPL , Figure S1 ) , a scap intron-exon boundary [12] , and a fasn intron-exon boundary were ordered from GeneTools ( Philomath , OR ) . Approximately 1–5 pmol were injected into 1–2 cell stage embryos . Morpholino sequences and amount injected per embryo ( ng ) are listed in Table S1 . The Tg ( fabp10:nAtf6-cherry; cmlc2:GFP ) transgenic line was created by injecting a vector containing 2813 bp of the fabp10 promoter [55] upstream of the predicted nuclear fragment of zebrafish Atf6 ( amino acids 1–366 ) as identified via DNA and protein alignments with human ATF6 ( NCBI Reference Sequence: NP_031374 . 2 ) . A cassette driving GFP expression in cardiomyocytes was included for rapid screening of transgenics ( cmlc2:GFP ) . The transgene was flanked by Tol2 sites and the vector was injected into fertilized eggs along with transposase mRNA . Larvae were selected for cmlc2:GFP expression and raised to adulthood , outcrossed to TAB14 adults and 4 germline founders were identified . The nATF6-pcDNA-DEST47 plasmid was created using the Invitrogen Gateway System in which human nATF6 ( amino acids 1–380 ) was amplified from a construct containing the full ORF ( pEGFP-hATF6 , from Dr . Aguirre-Ghiso ) and ligated into pcDNA-DEST47 . The nAtf6-GFP/pCI-Neo plasmid was generated by amplifying zebrafish nAtf6-GFP from a Tol2-generated plasmid ( fabp10:nAtf6-GFP ) with primers containing restriction sites for EcoRI and SmaI . The resulting PCR product was ligated into pCI-Neo via T4 ligase . pEGFP-C1 ( Clontech ) was used as a positive control for transfection . 293T and HepG2 cells were grown in 100 mm dishes ( BD Biosciences ) with DMEM ( Corning Cellgro , Manassas , VA ) containing 10% fetal bovine serum ( Invitrogen ) and penicillin/streptomycin ( Cellgro ) , and housed in a humidified incubator at 37 C with 5% CO2 . Cells were passaged at 80–90% confluency either into 100 mm dishes or 6 well culture plates ( Corning ) . Each well was transfected with 1–2 µg plasmid using Xtreme GENE 9 ( Roche ) or Lipofectamine 2000 ( Invitrogen ) for 24 hours . Following transfection , cells were collected for RNA extraction using TRIzol ( Invitrogen ) , used for oil red O staining , or labeled with 14C-acetate as described below . HepG2 cells were washed with PBS 24–28 hours after transfection and kept in serum free high glucose DMEM ( Corning Cellgro , Manassas , VA ) containing penicillin/streptomycin ( Cellgro ) , 100 nM insulin ( Lilly USA , Indianapolis , IN ) and 0 . 5 µCi 14C-acetate ( PerkinElmer , Waltham , MA ) for 17 hours . Media was harvested and spun to remove dead cells , while cells were scraped , washed twice and lysed in cold PBS . Protein concentration in the cell lysate was measured with a BCA protein quantification kit ( Thermo Scientific , Waltham , MA ) . Fractions of the media and cell lysate portion were used for lipid extraction as described below . WT and nATF6 5 dpf larvae were labeled with 1–5 µCi 14C-acetate for 48 hours . Larvae were washed twice and then sonicated in cold PBS . Protein concentration in the larval lysate was measured with a BCA protein quantification kit ( Thermo Scientific ) . Fractions of the larval lysate were used for lipid extraction as described below . Lipids from media , cells and 5 dpf larvae were extracted according to Bligh and Dyer [56] with the following modifications . Briefly , methanol and chloroform were added to 400 µl media or 50 µl of 100 µl total cell lysate in a ratio of 2∶1 . Samples were vortexed and chloroform and 0 . 45% NaCl were added at a ratio of 1∶1 , vortexed again and phases were separated by centrifugation . The aqueous phase was re-extracted with one part chloroform and combined with the lipid from the first second extraction , washed with methanol and 0 . 45% NaCl at a 1∶1 ratio , centrifuged and the lipid phase was recovered and dried under a stream of N2 gas . Lipids were reconstituted in a 2∶1 mixture of chloroform and methanol ( Thermo Scientific , Waltham , MA ) , dissolved in Ultima Gold and counted using a 2460 MicroBeta2 LumiJET ( PerkinElmer , Waltham , MA ) liquid scintillation counter . CCPM were normalized to protein concentration to account for variations in cell number . Total RNA isolated using TRIzol ( Invitrogen ) from pools of 15–20 dissected livers was reverse-transcribed using qScript cDNA SuperMix ( Quanta Biosciences , Gaithersburg , MD ) . qPCR using A Light Cycler 480 ( Roche ) with PerfeCTa SYBRGreen FastMix ( Quanta Biosciences ) was used as previously described [11] . Values for target gene were normalized to reference gene rpp0 , and dCts calculated using the comparative threshold method ( dCt = 2− ( Ct , gene – Ct , rpp0 ) ) . Primer sequences are listed in Table S1 . Approximately 20 livers from control and nAtf6 transgenic larvae were dissected and lysed in 0 . 5% Triton-X 100 and heated at 65°C for 5 minutes to inactivate hepatic lipases . Triglycerides were measured using the Infinity Triglyceride Liquid Stable Reagent ( Thermo Fisher Scientific , Waltham , MA ) following manufacturer's instructions , and were normalized to total protein concentration as determined by Bradford Assay ( Bio-Rad , Hercules , CA ) . Larvae were fixed in 4% paraformaldehyde ( PFA; Electron Microscopy Sciences , Hatfield , PA ) overnight at 4°C , stained with oil red O , and scored for steatosis as previously described [35] . Cryosections were stained by immersing slides in increasing concentrations of propylene glycol ( 85% , 100% ) for 10 minutes followed by an overnight incubation in oil red O ( 0 . 5% in propylene glycol , Polysciences , Warrington , PA ) . Excess oil red O was removed the next day by sequential washes in 100% and 85% propylene glycol for 5 minutes . Nuclei were counterstained with hematoxylin . HepG2 and 293T cells were stained as previously described [57] and counterstained with hematoxylin . Oil red O droplet area and number were calculated using ImageJ . In brief , images were split into the red , green , and blue channels and the green channel was further processed for quantification as described , as oil red O has an excitation at 510 nm [58] . Following this , the background was subtracted from each image to ensure only counting of oil red O droplets . Droplets equal or greater than five square pixels were counted and area quantified . The values were copied into Microsoft Excel and the average droplet area and number of droplets per nucleus were counted for each field . Twenty independent fields from nATF6 and GFP-transfected HepG2 and 293T cells were imaged at 60x magnification . Total RNA was isolated using TRIzol ( Invitrogen ) from pools of ∼40–50 livers dissected on 5 dpf . Two clutches of Tg ( fabp10:nls-mCherry ) larvae that were either untreated ( control ) or treated with 350 mM ethanol for 24 hours and one clutch of nAtf6 TG larvae were used . polyA-tailed mRNA was selected using oligo-dT beads and then fragmented . cDNAs were created using random-hexamers and ligated with bar-coded adaptors compatible with HiSeq 2000 . Single-end , 100 bp reads were sequenced at the Genomics Core of the Icahn School of Medicine at Mount Sinai . Custom-built software was used to map the reads to the zebrafish genome ( Zv9/DanRer7 ) and estimate the coverage of each gene . Briefly , the reads were split into three 32 bp parts after trimming 2 bp at each end and mapped to the genome using a suffix-array based approach . The median of coverage across the transcript was used as an estimate of gene expression . The expression values were quantile normalized and log-ratios were calculated by comparing nAtf6 TGs to the average of the controls . Each ethanol treated sample was compared to its paired control ( untreated siblings ) . Unique Gene Ontology terms ( GO terms ) were assigned to each gene by ranking the GO terms by relevance to the biology of the response , and using annotations from the human orthologues if the zebrafish annotations were lacking . The distribution of expression values were plotted to identify the peak in the distribution , which is the level of noise in the system . The values were regularized by adding the noise to each gene's expression level before the log-ratios were calculated . This ensures that genes with low expression do not contribute to list of genes with large fold-changes . An absolute natural log ratio of 0 . 2 was used as the cutoff ( known , non-responding genes are all below this threshold ) . The list of genes that show changes above this cutoff were analyzed for pathway enrichment using GO terms annotated as described above . This data is available via the Gene Expression Omnibus ( GEO ) at accession number GSE56498 . The controls for this dataset , untreated 5 dpf Tg ( fabp10:nls-mCherry ) , also serve as controls in GEO dataset GSE52605 [40] and were included in both sets for ease of comparison between genotypes and treatments . Images were cropped and minimally processed using Adobe Photoshop CS4 ( Adobe Systems , San Jose , CA ) . Graphs were plotted using Prism 5 . 0c ( GraphPad Software Inc . , La Jolla , CA ) . Heat maps for mRNA-Seq data were generated using GENE-E ( Broad Institute , Cambridge , MA ) . The global maximum and minimum for each gene set was set to orange and green , respectively , and zero was set to white . Metabolic pathway schematics were adapted from WikiPathways ( URL: www . wikipathways . org ) . Statistical tests were performed using GraphPad QuickCalcs ( GraphPad Software , URL: http://www . graphpad . com/quickcalcs/index . cfm ) . For oil red O staining of whole larvae , chi-square with Fisher's Exact test were performed . For qPCR , radiolabeling , and oil red O staining of HepG2 and 293T cells we performed unpaired and one-sample t-tests as appropriate . Methods for histological analysis , cryosectioning and staining , and liver circularity analysis are listed in Text S1 .
|
Fatty liver disease ( steatosis ) is the most common liver disease worldwide and is commonly caused by obesity , type 2 diabetes , or alcohol abuse . All of these conditions are associated with impaired hepatocyte protein secretion , resulting in hypoproteinemia that contributes to the systemic complications of these diseases . The unfolded protein response ( UPR ) is activated in response to stress in the protein secretory pathway and a wealth of data indicates that UPR activation can contribute to steatosis , but the mechanistic basis for this relationship is poorly understood . We identify activating transcription factor 6 ( Atf6 ) , one of three UPR sensors , as necessary and sufficient for steatosis and show that Atf6 activation can promote lipogenesis , providing a direct connection between the stress response and lipid metabolism . Blocking Atf6 in zebrafish larvae prevents alcohol-induced steatosis and Atf6 overexpression in zebrafish hepatocytes induces genes that drive lipogenesis , increases lipid production and causes steatosis . Fatty acid synthase ( fasn ) is a key lipogenic enzyme and we show that fasn is required for fatty liver in response to both ethanol and Atf6 overexpression . Our findings point to Atf6 as a potential therapeutic target for fatty liver disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"osteichthyes",
"animal",
"models",
"zebrafish",
"medicine",
"and",
"health",
"sciences",
"alcoholic",
"liver",
"disease",
"model",
"organisms",
"fishes",
"gastroenterology",
"and",
"hepatology",
"genetics",
"biology",
"and",
"life",
"sciences",
"vertebrates",
"genetics",
"of",
"disease",
"animals",
"liver",
"diseases",
"organisms",
"research",
"and",
"analysis",
"methods"
] |
2014
|
Activating Transcription Factor 6 Is Necessary and Sufficient for Alcoholic Fatty Liver Disease in Zebrafish
|
A major question in evolutionary biology is how natural selection has shaped patterns of genetic variation across the human genome . Previous work has documented a reduction in genetic diversity in regions of the genome with low recombination rates . However , it is unclear whether other summaries of genetic variation , like allele frequencies , are also correlated with recombination rate and whether these correlations can be explained solely by negative selection against deleterious mutations or whether positive selection acting on favorable alleles is also required . Here we attempt to address these questions by analyzing three different genome-wide resequencing datasets from European individuals . We document several significant correlations between different genomic features . In particular , we find that average minor allele frequency and diversity are reduced in regions of low recombination and that human diversity , human-chimp divergence , and average minor allele frequency are reduced near genes . Population genetic simulations show that either positive natural selection acting on favorable mutations or negative natural selection acting against deleterious mutations can explain these correlations . However , models with strong positive selection on nonsynonymous mutations and little negative selection predict a stronger negative correlation between neutral diversity and nonsynonymous divergence than observed in the actual data , supporting the importance of negative , rather than positive , selection throughout the genome . Further , we show that the widespread presence of weakly deleterious alleles , rather than a small number of strongly positively selected mutations , is responsible for the correlation between neutral genetic diversity and recombination rate . This work suggests that natural selection has affected multiple aspects of linked neutral variation throughout the human genome and that positive selection is not required to explain these observations .
A substantial amount of effort in human population genetics has been aimed at understanding how natural selection operates in the human genome . However , we lack a basic understanding of the importance of positive natural selection versus negative selection at shaping overall patterns of genome variation . Thus far , most of the attention has been aimed at locating genes that have been under positive selection [1]–[19] . These studies have identified several hundred candidates throughout the genome that may have been affected by positive natural selection . However , fewer studies have attempted to gauge the prevalence of positive natural selection in the human genome . Those that have attempted have come to very different conclusions . Several studies suggested that positive selection may be common , with around 10% of the genome having been affected by a recent selective sweep [9] , [10] , [14] , [16] . Other studies argued that selective sweeps were less common [20] , [21] . Finally , some have estimated that approximately 10% , but perhaps up to 40% , of nonsynonymous human-chimp differences have been fixed by positive natural selection [22] , [23] . Thus , there is little consensus regarding the importance of positive natural selection at shaping patterns of variability . Additionally , the role of negative selection at shaping broad patterns of genetic variation across the genome needs to be clarified . Many studies have suggested that nonsynonymous mutations and mutations in conserved noncoding sequences are weakly deleterious but may persist in the population due to genetic drift and other demographic phenomena [22] , [24]–[32] . The effect that these weakly deleterious mutations have on nearby patterns of genetic variation remains unclear . Furthermore , the importance of negative versus positive selection at shaping overall patterns of variation also remains ambiguous . If natural selection ( either positive or negative ) is common in the genome , it should affect patterns of genetic variation at linked neutral sites across the genome [33] , [34] . Selection may alter genetic variation in different ways . We review these ways , discuss the empirical evidence for these effects , and highlight open questions that our study seeks to address . First , selection may generate a correlation between levels of neutral diversity and recombination rate [35] , [36] . This can occur under models with strong positive selection ( selective sweeps ) or negative selection acting on many deleterious mutations ( background selection ) . Selective sweeps remove genetic diversity at linked neutral sites [33] , [37] . In a region of the genome with a low recombination rate , a large length of sequence will have the same genealogy as the selected site . As such , the selective sweep will remove neutral variation over a larger portion of the sequence in low recombination rate regions than in regions with higher recombination rates . Background selection against deleterious mutations can also generate this correlation [34] , [38]–[41] . Chromosomes carrying many deleterious mutations will be rapidly eliminated from the population . Any neutral variation linked to the deleterious mutations will also be eliminated from the population . This model predicts reduced variability in regions of the genome with low recombination rate because , as with the case of a selective sweep , a larger portion of the chromosome will share the same genealogy as the selected site ( s ) in regions of low recombination rather than in high recombination . Several studies have searched for a correlation between diversity and recombination rate in humans . Early studies based on a small number of genes came to conflicting conclusions . Nachman et al . [42] , [43] found a significant correlation between diversity and recombination rate , but found no correlation between divergence and recombination rate , suggesting the effects of natural selection . Hellmann et al . [44] , examining a different dataset , found that the correlation between diversity and recombination rate disappeared after correcting for human-chimp divergence . They suggested that recombination may be mutagenic and that the original correlation was driven by co-variation of mutation and recombination rates . Another study found that microsatellite diversity was not correlated with recombination rate [45] . More recent studies on larger datasets have found significant correlations between diversity and recombination rate [46]–[48] . These studies have found that the correlation between human diversity and recombination rate persists after controlling for human-chimp divergence . While this is suggestive of the effects of natural selection , important features of this correlation have yet to be characterized . For example , if natural selection is primarily driving the correlation , the correlation ought to be stronger in genic regions of the genome than in non-genic regions , because functional sites near genes are the most likely targets of selection . This feature has yet to be explored . Second , natural selection may generate a correlation between the allele frequency distribution and recombination rate . Specifically , models of selective sweeps predict a skew toward an excess of low-frequency single nucleotide polymorphisms ( SNPs ) near the target of selection [49]–[51] . Following the same logic as above , a larger region of the genome will be affected in areas with lower recombination rates , thus generating a correlation between allele frequency and recombination rate . The effect of background selection on allele frequencies is less clear . Simulation studies have suggested that intermediate strengths of background selection , especially in regions of low recombination , can generate a skew toward an excess of low-frequency SNPs [34] , [38] , [52]–[58] . Most of the analytical formulae that describe background selection model the process as a reduction in effective population size , which does not predict a skew of the frequency spectrum ( [34] , [38]–[41] , but see Santiago and Caballero [59] ) . Consequently , it has been argued that the effect of background selection on the frequency spectrum is rather weak , and as such , a skew toward low-frequency SNPs is more indicative of positive , rather than background selection [60]–[66] . It is unclear whether there is a correlation between allele frequency and recombination rate in the human genome , though several small studies have found suggestive evidence [6] , [67] . Furthermore , it is unclear which models of selection may be compatible with such a correlation . Third , if selection is common , it ought to primarily affect patterns of genetic variation near genes because genes are the likely targets of selection . Several studies have found that human-chimp divergence and human diversity were reduced near genes , suggesting the importance of selection at shaping overall patterns of variability throughout the genome [67]–[69] . It is less clear whether there is a skew toward low-frequency alleles near genes . Fourth , pervasive positive natural selection may generate a negative correlation between nonsynonymous divergence and levels of neutral genetic diversity ( [70]–[77] and reviewed in [78] ) . The reason for this is that selective sweeps acting on amino acid changing mutations generate nonsynonymous fixed differences between species . Regions of the genome that have been affected by these sweeps will likely also have reduced neutral polymorphism , thus generating the negative correlation between these two quantities . It is unclear whether such a correlation can be generated in the absence of positive selection and how strong the correlation might be under various models of positive selection . Here we further investigate these issues by studying patterns of genetic variation in three different genome-wide genetic variation datasets obtained from resequencing European individuals . We find that levels of diversity are positively correlated with recombination rate and negatively correlated with genic content . Minor allele frequency is also positively correlated with recombination rate and negatively correlated with genic content . Using simulations , we show that these correlations are best explained by a model where many sites are under weak negative selection . Models with numerous selective sweeps on nonsynonymous mutations predict too strong a negative correlation between neutral polymorphism and nonsynonymous divergence . Though not required to explain the data , some smaller fraction of sites may be under positive selection . Overall , this work points to the importance of weak negative selection at shaping patterns of variation throughout the human genome .
We analyzed genomic patterns of polymorphism from three genome resequencing datasets . First , we analyzed low-coverage next-generation sequence data obtained from an exome-capture study of 2 , 000 Danish individuals . Due to the non-specificity of the exome-capture arrays , portions of the genome outside of the targeted regions were sequenced , but at lower coverage . Given the shallow sequencing depth across most of the genome ( roughly 0 . 1× per individual ) , it would be impossible to infer genotypes for each individual with any appreciable accuracy . Instead , we implemented a statistical approach to estimate the population allele frequency of a SNP using the counts of different nucleotides at a particular site in the genome ( see Materials and Methods for a detailed description ) . When combining reads across all individuals , approximately 30–40% of the genome had a sequencing depth of at least 100 reads . We estimated the minor allele frequency ( MAF ) for all of these sites with a depth of at least 100 reads . Those sites with an estimated MAF>5% were considered to be SNPs in this dataset and were used for subsequent analyses . We used this conservative cut-off because of the difficulties in reliably estimating allele frequencies of rare alleles in low-coverage data [79] . In order to verify patterns found in our low-coverage resequencing dataset , we also analyzed two other complementary datasets . One dataset consisted of six European genomes that were sequenced to higher coverage ( denoted “higher coverage , ” see Materials and Methods for details ) . The other dataset consisted of five genomes from Utah residents with ancestry from northern and western Europe ( abbreviated CEU ) and one genome from a Toscan individual sampled from Italy ( abbreviated TSI ) sequenced to high coverage by Complete Genomics ( denoted “CGS , ” see Materials and Methods ) . Summaries of genetic variation were positively correlated across the three datasets ( Figure S1 and Figure S2 ) . Due to the stochasticity of the evolutionary process , even with perfect data , patterns of polymorphism will not be perfectly correlated across different datasets . To analyze correlations between different summaries of polymorphism and other genomic features , we divided the genome into non-overlapping 100 kb windows ( see Materials and Methods for further details ) . Within each window , we tabulated the number of SNPs , average MAF , number of human-chimp differences , GC content , recombination rate ( as estimated from the high-resolution deCODE map [80] ) , fraction of each window where sequencing data was available , and the fraction of the window that overlaps with a RefSeq gene . Since we wanted to examine the indirect effects of natural selection due to linkage , rather than assess the effects of natural selection on the selected sites themselves , all of our analyses removed the roughly 5% of the genome that was most conserved across species ( i . e . the phastCons regions [81] , see Materials and Methods ) . These were the regions most likely to be directly under negative selection in the human genome [81] . We then assumed that the remaining sequence that we analyzed was selectively neutral . Because many of the genomic features were correlated with each other ( Table S1 , Table S2 , Table S3 ) , we performed partial correlation analyses to remove the effects of possible confounding variables . The partial correlation can be thought of as the correlation between two variables when one or more other confounding variables are held constant . We used partial correlations , rather than a full multivariate analysis , because the partial correlations have a simpler biological interpretation and have been used in other recent evolutionary studies [82] . We found a strong positive correlation between the number of SNPs in a window and the recombination rate of the window ( Spearman's , Table S1 ) when looking at the low-coverage data . We also observed a strong correlation between the number of human-chimp differences within a window ( d ) and recombination rate ( Spearman's , Table S1 ) . When scaling diversity by divergence ( i . e . dividing the number of SNPs per covered base within a window by the number of human-chimp differences ) to potentially account for differences in mutation rate across the genome , we still found a strong correlation between scaled SNP diversity ( defined here as Snorm ) and recombination rate ( Spearman's , Table 1 , Table S1 ) . In particular , regions of the genome with low rates of recombination ( i . e . <0 . 5 cM/Mb ) had especially low levels of polymorphism . The rate of change of Snorm was less dramatic over the rest of the range of recombination rates . We also found a positive correlation between Snorm and recombination rate when analyzing the higher-coverage and CGS datasets ( Spearman's , Table 1 , and Table S2 for the higher-coverage data; Spearman's , Table 1 , and Table S3 for the CGS data ) . The correlation was even stronger than that observed in the low-coverage data . We discuss several possible reasons for this difference in the Discussion section . Nevertheless , the fact that we found the correlation in all three datasets strongly argues that it is a true biological correlation and not an artifact due to biases in the low-coverage Danish data . The correlation between Snorm and recombination rate remained significant even after controlling for GC content , d , the number of neutral bases covered by sequencing data , and the fraction of genic bases within a window ( Table 1 ) , suggesting that these factors cannot completely explain this correlation . Further , the average number of pairwise differences per window normalized by d was also positively correlated with recombination rate in both datasets ( Table S2 and Table S3 ) . If natural selection is responsible for this correlation between Snorm and recombination rate , it may be stronger in genic regions of the genome than in non-genic regions . The reason for this is that , all else being equal , genic regions will likely experience more natural selection than non-genic regions . Non-genic windows were defined to be those that did not overlap with a RefSeq transcript . Genic windows were those where at least half the window overlapped with a RefSeq transcript . Indeed , the correlation was significantly stronger in genic windows than in non-genic windows in all three datasets ( P<0 . 0001 by permutation test , Figure 1 , Table 2 , Figure S3 , and Figure S4 ) . This pattern holds even after controlling for confounding variables using a partial correlation analysis . Inspection of the lowess lines in Figure 1A illustrates the differences between the correlation in genic and non-genic regions . In genic regions with low recombination rates ( <0 . 5 cM/Mb ) , there is a sharp decrease in Snorm . However , non-genic regions with low recombination rates did not show such a pronounced decrease in Snorm ( Figure 1A ) . One concern with these analyses is that the low-coverage dataset was an exome resequencing dataset and the exome-capture process may have resulted in systematic differences between genic and nongenic regions . However , we found the same pattern in the higher-coverage dataset and the CGS dataset , which were not targeted toward genes or exons ( Figure S3 and Figure S4 ) . This argues that the differences between genic and non-genic regions were not due to systematic biases in the data , but rather to inherent differences between genic and non-genic regions of the genome . We then examined the correlation between average MAF within a window and recombination rate ( Table 1 and Table S1 ) in the low-coverage data . We found a weak , but statistically significant , positive correlation between these two variables ( Spearman's ) . In regions of low recombination , there was a skew toward lower average MAF . The correlation remained significant even after controlling for GC content , d , the number of neutral bases covered by sequencing data , and genic content , suggesting that it cannot be completely explained by these other factors ( Spearman's , Table 1 ) . Finally , we also found a positive correlation between average MAF and recombination rate in the higher-coverage and the CGS data ( Table 1 , Table S2 , Table S3 ) , again suggesting that it was not due to biases in estimating SNP frequencies from low-coverage data . A different summary of the frequency spectrum , Tajima's D [49] , also showed a correlation with recombination rate ( Table S2 and Table S3 ) , indicating that this correlation was not sensitive to the summary of the frequency spectrum employed . However , no clear pattern emerged when testing whether the correlation between average MAF and recombination rate was stronger in genic versus non-genic regions . For all three datasets , the pairwise correlation between average MAF and recombination rate was higher in genic regions than non-genic regions ( P<0 . 05 , by permutation test , Figure 1B , Figure S3B , Figure S4B , Table 2 ) . In the higher-coverage dataset , genic regions showed a stronger correlation between MAF and recombination rate than non-genic regions even after controlling for GC content , d , and the number of bases covered by sequencing data using a partial correlation analysis ( P<0 . 02 by permutation test , Table 2 ) . However , after controlling for the confounding variables , there was little difference in the partial correlation coefficients between genic and non-genic regions in the low-coverage and the CGS datasets ( Table 2 ) . Thus , there was no clear evidence suggesting that the correlation between MAF and recombination rate was stronger in genic than non-genic regions of the genome . This may not be surprising because this correlation was quite weak , making it difficult to detect subtle changes in its strength across the genome . If natural selection affects patterns of genetic variation across the genome , Snorm , average MAF , and d may be reduced in windows of the genome that contain more genic bases . These patterns would be expected if most of the selection in the genome occurs near genes , rather than in intergenic regions . Indeed , in all three datasets , we found a negative correlation between Snorm and the fraction of bases within a window that overlapped with a RefSeq transcript ( Table 1 ) . In other words , windows with a higher genic content tended to have fewer SNPs . These correlations became stronger when controlling for d , recombination rate , the fraction of the window with sequencing coverage , and GC content ( Table 1 ) . There was a weak , but significant , negative correlation between MAF and fraction of bases that overlapped with a RefSeq transcript in all three datasets examined ( Table 1 ) . Windows with a higher genic content tended to have lower average MAF than windows with lower genic content . In the low-coverage and higher-coverage datasets , the correlation became stronger when controlling for d , recombination rate , the fraction of the window with sequencing coverage , and GC content ( Table 1 ) . Finally , we found a very strong negative correlation between d and the fraction of genic bases within a window ( Spearman's , Table S1 , Table S2 , Table S3 ) . These results were in agreement with those from a study [67] which found reduced diversity and divergence near genes even after removing the regions of the genome most conserved across species ( i . e . the phastCons elements ) . We next tested whether there was a correlation between Snorm and the number of nonsynonymous human-chimp differences within a window ( DN ) . A negative correlation between these two variables has been interpreted as evidence of selective sweeps across the genome ( [70]–[77] and reviewed in [78] ) . When tabulating DN , we did not remove sites which were conserved across species . We observed weak negative correlations between Snorm and DN as well as between Snorm and the number of synonymous human-chimp differences ( DS ) for several of the datasets ( Table S4 ) . However , when we normalized DN by the number of nonsynonymous sites per window ( the normalized value is called dN ) or used a partial correlation analysis to control for the number of nonsynonymous sites per window , none of the datasets showed a significant negative correlation ( Table S4 ) . The same was true for synonymous human-chimp differences . Haddrill et al . [77] suggested that a negative correlation between Snorm and dN may be more apparent in genes with elevated dN . Thus , we also tested for a correlation between Snorm and dN using only the windows in the 90th percentile of dN . In general , the values of Spearman's were more negative in this subset of the data than when analyzing the entire dataset ( Table S5 ) . For example , in the CGS data , when controlling for d , GC content , recombination rate , the number of nonsynonymous sites , and the fraction of the window with sequencing coverage . However , Snorm was also negatively correlated with dS in the windows in the 90th percentile of dS ( , controlling for d , GC content , recombination rate , the number of synonymous sites , and the fraction of the window with sequencing coverage ) . The fact dS showed a similar negative correlation with Snorm as dN did , combined with the fact that synonymous sites are usually assumed to be neutrally evolving in humans , suggested that these correlations may have been driven by a neutral process , rather than positive selection . One possibility was that the recent fixations of neutral synonymous or nonsynonymous mutations led to a decrease in neutral diversity , as suggested by earlier theoretical work [83] . As such , regions with high dN ( or high dS ) would have lower Snorm , generating the negative correlation . Overall , these results suggest that regions of the genome that have more nonsynonymous human-chimp differences do not have lower levels of neutral polymorphism , beyond the reduction in diversity already expected in genic regions of the genome or surrounding neutral fixations . We next evaluated whether population genetic models including population size changes , recombination rate variation , and natural selection could generate the correlations that we observed in the empirical datasets . We simulated 100 kb regions consisting of exons , introns , and an intergenic sequence ( see Materials and Methods , Figure S5 ) . We examined several different models of selection ( see Table S6 for the specific parameter values ) and examined the correlation between patterns of genetic variation in the neutrally evolving intergenic sequence and other genomic attributes . Because many studies have found that nonsynonymous mutations are weakly deleterious [22] , [26] , [28] , [84] , one model included weak negative selection acting only on nonsynonymous sites ( shown in purple in Figure 2 ) . It had been suggested that conserved noncoding sites are also likely to be weakly deleterious [25] , [27] , [31] , so another model included negative selection acting on a fraction of intronic sites ( shown in blue in Figure 2 ) . In the third model ( shown in orange in Figure 2 ) , most mutations at nonsynonymous positions were negatively selected , but a small fraction was positively selected . Finally , the fourth model added weak negative selection at a fraction of intronic sites to a model where most mutations at nonsynonymous positions were negatively selected , but a small fraction was positively selected . Our simulations confirmed previous predictions that both hitchhiking and background selection [33] , [34] , [37]–[41] could generate a positive correlation between genetic diversity at linked neutral sites and recombination rate ( Figure 2A and Figure S6A ) . Importantly , these simulations demonstrated that the background selection effect can occur with weak negative selection acting on many sites simultaneously . Models with negative selection acting on noncoding and coding mutations , as well as models with positive selection , could generate positive correlations similar to those in the observed data ( red lines in Figure 2A and Figure S6A ) . Models of natural selection predicted a positive correlation between average MAF at linked neutral sites and recombination rate ( Figure 2B and Figure S6B ) . The strongest correlations seen for models with only negative selection were for intermediate strengths of selection ( e . g . 25% of intronic sites with s = 2 . 5×10−4 ) . Stronger selection ( s = 5×10−3 ) resulted in a weaker correlation ( Table S7 ) . Importantly , models that contained no sites under positive selection predicted a correlation between MAF and recombination rate roughly similar in magnitude to that seen in the observed data ( red lines in Figure 2B and Figure S6B ) . These results suggest that both positive and weak negative selection were capable of affecting allele frequencies at linked neutral sites . Thus , a correlation between allele frequency and recombination rate cannot be taken as unambiguous evidence of positive selection . In some cases , the correlation coefficients between MAF and recombination rate and diversity and recombination rate were significantly higher than zero under purely neutral models ( Figure 2 and Figure S6 ) . We performed coalescent simulations using ms [85] under the standard neutral model with different rates of recombination to further investigate this issue . Not only was the variance of the distribution of diversity ( or average MAF ) greater in simulations without recombination , but the shape of the distribution changed depending on the recombination rate . For example , in the case of a high recombination rate , the distribution of the number of segregating sites approached a Poisson distribution , and was symmetric about its mean . However , with no recombination , the distribution became less symmetric , with a higher mass below the mean and a longer tail to the right ( Figure S7 ) . Thus , the median of the distribution of diversity simulated with no recombination was lower than the median of the distribution with the high recombination rate . As such , a weak positive correlation between recombination rate and diversity may be expected . The same arguments hold for understanding the correlation between MAF and recombination rate ( Figure S7 ) and Tajima's D and recombination rate ( Figure S7 , see also [63] , [86] ) . Since we used simulations to interpret the correlations observed in the actual data , this effect did not alter our interpretation . Previous authors ( [70]–[77] and reviewed in [78] ) had suggested that a negative correlation between neutral polymorphism and nonsynonymous divergence may be a signature of positive selection that cannot be generated by negative selection and/or demographic processes . In our simulations , a model with negative selection acting on noncoding sites , but where a fraction of coding mutations were positively selected showed a negative correlation between Snorm and dN ( orange points in Figure 2C and Figure S6C ) . Models that did not include any positive selection , but included negative selection on a fraction of noncoding sites ( blue points in Figure 2C and Figure S6C ) , showed little correlation between these two variables . Thus , for the models investigated here , the negative correlation was specific to models of positive selection . As such , it may offer a way to distinguish between models of negative and positive selection . However , a significant negative correlation was not always seen in models that included some sites under positive selection ( green points in Figure 2C and Figure S6C ) . Instead , the correlation was influenced by the relative amounts of negative versus positive selection . Negative selection made the correlation more positive , while positive selection made the correlation more negative . The correlation ultimately observed was due to the net effect of both types of selection . We next used the simulations to evaluate what role positive selection may have played in shaping patterns of variability across the genome . We first examined models with only strong positive selection . A model where 0 . 5% of nonsynonymous mutations were positively selected ( s = 0 . 625% ) could generate the observed correlation between Snorm and recombination rate ( black , p+ = 100% , p− = 0% in Figure 3A; p+ denotes the proportion of simulated windows where positive selection could occur ) . However , this model predicted too strong a negative correlation between Snorm and dN to be compatible with the data ( black , p+ = 100% , p− = 0% in Figure 3B ) . Because several studies have suggested that 0–10% of the genome has been affected by a selective sweep [9] , [10] , [14] , [16] , [20] , [21] , we next examined a model where 5% of the simulated windows included positive selection . A model where the remaining 95% of the windows were neutral does not predict a correlation between Snorm and recombination rate strong enough to match the actual data ( black , p+ = 5% , p− = 0% in Figure 3A ) . This suggests that a small number of positively selected sites by themselves are not sufficient to generate this correlation . Further , this model still predicted a negative correlation between Snorm and dN ( black , p+ = 5% , p− = 0% in Figure 3B ) . However , a model where 5% of the simulated windows included positive selection and the remaining 95% of windows included negative selection on coding and noncoding sites predicted a correlation between Snorm and recombination rate similar to that observed in the actual data ( black , p+ = 5% , p− = 95% in Figure 3A ) . Because adding negative selection resulted in an increase in the strength of this correlation , we concluded that the correlation observed in the data has been primarily driven by negative selection . Also , under this model , the negative correlation between Snorm and dN was very weak and was compatible with that from the actual data ( black , p+ = 5% , p− = 95% in Figure 3B ) , presumably because most of the windows have been subjected to negative selection . A model where the strength of positive selection was weaker showed similar trends ( pink points in Figure 3 ) . This analysis indicated that the correlation between neutral diversity and recombination rate was primarily driven by many weakly deleterious polymorphisms across the genome , rather than by a small proportion of strongly positively selected mutations . Finally , our simulations ( Figure 4 ) suggest that negative or positive selection can generate a strong correlation between neutral human-chimp divergence ( d ) and recombination rate even when the mutation rate is constant across all simulation replicates . This correlation was likely driven by selection occurring in the ancestral population [67] , [87] . Thus , the correlation between d and recombination rate can be readily explained by mechanisms other than recombination itself being mutagenic [44] , [46] , [88] .
We have examined patterns of putatively neutral genetic variation in three genome-wide resequencing datasets to gauge the extent of natural selection throughout the human genome . To the best of our knowledge , this is the first report that the allele frequency spectrum is correlated with recombination rate across the human genome ( though suggestive evidence was found in smaller datasets [6] , [67] ) . As discussed below , these correlations are best explained by natural selection affecting linked neutral variation across the human genome , rather than artifacts in the data or other mutational processes . Through the use of population genetic simulations , we have shown that a model with negative selection acting on both coding and noncoding mutations fits the data . While we cannot rule out models that include some positive selection , models with abundant positive selection on nonsynonymous mutations and little negative selection predict too strong a negative correlation between neutral polymorphism and nonsynonymous divergence . In general , we observed qualitatively similar patterns in all three resequencing datasets . However , several of the correlations between different genomic attributes were stronger in the higher-coverage and CGS data than in the low-coverage data ( Table 1 and Table 2 ) . Several characteristics of the datasets may contribute to this difference . For example , the higher-coverage and CGS datasets are likely to be of higher quality than the low-coverage dataset . Additionally , a greater proportion of each window is covered in the higher-coverage and CGS datasets than in the low-coverage dataset ( Figure S8 ) . Both of these features lead to estimated correlation coefficients that are lower in the low-coverage data than in the higher-coverage and CGS data . Finally , the higher-coverage and CGS datasets contain a sample of a smaller number of chromosomes than the low-coverage data . Population genetic simulations suggest that some of the correlations are expected to be stronger in smaller samples than in larger samples ( compare Figure 2 to Figure S6 ) . Thus , the quantitative differences among the correlation coefficients across the different datasets are not too surprising . Instead , the fact that all three datasets show the same general trends is powerful evidence that the correlations are not technical artifacts specific to any one type of data . Thus , it is our conclusion that these correlations were , at least in part , driven by natural selection across the human genome . Several lines of evidence support this conclusion . First , the correlations remain significant after filtering repetitive sequence and CpG islands ( Table S8 ) , and after controlling for the effects of GC content , suggesting that base composition or mutational patterns associated with base composition are not entirely responsible for the correlations . Second , we have evaluated whether biased gene conversion , a neutral alternative sometimes invoked to explain signatures of natural selection [89] , [90] , can generate the correlations we have identified . Our simulations show that neutral models with biased gene conversion cannot generate a correlation between Snorm and recombination rate similar in magnitude to that observed in our datasets ( Table S6 and Table S7 ) . The third line of evidence is that the correlation between neutral polymorphism and recombination rate is stronger in genic regions compared to non-genic regions . Natural selection would predominately occur closer to genes , while mutational effects would be distributed throughout the genome [88] . We have also found that both diversity and minor allele frequency are negatively correlated with genic content , suggesting a difference in patterns of variability between genic and non-genic regions of the genome . As discussed further below , models that include natural selection can readily account for these observed patterns . We have explored which models of selection can generate the correlations that we observed in the actual data . While we have found population genetic models that qualitatively predict the correlations that we have observed in the data , it is more difficult to translate these models into specific statements about the absolute amount of selection in the genome . For example , many of our simulations of negative selection on noncoding sites assume that 25% of intronic sites were under weak negative selection . This is likely to be a substantial over-estimate of the proportion of sites under negative selection [81] , [91] . One explanation for this discrepancy is that , for computational convenience , we simulated 100 kb windows independently of each other , rather than whole chromosomes . In reality , each 100 kb window of the genome is linked to other selected mutations outside of the window that may affect patterns of diversity within the window . In fact , simulations of larger windows ( 348 kb ) provide similar values of Spearman's when only 5% of intronic sties are under negative selection ( Table S6 and Table S7 ) . This may explain why the models that fit the data include so many selected sites . Simulating larger regions would only yield more biologically relevant simulations if we were able to simulate the correct magnitude of selection at noncoding sites , as well as the correct spatial distribution of sites under selection across the genome . Though there has been some progress from comparative and population genomic studies [25] , [27] , [31] , [81] , [91] , further work is needed in this area . Additionally , there are nearly an infinite number of possible models for how selection can operate in the genome . For example , selection coefficients within a given window may be correlated with each other , and windows may not be exchangeable ( i . e . each window may have its own distribution of selective effects ) . Our simulations do not capture these phenomena and instead merely illustrate the types of correlations predicted for very basic models of certain types of selection . Nevertheless , our simplified models do allow some important qualitative statements regarding the relative importance of negative versus positive selection in the human genome . First , all of the correlations observed in all three datasets can be explained without invoking positive selection . Different models of negative selection can readily account for these correlations ( Table S6 and Table S7 ) . Second , based on the lack of a negative correlation between Snorm and dN in any of our datasets ( Table S4 , Figure 2C , Figure S6C ) , we can reject models with an abundance of selective sweeps acing on nonsynonymous mutations in the presence of few negatively selected sites ( Figure 2 , Figure 3 , Figure S6 ) . This finding is complementary to what was found in a recent study by Hernandez et al . [21] . However , we cannot rule out the presence of some positively selected mutations in the presence of many negatively selected ones . It is difficult to precisely estimate the fraction of the genome that has been affected by positive selection because such inferences are likely to be highly model-dependent and influenced by many unknown variables . Yet , for the model shown in Figure 3 , which fits the actual data ( black , p+ = 5% , p− = 95% ) , 5% of the simulated windows included positively selected mutations . This model predicts that roughly 2 . 3% of the windows will have at least one positively selected nonsynonymous mutation that fixed in humans within the last Ne generations ( here 20 , 000 generations , or 500 , 000 years , assuming 25 years per generation ) . This is likely to be an upper bound on the fraction of the genome subjected to such strong positive selection because a higher fraction would predict a negative correlation between Snorm and dN that is too strong to match the data . However , if the strength of positive selection on individual mutations is weaker , if selection operates on standing variation , predominantly on noncoding mutations , or on multiple mutations simultaneously , then a much greater fraction of the genome could have been subjected to positive selection [20] , [92]–[94] . Nonetheless , even if a small fraction of the genome was linked to a selective sweep , this amount of selection is not sufficient to generate the correlation between diversity and recombination rate seen in the actual data ( Figure 3A ) . The widespread presence of weakly deleterious alleles , however , can generate this correlation , even in the presence of some positively selected sites ( Figure 3A ) . Taken together , our results suggest that selective sweeps were not the dominant factor explaining the distribution of variability across the human genome . The notion that sites under natural selection can affect linked neutral variation in the human genome has several important implications for learning about human history using genetic variation data . Most methods to infer parameters in population genetic models assume that all of the SNPs being analyzed are selectively neutral and are not linked to other sites that are affected by selection [17] , [95]–[100] . Many of these methods summarize the genetic variation data by the number or proportion of SNPs at different frequencies in the sample ( i . e . the frequency spectrum ) and then find the demographic parameters that can generate the observed frequency spectrum . Compared to other regions of the genome , we found an excess of low-frequency SNPs in regions near genes and with low recombination rate . It is unlikely that these regions provide an accurate picture of the selectively neutral frequency spectrum for the population of interest . It is unclear what effect including such regions in demographic studies will have on the final parameter estimates . Further investigation of this topic is warranted . In the meantime , one way of circumventing the potential problem of natural selection confounding studies of demography would be to study regions of the genome far away from genes and with high recombination rate [101] . Finally , our study illustrates the utility of low-coverage sequencing data for population genetic studies . Here we have shown that analyzing the low-coverage data without first inferring individual genotypes provides estimates of allele frequency across the genome that are in broad agreement with estimates made from higher-coverage sequencing of a smaller number of individuals . Another unique feature of the low-coverage dataset was that it was generated as part of an exome-capture experiment [84] . Because the capture process is not completely specific and only enriches for sequences within the targeted regions , portions of the genome outside of the targeted regions were sequenced at a lower rate . Such data from a large number of individuals can be used to study patterns of genetic variation across the non-targeted regions of the genome , provided that one analyzes it using an approach that is appropriate for low-coverage data . Such studies promise to yield new insights in population and medical genetics .
The low-coverage dataset that we used here was an augmented version of the dataset published in Li et al . [84] . The sequencing was performed on 2 , 000 Danish individuals ascertained from three sources: 1 ) the population-based Inter99 study [102] ( ClinicalTrials . gov ID-no: NCT00289237; n = 887 ) , 2 ) the ADDITION study [103] ( ClinicalTrials . gov ID-no: NCT00237548 ) ; n = 354 ) and 3 ) the Steno Diabetes Center ( n = 759 ) . All participants ( mean age of 54 . 5 years ) were of self-reported Danish nationality . All study participants provided written informed consent , and the study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Copenhagen County , Denmark . DNA from these individuals was analyzed in an exome-capture resequencing experiment . Each individual was sequenced separately without any pooling . NimbleGen2 . 1M HD arrays were used to enrich for exome sequences . These arrays contain probes complementary to exonic DNA fragments . Exonic DNA hybridized to the array while non-exonic DNA was washed away . However , this hybridization process was not perfect , and some non-exonic DNA remained bound to the array and was sequenced . The Illumina Genome Analyzer II was used to perform the sequencing . Further methodological details can be found in Li et al . [84] . The bioinformatic pipeline used for these data is similar to the one previously published [84] . First , reads were aligned to the NCBI human genome reference assembly ( build 36 . 3 ) using SOAPaligner [104] , [105] . Reads that mapped outside of the exome target regions were retained for further analyses , but bases with a Q score <20 were removed . Ideally , since we wish to compare allele frequency estimates for different regions of the genome , we would like to have a similar depth of coverage across the genome . However , depth of coverage varied greatly across the genome with the target regions having very high coverage and the non-target regions having substantially lower coverage . To circumvent this problem , at each position in the genome , we selected a random subset of 100 reads ( from the 2 , 000 individuals ) to be used for the frequency estimation process . We chose a cutoff of 100 reads since about 35–40% of the total genome was covered by at least 100 Q >20 bases . Decreasing this cutoff would increase the number of bases that were covered , but it would also make it harder to accurately estimate the frequency of lower-frequency SNPs . We estimated allele frequencies directly from the read counts without attempting to call SNPs or individual genotypes from these data . For each site in the genome with at least 100 reads , we first estimated the population minor allele frequency ( MAF ) using the method-of-moments estimator [84] . For sites that had an estimated MAF >1% using , we obtained a more precise estimate of the MAF using the maximum likelihood approach described by Kim et al . [79] , [106] . Due to computational constraints on analyzing a dataset of this size , we did not use the genotype likelihood files from soapSNP [107] . Rather , we used the binomial distribution to compute the probability of the read counts for each individual , taking the base-specific sequencing error probabilities into account . We treated the second-most common base at each site as the minor allele . Finally , only sites with estimated MAF >5% were considered as SNPs and were used in subsequent analyses . Given the low depth of coverage ( 100 reads ) , it would be difficult to distinguish lower-frequency SNPs from sequencing errors . For example , for a SNP with a MAF of 1% , the less common allele would only be seen approximately one time across all individuals . We also analyzed a dataset of six European individuals whose genomes were sequenced to higher coverage . This dataset is complementary to the low-coverage dataset because each individual in this dataset was sequenced to higher coverage , coverage was more uniform across the genome , and a higher fraction of bases were covered . But , the sample depth at any particular site in the genome was substantially lower ( only 12 chromosomes at most ) . This dataset included the genomes of James Watson [108] , Craig Venter [109] , the two parents from a CEU trio ( NA12891 and NA12892 ) that was sequenced to high coverage in pilot 2 of the 1000 Genomes project [69] , and two European genomes ( NA07022 and NA20431 ) sequenced by Complete Genomics [110] . Since each individual's genome was sequenced to higher coverage , we treated the called genotypes as though they were the true genotypes throughout subsequent analyses . For the Venter and Watson genomes , we downloaded SNP genotypes from the “Genome Variants” table of the UCSC browser . Coverage information across these two genomes was obtained from “emf” files from the Ensembl database . Sites with a score of 1 or greater were considered covered . SNPs overlapping regions with a lower score as well as indels and other structural variants were dropped from the analysis . Sites that were covered by reads , but did not have a SNP genotype were considered to be homozygous for the reference genotype . We downloaded the “ . vcf” and “mask” files for the CEU trio of the 1000 Genomes Project . Genotypes for variable positions were obtained from the . vcf files . For the rest of the genome , the individuals were assumed to be homozygous for the reference allele if SNP calling was attempted at the position ( i . e . the position had a score of “0” in the mask file ) . A small number of reported SNPs in the . vcf files that fell in masked positions of the genome were removed from subsequent analyses . Coverage and SNP genotype information could be directly obtained from the Complete Genomics “variations” files . SNPs and positions that were within 2 bp of indels or structural variants were removed from subsequent analyses . We intersected the variant genotypes and coverage information from all six genomes and called genotypes for each individual . SNPs with more than two different alleles across all individuals or SNPs where one of the two alleles did not match the reference sequence were removed from subsequent analyses . For sites where one individual had a variant genotype , the genotypes for the other individuals who did not have a variant allele were considered to be homozygous for the reference if they had coverage at that particular site , or were considered to be missing if they did not have any coverage . Subsequent analyses of diversity levels and MAF only used those SNPs and sites that were covered in all six individuals . We also analyzed six European genomes sequenced by Complete Genomics ( CGS ) . Five of the genomes were from the CEU sample ( NA06985 , NA06994 , NA07357 , NA10851 , and NA12004 ) and one was from a TSI individual ( NA20502 ) . We used the genotype calls made by CGS that were found in the “masterVarBeta” files . SNPs with more than two different alleles across all individuals , SNPs where one of the two alleles did not match the reference sequence , and sites that were within 2 bp of structural variants called in any one of the individuals were removed from subsequent analyses . Later analyses of diversity levels and MAF only used those SNPs and sites that were covered in all six individuals . We noted that some windows of the genome appeared to have an unusually high number of SNPs where many individuals were heterozygous ( Figure S9 ) . We removed windows which had at least 10 SNPs where the average number of heterozygous genotypes per SNP was greater than 3 ( out of 6 ) . This filtering resulted in dropping 3 . 8% of the windows and appeared to remove the outlier regions ( Figure S9 ) . We divided the genome into non-overlapping 100 kb windows . Windows that were within 10 Mb of an annotated centromere , telomere , or end of a chromosome were omitted from further analyses . For each window , we tabulated several genomic features . First , we obtained the recombination rate for each window using the high-resolution pedigree-based genetic map assembled by deCODE [80] . Second , we tabulated the number of sites within each window where the hg18 base differed from the pantro2 base . This was done using the . axt alignments obtained from the UCSC browser . Importantly , bases in RepeatMasked parts of the genome or where the hg18 or pantro2 alleles were missing were not counted . Since we wanted to examine putatively neutral sites , bases falling in the 17-way phastCons regions were also not counted [81] , except when analyzing synonymous and nonsynonymous human-chimp divergence ( see below ) . Third , we tabulated GC content within each window as the fraction of bases where the hg18 sequence was a G or a C . Only those bases that met the inclusion criteria described above were counted in this analysis . Fourth , as a measure of genic content , we tabulated the proportion of bases within each window that overlapped with a RefSeq transcript . We then tabulated the number of SNPs within each window and the number of bases that had sequencing coverage ( see above for the criteria used to define covered bases ) . Importantly , SNPs falling RepeatMasked regions or phastCons regions were dropped from the analysis . Similarly , these bases were not counted as covered bases . The number of SNPs per covered base was used as a summary of diversity within each window . Finally , we summarized the frequency spectrum within each window by the average MAF over all the SNPs within each window . We tested for correlations between the variables described above using non-parametric correlation tests . Specifically , we tested for pairwise correlations between variables using Spearman's . Since many of the variables were correlated with each other ( Table S1 ) , we calculated partial correlations to remove the effects of confounding variables on the variables of interest . Partial correlation statistics were calculated using the pcor function in R [111] . We tested whether the correlations were stronger in genic windows compared to non-genic windows using a permutation test . For each permutation , windows were randomly assigned to a genic and a non-genic group , keeping the number of genic and non-genic windows equal to that in the observed data . We recorded the difference in the correlation coefficient between each permuted genic and permuted non-genic dataset . The P-value for the test was the proportion of 10 , 000 permuted datasets with differences larger than those seen in the non-permuted data . To test for a correlation between neutral polymorphism ( Snorm ) and nonsynonymous divergence , we found the number of nonsynonymous hg18-pantro2 alignment differences in each window ( DN ) . This was done by putting those alignment differences that were not in RepeatMasked sequence and overlapped with an exon in the Consensus Coding Sequence ( CCDS ) table from the UCSC Table Browser into the SeattleSeq SNP annotation pipeline ( http://gvs . gs . washington . edu/SeattleSeqAnnotation/ ) . The human and the chimp bases were used as the two alleles . If multiple CCDS genes overlapped , we selected the longest one and discarded the remainder . We used the Nei-Gobjori [112] approach with the CCDS gene model to count the number of synonymous ( LS ) and nonsynonymous ( LN ) sites per window . LN and LS were only counted from the hg18 sequence , rather than averaged between the hg18 and pantro2 sequences . Only those sites that were not Repeat-Masked and were aligned with pantro2 were counted . The number of nonsynonymous differences per nonsynonymous site ( dN ) was then calculated as DN/LN . Similarly , the number of synonymous differences per synonymous site ( dS ) was then calculated as DS/LS . To determine which models of selection could generate the correlations we observed in the resequencing data , we performed forward-in-time population genetic simulations using the program SFS_CODE [113] . Specifically , we simulated 100 kb regions that included exons and introns separated by an intergenic spacer region ( Figure S5 ) . We assumed a Jukes-Cantor mutation model [114] with a per-base pair mutation rate of 2 . 5×10−8 . Figure S10 shows the demographic model used for the simulations . Briefly , we simulated a human population with a chimp outgroup where the chimp population split from the human population 5 million years ago ( assuming 25 years per generation ) . The ancestral human-chimp population was assumed to be of size 20 , 000 because previous studies have found that the ancestral human-chimp population was likely 2–10-fold larger than the current human effective population size [115]–[119] . At the human-chimp speciation event , the both the chimp and human populations underwent an instantaneous 2-fold contraction to their current sizes . Since our data consisted of European individuals , we also included a bottleneck in the human population with parameters from Lohmueller et al . [120] , but using an ancestral population size of 10 , 000 between the human-chimp split and the more recent bottleneck . The recombination rates for the simulated windows were chosen to approximately match the distribution of estimated recombination rates of the genic windows from the low-coverage dataset . This was done by assigning each window in the low-coverage data to one of 100 different bins based on its recombination rate . A single recombination rate was chosen for each bin ( the mid-point of the bin ) , and this rate was used to simulate the number of replicates proportional to the number of windows in the actual data falling into the bin . A total of 20 , 000 simulated windows were generated for each model of selection . Recombination hotspots were added to each window . Hotspots were assumed to have a width of 2 kb and the inter-hotspot distances were drawn from an exponential distribution with a mean of 20 kb . To specify the intensities of the hotspots in SFS_CODE , one needs to provide the proportion of the total amount of recombination that occurred within each of the hotspots and coldspots . We set the proportion of recombination that occurred in hotspot i to be 0 . 8xi , where xi was drawn from a Dirichlet distribution with parameter k equal to the number of hotspots within the window , and . This framework allowed hotspots to have different intensities and kept the total proportion of recombination that occurred in hotspots in each window at 80% [121] . A similar approach was used to determine the background recombination rates for each part of the sequence outside of the hotspots , except 0 . 2 was used instead of 0 . 8 . We examined several different models of natural selection ( Table S6 ) . In most models , nonsynonymous mutations were weakly deleterious with their selection coefficients drawn from a gamma distribution of selective effects , the parameters of which had been estimated from human resequencing data [22] . Some models also included positive selection acting on a fraction of nonsynonymous mutations , or a fraction of intronic mutations that were weakly deleterious . We then tabulated diversity and divergence summary statistics from the simulations . Importantly , we only analyzed SNPs and human-chimp differences that occurred in the neutral intergenic sequence . For comparison to the low-coverage Danish data , we used the population MAFs from the simulations , counting only those SNPs with MAF >5% as we did in the observed data . From these same simulations , we took a sample of six individuals to analyze and compare to the higher-coverage data . The strength of some correlations may depend on how precisely diversity statistics could be estimated , and these estimates likely depend on the amount of sequence analyzed within each window ( or , in other words , the fraction of bases within the 100 kb window that were covered ) . Therefore , we sampled the amount of intergenic sequence to be analyzed in each simulated window from the empirical distribution of the number of bases covered in each window . This was done separately for the low and higher-coverage datasets because the number of bases covered differed between the two datasets . To compute the number of human-chimp differences from the simulations , we compared the sequence of a single chimp individual to a single human individual . Sites where the two individuals were homozygous for different alleles were counted as differences . Sites where both were homozygous for the same allele were not counted as differences . All other sites ( e . g . chimp was heterozygous and human was homozygous , chimp was heterozygous and human was heterozygous , chimp was homozygous and human was heterozygous ) were counted as half a difference . For computational efficiency , we simulated an ancestral population of 500 individuals while keeping the population-scaled mutation and recombination rates and selection coefficients equal to their original values . This approach increased computational efficiency , but should result in the same patterns of variation as larger population sizes since the patterns of variation depend only on the scaled population parameters .
|
While researchers have identified candidate genes that have evolved under positive Darwinian natural selection , less is known about how much of the human genome has been affected by natural selection or whether positive selection has had a greater role at shaping patterns of variation across the human genome than negative selection acting against deleterious mutations . To address these questions , we have combined patterns of genetic variation in three genome-wide resequencing datasets with population genetic models of natural selection . We find that genetic diversity and average minor allele frequency are reduced in regions of the genome with low recombination rate . Additionally , genetic diversity , human-chimp divergence , and average minor allele frequency have been reduced near genes . Overall , while we cannot exclude positive selection at a fraction of mutations , models that include many weakly deleterious mutations throughout the human genome better explain multiple aspects of the genome-wide resequencing data . This work points to negative selection as an important force for shaping patterns of variation and suggests that there are many weakly deleterious mutations at both coding and noncoding sites throughout the human genome . Understanding such mutations will be important for learning about human evolution and the genetic basis of common disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neutral",
"theory",
"population",
"genetics",
"evolutionary",
"selection",
"mutation",
"effective",
"population",
"size",
"genetic",
"polymorphism",
"comparative",
"genomics",
"biology",
"evolutionary",
"theory",
"evolutionary",
"genetics",
"genetic",
"drift",
"adaptation",
"natural",
"selection",
"genetics",
"evolutionary",
"biology",
"evolutionary",
"processes",
"genetics",
"and",
"genomics"
] |
2011
|
Natural Selection Affects Multiple Aspects of Genetic Variation at Putatively Neutral Sites across the Human Genome
|
Stochastic simulation has been a powerful tool for studying the dynamics of gene regulatory networks , particularly in terms of understanding how cell-phenotype stability and fate-transitions are impacted by noisy gene expression . However , gene networks often have dynamics characterized by multiple attractors . Stochastic simulation is often inefficient for such systems , because most of the simulation time is spent waiting for rare , barrier-crossing events to occur . We present a rare-event simulation-based method for computing epigenetic landscapes and phenotype-transitions in metastable gene networks . Our computational pipeline was inspired by studies of metastability and barrier-crossing in protein folding , and provides an automated means of computing and visualizing essential stationary and dynamic information that is generally inaccessible to conventional simulation . Applied to a network model of pluripotency in Embryonic Stem Cells , our simulations revealed rare phenotypes and approximately Markovian transitions among phenotype-states , occurring with a broad range of timescales . The relative probabilities of phenotypes and the transition paths linking pluripotency and differentiation are sensitive to global kinetic parameters governing transcription factor-DNA binding kinetics . Our approach significantly expands the capability of stochastic simulation to investigate gene regulatory network dynamics , which may help guide rational cell reprogramming strategies . Our approach is also generalizable to other types of molecular networks and stochastic dynamics frameworks .
In multicellular organisms , differentiation of pluripotent stem cells into tissue-specific cells was traditionally considered to be an irreversible process . The discovery of cell reprogramming revealed that the identity of a cell is not irreversibly stable , but rather plastic and amenable to control by perturbation of gene regulatory interactions—for example , through over-expression of key transcription factors [1] . Cellular plasticity has also been observed in other contexts , where cells appear to spontaneously transition among phenotypically distinct states . For example , in embryonic stem cells , expression levels of key transcription factors show dynamic heterogeneity , which is thought to enable diversification of the population prior to lineage commitment [2–6] . This heterogeneity may result at least in part from stochastic state-transitions between functionally distinct , metastable subpopulations [4 , 7–9] . Stochastic state-transitions have also been proposed to play a role in cancer , by enabling cancer stem cells to arise de novo from non-stem subpopulations [10] , or by enabling cells to reversibly transition to a drug-tolerant phenotype [11] . In microbial systems , stochastic phenotype switching has been identified as a survival mechanism for populations subjected to fluctuating environments [12 , 13] . Mathematical modeling has provided a basis for understanding how gene regulatory mechanisms and network interactions control cellular identity , stability , and phenotype-transitions . These approaches yield a quantitative means of reinterpreting the long-standing conceptual framework known as Waddington’s epigenetic landscape [14–17] . In a mathematical framework , the “valleys” in the landscape that stabilize cell identities within distinct lineages correspond to attractor basins of a high-dimensional nonlinear dynamical system [18] . The nonlinearity results from positive feedback in transcriptional regulation and epigenetic barriers to chromatin remodeling , for example . These feedback mechanisms give rise to multiple , stable ( or metastable ) phenotype-states accessible to a given genome . Given the “bursty” nature of gene expression and ever-present molecular fluctuations in the cell [19 , 20] , an active area of research is in modeling the effects of so-called intrinsic noise on gene regulatory network ( GRN ) dynamics . These mathematical models support the idea that intrinsic noise can drive stochastic phenotype-transitions [21–25] , which , though likely to be exceedingly rare in general cellular contexts , may explain the heterogeneity observed in embryonic stem cells where epigenetic barriers appear to be lowered [26] . Mathematical models of GRN dynamics that treat stochastic molecular processes are often formulated as probabilistic Master Equations , in which the system evolves probabilistically over a discrete state-space of molecular species and configurations according to a defined set of biochemical reaction rules . Another common framework is that of a coupled system of ODEs describing the expression levels of genes in the network , with the inclusion of additive noise terms . The Master Equation framework is well-suited to studying how “local” stochastic molecular events ( e . g . , transcription factors interacting with DNA or chromatin state-transitions near promoters ) impact “global” dynamics of phenotype stability and state-switching [23–25 , 27 , 28] . These molecular fluctuations affecting promoter activity have been shown to significantly impact the structure of epigenetic landscapes , motivating the use of Master Equation-based approaches . That is , the number and stability of phenotype-states accessible to a given GRN varies depending on the kinetic parameters governing these fluctuations [23 , 24 , 29] . Furthermore , ODE or “mean-field” models that average over these fluctuations can show qualitatively different landscape features [30–32] . Master Equation approaches face the well-known challenge of the “Curse-of-Dimensionality” , as solving them requires enumeration of a state-space that grows exponentially with the number of molecular species in the network . For this reason , discrete stochastic models of GRNs are often studied by stochastic Monte Carlo simulation , via the Gillespie algorithm [33] . However , stochastic simulation can also be problematic: in systems with metastability , such as GRNs , stochastic simulation becomes highly inefficient . Transitions between metastable states are rare events ( i . e . , rare relative to the timescale of fluctuations within a metastable attractor basin ) , and thus difficult or impossible to observe . Often , these rare events are precisely the events of interest , such as in GRNs where infrequent state-transitions represent critical cell-fate transitions . Rare-event sampling algorithms are designed to overcome these challenges , by redirecting computational resources towards events of interest , while maintaining statistical accuracy to global system dynamics [34 , 35] . In this work , we present a rare-event simulation-based method for computing and analyzing epigenetic landscapes of stochastic GRN models . We combine rare-event methods with coarse-graining and analysis by Transition Path Theory—adopted from the field of Molecular Dynamics of protein folding [36]–and show that this unified framework provides an automated approach to map epigenetic landscapes and transition dynamics in complex GRNs . The method quantifies the number of metastable phenotype-states accessible to a GRN , calculates the rates of transitioning among phenotypes , and computes the likely paths by which transitions among phenotypes occur . We apply the method to a model of pluripotency in mouse Embryonic Stem Cells . Our results reveal rare sub-populations and transitions in the network , demonstrate how global landscape structure depends on kinetic parameters , and reveal irreversibility in paths of differentiation and reprogramming . Our approach is not limited to gene regulatory networks; it is generalizable to other stochastic dynamics frameworks and is thus a potentially powerful tool for computing global dynamic landscapes in areas such as signal-transduction , population dynamics , and evolutionary dynamics .
We demonstrate the rare-event sampling method for two representative GRN models . A small , two-gene network serves as a model system to validate the simulations . We then apply the method to a more complex model of pluripotency in mouse Embryonic Stem Cells ( mESCs ) . The mathematical framework of the network models is the discrete Chemical Master Equation ( CME ) [33] , which gives the time-evolution of the probability to observe the system in a given state . In vector-matrix form , the CME can be written d p ( x , t ) d t = K p ( x , t ) ( 1 ) where p ( x , t ) is the probability over the system state-space ( x ) at time t , and K is the reaction rate-matrix containing stochastic reaction propensities ( diagonal elements kjj = −∑i kij , i . e . , columns sum to 0 ) . Eq 1 assumes a well-mixed system of reacting species , and assumes that the technically infinite state-space described by x ( containing molecular species numbers/configurations ) may be limited to some finite number of “reachable” states , ( i . e . , with non-negligible probability ) for an enumeration of N states of the system , K ∈ R N × N . The steady-state probability π ( x ) ≡ p ( x , t → ∞ ) over N states satisfies K π ( x ) = 0 . ( 2 ) Thus , π ( x ) can be obtained from K as the normalized right-eigenvector corresponding to the zero-eigenvalue . It is sometimes desirable to work with the time-dependent stochastic transition-matrix T ( τ ) rather than the time-independent stochastic rate matrix K [42] . For example , T ( τ ) may be more amenable to estimation by sampling ( as we demonstrate in this work for the pluripotency network , for which K is impractical to enumerate ) . For a CME with rate matrix K , T ( τ ) is given by T ( τ ) = exp ( τ K T ) ( 3 ) where exp denotes the matrix exponential . T ( τ ) ∈ R 0 ≤ x ≤ 1 N × N then gives the conditional probability for the system to transition between each pair of states within a lagtime τ . That is , the elements Tij give the probability that the system , if found in state i , will then be found in state j at a time τ later , and rows sum to 1 . Using T ( τ ) , the evolution of probability over discrete intervals of the lagtime τ is given by the Chapman-Kolmogorov equation: p T ( x , t + k τ ) = p T ( x , t ) T k ( τ ) . ( 4 ) Eigenvectors corresponding to dominant eigenvalues of the stochastic transition-matrix are associated with slow system processes . By Perron-Frobenius , for an irreducible stochastic matrix T ( τ ) with eigenvalues λi , there exists λ1 = 1 , and all other eigenvalues satisfy |λi| < 1 . Analogous to Eq ( 2 ) for K , the steady-state probability can be obtained directly from T ( τ ) according to πT ( x ) = πT ( x ) T ( τ ) , i . e . , as the normalized left-eigenvector corresponding to λ1 . Eigenvalues λi are related to global system timescales ti by t i = - τ ln|λ i ( τ ) | , ( 5 ) ( with t1 giving the infinite-time , stationary result ) [42] . Additionally , the Mean First Passage Time for transitions from an individual state i to a region Y ( MFPTi , Y , where Y may be an individual state or a set of states ) can be computed using the matrix elements Ti , j by [43 , 44]: MFPT i , Y = { 0 i ∈ Y 1 + ∑ j ∉ Y T i , j MFPT j , Y i ∉ Y . ( 6 ) MFPTi , Y is defined as the expected time for the system to reach Y for the first time , having started in state i . The MFPTs may be computed by solving the linear system in Eq 6 . Eq 6 computes the MFPT as a dimensionless quantity , the expected number of “steps” ( of duration τ ) required for the transition; multiplication by τ gives the MFPT in units of time . The MFPT starting from a region X ( i . e . , a set of states , rather than an individual state ) and ending in a region Y is given by the stationary-probability-weighted sum: MFPT X , Y = ∑ i ∈ X π i MFPT i , Y ∑ j ∈ X π j . ( 7 ) Stochastic reaction kinetics can be simulated by the Stochastic Simulation Algorithm ( SSA ) [33] , which produces numerically exact realizations of the CME ( Eq 1 ) . Simulation circumvents the need for enumerating the exceedingly large system state-spaces typical of gene network models , but suffers from inefficiency due to rare events . The Weighted Ensemble ( WE ) rare-event sampling algorithm [45] redistributes computational resources from high-probability regions of state-space to low-probability regions , which tend to be under-sampled in conventional simulation . The method thereby reduces computational effort in sampling rare transitions and improves accuracy of estimating probability density in , e . g . , barrier-regions or tails of distributions . The method can be applied to any stochastic dynamics framework; in recent years , it has been widely applied to atom-scale Molecular Dynamics . Details of the methodology are discussed in a recent review [35] and references therein . Both WE and a related method , Forward Flux Sampling , have been applied previously to the study of 2-gene networks [46 , 47] . Briefly , the algorithm works as follows: state-space is divided up into bins that span transitions of interest . The number of bins , Nbins , is typically O ( 100 ) , and a variety of binning procedures can be used ( we use an adaptive procedure described below ) . Initially , a single simulation trajectory , or “replica” , is assigned a weight of 1 and allowed to freely move within and between bins for a user-defined lagtime τWE . After each iteration of τWE , a splitting and culling procedure divides and/or combines replicas and their associated weights in such a way as to reach and maintain an equal target number of weighted replicas , Mtarg , in each bin . Over the course of the simulation , the combined weights of the replicas in a bin ( averaged over successive iterations ) will evolve toward the probability of the system to reside in that bin . By maintaining the same number of replicas in each bin ( Mtarg ) , with weights proportional to probability , the algorithm devotes comparable computational time to low- and high-probability regions . Effectively , the algorithm computes long-time processes on the basis of many short-time simulated trajectories . While the sampled Nbins × Nbins transition-matrix provides a global approximation of the epigenetic landscape and state-transitions , we apply a method to further coarse-grain dynamics , known as the Markov State Model framework [29 , 36 , 42] . This automated procedure produces a highly simplified representation of global dynamics in terms of a few ( generally < 10 ) clustered sets and the transitions among them . Such highly-reduced models can be beneficial in terms of human intuition of system dynamics , comparison to experiments , and—in this application—automated designation of dynamic phenotype-states . The method utilizes the concept of metastability , i . e . , system states that experience relatively fast transitions among them are clustered together into the same coarse-grained set . Collectively , the coarse sets experience relatively rare inter-cluster transitions and frequent intra-cluster transitions . We employ the metastability concept as a definition of cell phenotype , reasoning that a phenotype should be a relatively stable attribute of a cell , and stochastic inter-phenotype transitions should be relatively rare . In practice , we employ the Markov State Model framework to further reduce the sampled row-stochastic transition-matrix T ˜ ( τ ) from size Nbins × Nbins down to C × C , where C is the number of coarse-grained clusters chosen by the user . As the Markov State Model ( MSM ) is itself a stochastic transition-matrix on a coarse-grained space , it implies a more severe Markovian approximation . It provides a way to describe global system dynamics in a highly simplified way while maintaining high accuracy to the slowest system dynamics as sampled by T ˜ ( τ ) . In previous work , we demonstrated the application of this coarse-graining approach to automatically designate phenotypes in small gene networks [29]; here , we extend the applicability of the coarse-graining to large , complex networks by combining it with rare-event sampling . The coarse-graining procedure is a spectral clustering method based on the Perron Cluster Cluster Analysis ( PCCA+ ) algorithm [53] , which optimizes the ( nearly ) -block-diagonal structure of T ˜ ( τ ) for systems with metastability . The signature of such metastability is a separation-of-timescales for intra- and inter-basin dynamics , which may be seen as gaps in the eigenvalue spectrum [42] . As noted above , T ( τ ) ( or its sampled counterpart , T ˜ ( τ ) ) has λ1 = 1 , corresponding to the infinite time-limit . If a set of m dominant eigenvalues exists , such that for decreasing eigenvalues λi ⪅ 1 , i ∈ {2 , … , m} , and a gap is present , λj ≪ λm for j > m , this indicates the presence of m slow-timescale processes in the system , and further indicates that T ˜ ( τ ) may be re-ordered to give m nearly-uncoupled blocks . In practice , the algorithm attempts to find a coarse-graining onto C clusters , where C may be user-defined , or may be determined algorithmically , e . g . , according to the spectral gap [53] . Here , we choose C clusters , where the last significant gap in the spectrum is seen between λC and λC+1 . For the GRNs studied here , this corresponds to choosing C such that λC/λC+1 > 10 . Both the sampled transition-matrix T ˜ ( τ ) and the coarse-grained MSM encode stationary and dynamic information about global dynamics—that is , they quantify the epigenetic landscape . For visualization , we use Gephi graph visualization software [54] using the Force Atlas algorithm . Every circle ( or node ) in the graph corresponds to a sampling bin or to a coarse-grained phenotype , and the area of a circle is proportional to its relative steady state probability according to ln ( γPSS ) , where PSS is the steady state probability of the node and γ is a constant chosen to improve visibility of low probability regions of the landscape . Lines between circles ( edges ) correspond to transitions between sampling regions or coarse-grained phenotype . Their thickness and coloring correspond to their relative transition probability and source state , respectively . To validate the simulation method , we compare the simulated dynamics to the numerical solution to the CME . We choose the parameters of the ExMISA model in such a way as to restrict the effective state-space , so that a numerical solution of the CME is tractable . Building the reaction rate matrix K ∈ R N × N requires enumeration of N system states . In general , if a system of S molecular species has a maximum copy number per species of nmax , then N ≈ n m a x S . In the ExMISA model , the state-vector is given by x = [Aij , Bij , na , nb] . For enumeration , we neglect states with protein copy-numbers larger than a cutoff value which exceeds g10/k ( corresponding to the average number of transcription factors maintained in the system from a gene while in its active state ) . For example , with model parameters g10 = 18 and k = 1 , we truncate at na , max = nb , max = 41 and assume that probability flux between states with na , nb ≤ 41 and states with na , nb > 41 is assumed to be 0 ( i . e . , the boundaries of the state-space are reflective ) . Including the gene-binding states , this gives N = 3 × 3 × 42 × 42 = 15876 states . This size is tractable for complete solution of the CME using matrix methods in MATLAB [55] . This truncation of the state-space introduces a small approximation error ( see S2 Fig ) . The pluripotency network has 8 genes with copy numbers of O ( 10 3 ) ( determined by the parameters gon/k = 3900 ) . The number of distinct binding-promoter states for each gene are 16 , 32 , 8 , 8 , 2 , 8 , 4 , and 2 for GATA6 , NANOG , CDX2 , OCT4 , SOX2 , KLF4 , GCNF , PBX1 , respectively ( see S2 Table ) . Together these combinations enumerate a state-space of N > 1030 ≈ 10008 × 16 × 32 × 8 × 8 × 2 × 8 × 4 × 2 . This size precludes solution of the CME , and we instead estimate the dynamics by WE sampling . Where possible , we validate the WE-sampling results by “conventional” , i . e . , by direct simulation using SSA . Stochastic Gillespie ( SSA ) simulations were carried out using BioNetGen [56] . WE sampling was implemented with in-house software code written in MATLAB . Simulations were run on the high performance computing cluster ( HPC ) at the University of California , Irvine , and parallelization of BioNetGen SSA simulations was performed using the Sun Grid Engine scheduler . The coarse-graining procedure and transition path analysis was implemented in python scripts , adapted from MSMBuilder [57] and Pyemma [43] , respectively . Transition-matrix and MSM visualization was carried out using Gephi software and the Force Atlas layout [54] . All simulation parameters can be found in the supplement S4 Table . Pseudo-code for the adaptive binning procedure can be found in S2 File and software can be found in https://github . com/Read-Lab-UCI/Rare-Event-Sampling-Gene-Networks .
We first apply the computational pipeline to a small two-gene model ( the exclusive Mutual Inhibition , Self-Activation model , ExMISA , see Methods ) , exhibiting an archetypal motif for cell fate-decisions [37 , 38] . The model is tractable for computation of full , discrete stochastic dynamics to within a small approximation error using matrix methods . Thus , the model provides a numerical benchmark for assessing the accuracy of the simulation method , before extension to larger systems where solution of the Chemical Master Equation ( CME ) is intractable . For the chosen parameters , the ExMISA model shows four peaks in the steady-state probability distribution ( projected onto protein copy numbers , na and nb ) . Peaks in probability correspond to basins in the so-called quasipotential landscape , defined by U = −ln ( π ( x ) ) ( Fig 2 ) . The four peaks/basins corresponds to four possible combinations of binarized A/B gene expression: hi/hi , hi/lo , lo/hi , and lo/lo . These four phenotype-states arise due to the combination of balanced repression and self-activation in the network , and the slow kinetic parameters ( Supplementary S1 Table ) for transcription factor binding and unbinding to promoters that effect changes in individual gene-activity states between low and high expression rates [29 , 58] . The WE-based simulation method enabled estimation of global dynamics of the ExMISA model . By redistributing computational resources from relatively high-probability to low-probability regions ( see Methods ) , the WE method enabled uniform sampling of the quasipotential landscape , i . e . , mapping basins ( high-probability regions ) along with high barriers ( low probability regions ) ( Fig 2a ) . The simulation estimated individual steady-state bin-probabilities as low as 1 . 3 × 10−6 and showed good global agreement with the numerical CME benchmark ( see Fig 2 and Supplement , S3 Fig ) . In addition to sampling global dynamics , the WE method can be used to estimate rate constants for individual , rare transitions of interest . The Mean First Passage Time of the global network switch from the center of one polarized phenotype-state to another , i . e . , MFPTX→Y from protein a/b expression level hi/lo to lo/hi was estimated from WE to be 1 . 82 × 105 ( units of k−1 ) ( see S6 Table ) , in agreement with the CME result . A network transition-matrix T ˜ ( τ ) over sampled bins ( Nbins = 300 ) was constructed from WE sampling for ExMISA and used for subsequent analysis of global system dynamics . By comparison , a full network transition-matrix T ( τ ) over the enumerated system state-space was constructed from the CME ( N = 15876 , see Methods ) . The full , computed ( T ( τ ) ) and simulated ( T ˜ ( τ ) ) transition-matrices showed qualitatively similar eigenvalue spectra with four dominant eigenvalues , indicating the presence of metastability ( separation-of-timescales between intra-basin and inter-basin transitions ) ( Fig 2b ) . The slow system-timescales predicted by the full CME model corresponding to eigenvalues λ2 , λ3 , λ4 were t2 , t3 , t4 = 6 . 8 × 104 , 4 . 2 × 104 , 1 . 0 × 104 respectively , in units of k−1 where k is the protein degradation rate ( the Perron eigenvalue λ1 = 1 is associated with the infinite-time ( stationary ) distribution ) . The corresponding values given by the WE-simulated T ˜ ( τ ) were 6 . 1 × 104 , 3 . 5 × 104 , 9 . 4 × 103 , respectively . These numbers demonstrate how the sampled T ˜ ( τ ) enables global approximation of slow system timescales to < 20% relative error . Error in these values ( relative to the slowest timescales implied by the true eigenvalues ) depends on both “spectral” ( lagtime ) and discretization error , i . e . , improvements can be achieved only with a larger number of bins ( finer discretization ) and/or longer lagtime [42] ( see S4 Fig ) . In contrast , WE sampling in “rate mode” ( see Methods ) enabled highly accurate estimation of MFPTX→Y to within 2% error ( S6 Table ) . According to the Markov State Model framework , the presence of timescale separation indicates that a simplified model , retaining a few coarse-grained metastable states with Markovian transitions among them , can reasonably approximate the full system dynamics . Using this approach , we label the metastable sets as phenotypes accessible to the network , reasoning that a useful classification of cell phenotypes should be one that gives relatively stable , rather than transient , cell types . We apply the Markov State Model coarse-graining procedure to both the full T ( τ ) and simulated T ˜ ( τ ) , yielding similar results . The coarse sets ( or metastable phenotype-states ) in the reduced models for both cases are generated automatically , and map directly onto the four basins seen in the quasipotential landscape ( i . e . , the gene A/B expression hi/hi , hi/lo , lo/hi , and lo/lo cell phenotypes ) . The reduced models are visualized by network graphs , in which node sizes are proportional to steady-state probability , and the thicknesses and lengths of edges are proportional to the transition probability between them ( on lagtime τ ) ( Fig 2c ) . Some discrepancies can be seen visually in the network graphs . These discrepancies likely result in part from the slightly different mappings of the full state-space onto the four clusters ( see S11 Fig for details ) , which could in turn result from the distance-metric-based binning , which is relatively insensitive to changes in promoter configuration . Numerical values for the reduced models can be found in S5 Table . The network graph can be considered to be an alternative representation of the global epigenetic landscape , which contains both stationary and dynamic information . ( In contrast , the epigenetic landscape plotted as a quasipotential function does not explicitly contain dynamic information , due to non-gradient dynamics [16] ) . Validation of the coarse-grained model can be carried out according to the Chapman-Kolmogorov test [42] , which tests how well the relaxation dynamics initialized in the metastable phenotypes approximate the dynamics that are predicted either by the full model ( CME ) or simulated trajectories . According to this test , relaxation dynamics out of metastable phenotypes from WE sampling was predicted with relative error values between 0 . 02 and 0 . 12 for all phenotypes ( S5 Fig ) . Together , these results indicate ( i ) that a Markovian model of phenotype transitions is a good approximation of the full system dynamics for the ExMISA model , and ( ii ) that the WE-simulation based computational pipeline predicts a quantitatively similar coarse-grained phenotype-network to the full CME model . We apply the computational pipeline to a pluripotent fate-decision network from mouse Embryonic Stem Cells ( mESCs ) introduced by Zhang et al . [28] ( Fig 3A ) . The network comprises eight interacting genes: NANOG , GATA6 , CDX2 , SOX2 , OCT4 , GCNF , and PBX1 . Three of these genes , NANOG , SOX2 , and OCT4 have been suggested to maintain pluripotency [59] , and NANOG inhibits the expression of differentiation markers [60] . The GATA6 and CDX2 genes have been used in experiments as markers of differentiation , with the GATA6 transcription factor being a marker of the primitive endoderm cell lineage , and the CDX2 transcription factor being a marker of the trophectoderm lineage [61] . Using the WE-based computational pipeline , we estimate T ˜ ( τ ) with a resolution of Nbins = 250 . To visualize the global landscape as a graph network at this resolution , we plot the converged T ˜ ( τ ) using a force-directed automated graph layout [54] ( Fig 3B ) . The barbell shape of the network reflects the broad antagonism between pluripotency and differentiation genes , which is a general feature of the overall network topology . At the same time , each “pole” comprises multiple distinct patterns of gene expression ( seen in the graph as different colors with full compositions in Fig 3C ) , hinting at the existence of multiple phenotypes associated with both pluripotency and lineage-specification . Moreover , the network representation reveals numerous links between pluripotent and differentiated states , pointing to both direct and indirect transitions , through a network of relatively transient intermediate states . To further analyze the global dynamics of the pluripotency network , we apply the Markov State Model coarse-graining framework . The simulated T ˜ ( τ ) shows gaps in the eigenvalue spectrum after four and after six eigenvalues ( Fig 4a ) . The corresponding approximate timescales are given by t2 , t3 , t4 , t5 , t6 = 1 . 1 × 105 , 95 , 51 , 12 , 12 ( k−1 ) , respectively . These values , though only approximate , indicate the presence of a single long timescale process ( t2 ) corresponding to transfer between differentiated and pluripotent states , while transitions within those basins ( t3 , etc . ) occur at least four orders of magnitude more quickly . Applying the coarse-graining algorithm to achieve six clusters results in a reduced model ( Fig 4b ) , with the clusters representing metastable phenotypes . The phenotypes can largely be distinguished in the subspace of NANOG , GATA6 , and CDX2 expression levels; the differentiated phenotypes show expression of either GATA6 ( primitive endoderm , PE ) , CDX2 ( trophectoderm , TE ) , or both ( denoted an intermediate cell type , IM ) . Phenotypes associated with pluripotency do not express high levels of GATA6 or CDX2 , and may express high levels of NANOG ( stem cell , SC ) . The coarse-grained model reveals two separate pluripotent phenotypes that are low in NANOG expression: one which expresses other pluripotent factors OCT4 , SOX2 , and KLF4 ( “Low NANOG 1” LN1 ) , and one which has low expression of all factors ( “Low NANOG 2” LN2 ) ( Fig 4c ) . Overall , these phenotypes broadly match experimentally-determined categories , coincide with steady-states of the stochastic model computed previously by a CME-approximation method [28] , and coincide with phenotype-states identified in related pluripotency GRN models [62] . The steady-state probabilities associated with the phenotypes are highly nonuniform , with 95% of the population divided nearly evenly between the IM and LN1 phenotypes , which are associated with differentiation and pluripotency , respectively . The LN2 state is rarest , comprising only 8 × 10−4% of the population , and was not identified previously [28] . Together , these results indicate that the clustering method identifies both common and exceedingly rare phenotypes in the in silico cell population modeled by simulation trajectories . Furthermore , the automated method identifies both expected phenotypes and one novel ( albeit low probability ) phenotype . Previously , Markov State Models constructed on the basis of Molecular Dynamics simulations were used to analyze the ensemble of distinct pathways of protein-folding [36] . Here , we utilize the coarse-grained model of phenotype transitions in the pluripotency GRN in a similar manner , to analyze pathways of cell differentiation and dedifferentiation . Using Transition Path Theory , the method identifies the pathways that carry the greatest fraction of net probability flux , among sequences associated with successful SC→TE transitions ( and reverse ) ( Fig 4d and 4e ) . Transition paths between the stem cell ( SC ) and PE phenotypes can be found in S6 Fig . For Parameter Set I , the method identifies three pathways encompassing > 98% of the probability flux for both forward and reverse transitions . While the SC→ TE transition is most likely to occur directly through the LN1 state ( i . e . , NANOG expression will shut off , followed by turning on CDX2 ) , the reverse transition shows a different route through the IM and PE states ( i . e . , GATA6 expression turns on , then CDX2 turns off , then GATA6 turns off , and finally NANOG turns on ) . Dynamic analysis of the coarse-grained model , including analysis of transition paths , relies on the Markovian approximation for inter-phenotype transitions . In the pluripotency network , stochastic transitions between pluripotency ( SC , LN1 , LN2 ) and differentiation ( TE , IM , PE ) basins are infrequent relative to transitions within those basins , justifying the Markovian assumption , since the system equilibrates within those basins much more rapidly than inter-basin transitions occur . However , the Markovian assumption may be less accurate for describing intra-basin transitions between phenotypes , which occur much more frequently . Despite the coarse-grained model encompassing transitions on highly disparate timescales , the qualitative results of transition path analysis were validated by collected conventional simulation trajectories ( not subject to any Markovian assumption ) , which identified the same dominant transition paths ( S7 Fig ) . Overall , these results indicate that a stochastic excursion of a cell from the SC to TE phenotypes and back maps a cycle in gene-expression space , echoing previous studies indicating nonequilibrium dynamics in GRNs [16 , 23] . The results further indicate that the Markov State Model , while a highly coarse-grained approximation , can provide an accurate estimation of inter-phenotype transition dynamics . We applied the computational pipeline to the pluripotency network using two different rate parameters sets ( see S1 File ) , which differ in rates of transcription factor binding and unbinding to DNA . In line with previous studies [23 , 24 , 29] , we found that increasing the so-called adiabaticity ( i . e . , increasing h and f , or the rates of TF-binding relative to protein production and degradation , Parameter Set II ) led generally to rarer inter-phenotype transitions ( see Table 1 ) . For example , in Parameter Set I , the Mean First Passage Time ( MFPT ) for transitions from SC → TE was calculated to be 1 . 36 × 105 in units of k−1 , as compared to 8 . 13 × 108 for Parameter Set II . The MFPTs of the reverse transition TE → SC for each set were 2 . 70 × 105 and 5 . 82 × 109 , respectively ( see Table 1 and S7 Table ) . These differences in magnitude broadly reflect that moving toward the adiabatic regime leads to increased epigenetic barriers between phenotypes . In addition to generally slowing transitions , the increased adiabaticity of Parameter Set II gives rise to an epigenetic landscape structure that is distinct from that of Parameter Set I , with altered steady-state phenotype probabilities ( Fig 5a ) . The eigenvalue spectrum shows qualitatively distinct features as well , with a gap after five values ( Fig 6a ) . As such , the Markov State Model framework identifies five dominant phenotypes in the network , which correspond broadly to those of Parameter Set I , except that only a single Low-NANOG ( LN ) phenotype is identified ( Fig 6b ) . Most of the steady-state probability is contained in the IM state ( Fig 6c ) . In addition to altering the transition rates and relative phenotype probabilities , the kinetic parameters altered the dynamics of differentiation and dedifferentiation . The two likeliest pathways of forward ( and reverse ) SC → TE transitions follow the same route through LN and IM phenotypes ( Fig 6d and 6e ) . Alternative differentiation pathways of forward ( and reverse ) SC → PE transitions can be found in S9 Fig . These results indicate that , while the same GRN model with different kinetic parameters may give rise to qualitatively similar phenotypes , they differ in quantitative stationary and dynamic features , including relative steady-state probabilities , transition times , and likeliest transition pathways . Rare phenotype transitions can be difficult to observe with conventional SSA simulation . We compared simulated landscapes ( based on estimated T ˜ ( τ ) ) from the computational pipeline for the Pluripotency network ( Parameter Set II ) to those obtained from an equivalent ( large ) number of SSA simulation steps ( Fig 5a and 5b ) . This comparison revealed that the WE-based method uncovers multiple phenotypes and associated transitions that are effectively invisible to conventional simulation due to the rarity of exiting metastable basins . Quantitative estimates of efficiency gains for WE have been based on comparing the number of simulation steps required to estimate a desired quantity ( such as a rate constant ) using WE versus conventional simulation [47] . Treating T ˜ ( τ ) as the desired output ( as it contains holistic dynamic information for the system ) , we estimate the efficiency gain of our pipeline by computing: E = Sim . steps to estimate T ˜ ( τ ) , Conv . Sim . steps to estimate T ˜ ( τ ) , WE . ( 10 ) The denominator of Eq 10 is given by Nbins × Niterations × τ × Mtarg , thus accounting for all individual replica-steps in the total WE simulation time . The numerator is computed by asking how many steps of a conventional simulated trajectory are required to estimate T ˜ ( τ ) . It is generally prohibitive to collect enough conventional simulation steps to estimate T ˜ ( τ ) to a similar resolution as WE . However , given a T ˜ ( τ ) estimated from WE , it is in principle possible to estimate how many steps would be necessary to achieve the same T ˜ ( τ ) by conventional simulation . We used an approximate , conservative estimate given by: [ Sim . steps to estimateT ˜ ( τ ) , Conv . ] ⪆ τ ∑ i ( P 5 % , i { T i j } ) - 1 , ( 11 ) where P5% , i denotes the 5th percentile over nonzero elements of row i . Justification of Eq 11 is given in the Supplement , S3 File . Briefly , Eq 11 reflects the fact that the required simulation time should be dominated by the rare transitions ( i . e . , the smaller elements of T ˜ ( τ ) ) , while attempting to avoid over-dependence on individual estimates of small Tij , which generally have unknown error . The error versus simulation time in WE- and Conv . -estimated T ˜ ( τ ) are plotted in S12 Fig . According to Eq 11 , we estimate that our pipeline provided efficiency gains of 2 for ExMISA ( Fig 2 ) , 900 for Pluripotency Parameter Set I ( Fig 3 ) , and 1 × 106 for Parameter Set II ( Fig 6 ) . These numbers show that the pipeline can afford a significant speedup over conventional simulation in providing global dynamic information . The numbers further show that the efficiency gain is most pronounced for the Pluripotency network with exceedingly rare inter-phenotype transitions .
We used the pluripotency network as a model system to develop and demonstrate the simulation approach , but the results also yielded biological insights . For example , the simulations revealed a hierarchical structure of the epigenetic landscape . The network—exhibiting 5-6 metastable phenotypes—occupies a limited subspace from the vast possible gene combinations ( e . g . , 28 = 256 possible distinct on/off combinations of gene expression states ) . The dominant feature of the global landscape is a high barrier/slow timescale between pluripotent and differentiated phenotypes . Within each of these categories , further sub-states were identified . The model revealed multi-timescale dynamics of phenotype transitions; the pluripotency network showed relatively rapid transitions between phenotype-states that differed in the expression-level ( high vs . low ) of a single gene , e . g . the high NANOG to low NANOG transition , whereas phenotype transitions involving a change in expression level of seven genes , e . g . the SC macrostate to the TE macrostate , occurred five orders of magnitude more slowly on average . While the accessible phenotypes appear broadly similar across parameter sets , the relative stability and transition dynamics among phenotypes were sensitive to kinetic parameters governing transcription factor binding/unbinding . A global change in these parameters ( affecting all individual transcription factor-DNA interactions equally ) changed the shape of the landscape , altering the relative steady-state probabilities of different phenotypes and the likely transition pathways linking them . The DNA binding parameters capture the local epigenetic mechanisms that enable/disable transcription factors from accessing regulatory elements . A global rate change nevertheless has a varying influence on different genes because the number of regulators differs , as does the molecular logic by which activators and repressors exert combinatorial control on different genes . These results echo findings that global modification of chromatin regulators often have lineage-specific effects [63] . These results highlight both the need for , and the challenge of , informing cell reprogramming strategies with quantitative network models , as they suggest that the dynamic response of cellular networks to perturbations is governed by the detailed kinetics of molecular regulatory mechanisms , which are generally difficult to parameterize . The Markov State Model framework implicitly imposes a dynamic definition of cell phenotypes; the number of phenotypes was determined using spectral gap-analysis , and the coarse-graining algorithm automatically identified metastable aggregates ( i . e . , grouped sampled network states into larger clusters ) . This is different from the classifications of phenotypes that are generally used in analyzing experimental data , where gene expression or marker levels are often used to categorize cells . However , experiments have also revealed the potential need for a dynamic definition of cell phenotype , based not only on single-timepoint measurements of gene expression or phenotype-markers , but also on information from past or future timepoints [4 , 8] . For example , Filipczyk et al . [8] identified distinct subpopulations within a compartment of NANOG-negative cells in mESCS , which differed in their propensity to re-express NANOG . At the same time , fluctuations between low- and high-NANOG expressing cells were not necessarily associated with any functional state change . The Markov State Model approach , based on kinetic/dynamic coarse-graining , thus provides a quantitative approach for classifying phenotype-states that is both completely generalizable rather than ad hoc ( it requires no a priori knowledge or designation of markers/genes ) and is in line with these recent experiments revealing the need for a dynamic definition of phenotype . Markovian transitions ( i . e . , memoryless “hops” ) among cell phenotypes have been observed experimentally: examples include transitions among phenotypes in cancer cells , as measured by flow cytometry [10] , and among pluripotency-states in mESCs , as measured by time-lapse microscopy of fluctuating gene expression [7–9] . The compact nature of these data-inferred networks—showing hops among a limited set of broad phenotypes—suggests that the computed MSM framework advanced in this study provides an appropriate level of resolution at which to analyze GRN dynamics and may serve as a useful tool for comparing models to experimental data . Experimental studies have quantified the timescales of Markovian transitions between NANOG-high and NANOG-low states in mESCs [8 , 9] . From Hormoz et al . , the probability of transitioning from NANOG-high to NANOG-low in mESCs is 0 . 02 per cell cycle , while that of the reverse transition is 0 . 08 . These values represent a relatively rapid transition rate , since NANOG expression is known to be particularly dynamic [60] . Similarly , plasticity has been observed in cancer cells where quantitative estimates of stochastic cell transitions between a stem cell cancer cell phenotype to a basal cancer cell phenotype were observed to be roughly on the order of 0 . 01 to 0 . 1 per cell cycle [10] . We can translate our model results to approximate biological timescales: the degradation rate , which sets the timeunit for model results ( i . e . , k is taken to be 1 ) was experimentally determined to be on the order of a few hours ( in the E14 mouse embryonic stem cell line , the half-lives of NANOG , OCT4 , and SOX2 are approximately 4 . 7 , > 6 , and 1 . 6 hours , respectively [64] ) . Assuming that degradation is unimolecular , k = ln ( 2 ) /t[NANOG]1/2 , and the half-life of NANOG , t[NANOG]1/2 = 5 hours , the degradation rate is k = 0 . 1 . Using a mESC cell cycle time of 12 hours [65] , the simulations for Parameter Set I then predict NANOG-high to NANOG-low transitions occurring with a rate of 0 . 03 per cell cycle , and of 3 × 10−3 for the reverse . For Parameter Set II , the computed rates were 8 × 10−6 and 5 × 10−5 , respectively . Comparison of these computed and experimental rates of NANOG transitions indicates that Parameter Set I ( f = 10 ) is more in line with experimental observations , while Parameter Set II ( f = 50 ) gives transition rates that are three orders of magnitude too slow . These results are in agreement with previous findings from theoretical studies that GRNs in pluripotency networks operate in a so-called “weakly-adiabatic” regime [24 , 27 , 28] , in which the timescale of DNA-binding by transcription factors is on the order of transcription factor production and degradation . A number of theoretical studies have elucidated dynamics of stochastic molecular-detailed GRN models ( i . e . , models that include molecular fluctuations and regulatory mechanisms , in contrast to Boolean models [66] ) . These studies have largely focused on small 1- or 2-gene motifs[[21–25 , 32 , 39]] . In the limit of slow DNA-binding/unbinding , it was shown that the stationary distribution of the stochastic model can be solved exactly [41] . Recent years have seen extension of stochastic methods to studies of more complex , experimentally derived GRN models encompassing O ( 10 ) genes . For example , determination of global dynamic properties of such networks has been achieved by combining information from long stochastic simulations of discrete models [27 , 62] , or of continuum SDE models , in combination with path integral approaches [58 , 67] . The pluripotency network studied herein was developed by Zhang and Wolynes [28]; in their work , the authors developed a continuum approximation to the Chemical Master Equation that enabled quantitative construction of the epigenetic landscape . Here , we present an alternative approach that is unique in two major aspects: ( 1 ) the use of stochastic simulations ( i . e . , SSA [33] ) , which is enabled by use of the WE rare-event sampling algorithm , and ( 2 ) the automated Markov State Model framework for designating phenotypes and constructing a coarse-grained view of the epigenetic landscape . While we utilize a different framework ( that of coarse-grained , discrete stochastic models ) from Zhang and Wolynes to approximate and interpret dynamics , our results are broadly consistent with theirs . For example , the dominant identified phenotypes we found are the same as in their work ( the only exception being the exceedingly rare LN2 phenotype identified by the coarse-graining algorithm for Parameter Set I ) . Our approach is uniquely suited to extracting global dynamics information for stochastic systems with metastability , using simulations . An advantage of this approach is that both the WE and coarse-graining algorithms are“dynamics-agnostic” [47] , meaning that they can be applied to any type of stochastic dynamics framework . In the context of computational biology , our pipeline could be extended to other types of stochastic biochemical systems , such as systems with hybrid discrete-continuum dynamics [68] , systems with spatial heterogeneity [69] , or multi-level models [70] . In addition to this flexibility , simulation-based methods have the advantage of being able to leverage existing , widely-used open-source packages , which in turn facilitate model specification and model sharing . For example , BioNetGen [56] can interpret models specified in the Systems Biology Markup Language [71] . Several challenges and potential weaknesses with the pipeline exist , both with regard to sampling rare events , and in determining an appropriate coarse-grained model . Potential challenges with the WE algorithm itself have been described elsewhere [35 , 69] , and include the difficulty of determining a binning that captures slow degrees of freedom and the existence of time-correlations between sampled iterations of the simulation , which can impede unbiased sampling . The Voronoi-based binning procedure we employ here is related to a number of similar approaches [24 , 48–50] , and has the advantage of effectively tiling a high-dimensional space without the need for a priori knowledge . However , in practice , according to others and our own studies , the method is effective up to about 10 degrees of freedom . Therefore , in larger gene networks ( as in other complex systems ) an ongoing challenge will be to identify optimal binning methods to effectively partition slow degrees of freedom and thus enable efficient enhanced sampling . New adaptive partitioning methods could also have the effect of improving the accuracy of coarse-grained Markov models , as finer partitioning of transition regions has been found to reduce errors in the Markovian approximation [42] . Additional improvements to efficiency , which could aid in scaling the method to larger networks , could be achieved in the future by using alternatives to the direct SSA algorithm ( see e . g . , [72] ) or improved parallelization techniques .
|
Cell phenotypes are controlled by complex interactions between genes , proteins , and other molecules within a cell , along with signals from the cell’s environment . Gene regulatory networks ( GRNs ) describe these interactions mathematically . In principle , a GRN model can produce a map of possible cell phenotypes and phenotype-transitions , potentially informing experimental strategies for controlling cell phenotypes . Such a map could have a profound impact on many medical fields , ranging from stem cell therapies to wound healing . However , analytical solution of GRN models is virtually impossible , except for the smallest networks . Instead , time course trajectories of GRN dynamics can be simulated using specialized algorithms . However , these methods suffer from the difficulty of studying rare events , such as the spontaneous transitions between cell phenotypes that can occur in Embryonic Stem Cells or cancer cells . In this paper , we present a method to expand current stochastic simulation algorithms for the sampling of rare phenotypes and phenotype-transitions . The output of the computational pipeline is a simplified network of a few stable phenotypes , linked by potential transitions with quantified probabilities . This simplified network gives an intuitive representation of cell phenotype-transition dynamics , which could be useful for understanding how molecular processes impact cellular responses and aid interpretation of experimental data .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"markov",
"models",
"gene",
"regulation",
"regulatory",
"proteins",
"dna-binding",
"proteins",
"simulation",
"and",
"modeling",
"mathematics",
"stem",
"cells",
"network",
"analysis",
"transcription",
"factors",
"epigenetics",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"cell",
"potency",
"animal",
"cells",
"proteins",
"gene",
"expression",
"pluripotency",
"probability",
"theory",
"biochemistry",
"biochemical",
"simulations",
"cell",
"biology",
"phenotypes",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"computational",
"biology"
] |
2018
|
Rare-event sampling of epigenetic landscapes and phenotype transitions
|
Praziquantel ( PZQ ) is the only widely available drug to treat schistosomiasis . Given the potential for drug resistance , it is prudent to search for novel therapeutics . Identification of anti-schistosomal chemicals has traditionally relied on phenotypic ( whole organism ) screening with adult worms in vitro and/or animal models of disease—tools that limit automation and throughput with modern microtiter plate-formatted compound libraries . A partially automated , three-component phenotypic screen workflow is presented that utilizes at its apex the schistosomular stage of the parasite adapted to a 96-well plate format with a throughput of 640 compounds per month . Hits that arise are subsequently screened in vitro against adult parasites and finally for efficacy in a murine model of disease . Two GO/NO GO criteria filters in the workflow prioritize hit compounds for tests in the animal disease model in accordance with a target drug profile that demands short-course oral therapy . The screen workflow was inaugurated with 2 , 160 chemically diverse natural and synthetic compounds , of which 821 are drugs already approved for human use . This affords a unique starting point to ‘reposition’ ( re-profile ) drugs as anti-schistosomals with potential savings in development timelines and costs . Multiple and dynamic phenotypes could be categorized for schistosomula and adults in vitro , and a diverse set of ‘hit’ drugs and chemistries were identified , including anti-schistosomals , anthelmintics , antibiotics , and neuromodulators . Of those hits prioritized for tests in the animal disease model , a number of leads were identified , one of which compares reasonably well with PZQ in significantly decreasing worm and egg burdens , and disease-associated pathology . Data arising from the three components of the screen are posted online as a community resource . To accelerate the identification of novel anti-schistosomals , we have developed a partially automated screen workflow that interfaces schistosomula with microtiter plate-formatted compound libraries . The workflow has identified various compounds and drugs as hits in vitro and leads , with the prescribed oral efficacy , in vivo . Efforts to improve throughput , automation , and rigor of the screening workflow are ongoing .
Treatment and control of the flatworm disease , schistosomiasis , relies on a single drug , praziquantel ( PZQ ) . Since the first clinical trials in the late 1970's [1] , PZQ has proven safe and effective against all three major forms of the disease , and today , declining costs make the drug more affordable , currently at around 7–19 US cents per 600 mg tablet [2] . A single oral dose of 40–60 mg/kg is sufficient to achieve cure rates of 60–90% [3] while facilitating patient compliance , especially among children . Clinically relevant and widespread resistance , despite occasional and isolated incidences [4] , has yet to occur . This fortuitous situation stands in contrast to the situation for some other ‘neglected tropical diseases’ , ( NTDs; [5] ) for which antiquated and often toxic drugs must be parenterally administered over a number of days or weeks and which increasingly have problems associated with drug resistance [6] . Thankfully , concerted pharmaceutical discovery efforts via ‘public-private partnership’ ( PPP ) consortia [7] , [8] are ongoing to address this desperate situation and robust ‘drug pipelines’ have been established which , hopefully , should yield new therapies over the next ten to 15 years . All of this recent activity has bypassed schistosomiasis , due in part to the tremendous success of PZQ . Yet , reliance on a single drug to treat a population of over 200 million people infected and over 700 million people at risk over three continents [9] seems particularly perilous when considering the threat of drug resistance . Also , PZQ is not without problems . Principal among these is its relative inactivity against migratory juvenile and sub-adult worms [10] , [11] meaning that , for effective treatment and sustainable control , PZQ must be given on a regular basis . Thus , recent discussions , as part of treatment landscape for human helminthiases in general [12] , have focused on reawakening the need to search for alternatives to PZQ , including the development of combinations of drugs incorporating PZQ [13] , [14] . The latter option , if more difficult and costly to develop , has the longer term benefit of extending the availability of PZQ while hindering the onset of resistance to this most valuable of drugs . Only to a limited extent has the underlying rationale for inquiry of anti-schistosomal compounds ( and anthelmintics in general ) involved detailed knowledge of the molecular drug target or mechanism of action although there are some notable advances , e . g . , inhibition of redox [15] and proteolytic enzymes [16] , and heme aggregation [17] . The relative lack of validated molecular drug targets for this parasite is in stark contrast to those underpinning entire drug development portfolios of PPPs tackling other infectious diseases of global import such as malaria and the trypanosomiases [7] , [8] . Hopefully , this paucity of targets can be better addressed with the recent availability of the draft genomes of both Schistosoma mansoni [18] and S . japonicum , and the first attempts to prioritize those targets ( [19]; TDR Drug Targets Prioritization Database [20] ) . More common in schistosome drug discovery has been the complementary approach of phenotypic ( whole organism ) screening in vitro ( usually with adult worms ) and/or animal models of disease to measure compound efficacy [21] . These strategies are usually without specific knowledge of the target and/or mechanism of action ( e . g . , [22] , [23] ) , or for which bioactivity has been characterized in other parasitological or biomedical settings [24] , [25] , [26] . They are of proven value . For example , PZQ was first developed as a veterinary cestocide before being tested in an animal model of schistosomiasis [27] and long before data regarding its mechanism of action was gathered . However , the pace of discovery with these techniques is somewhat slow , relying on a small number of research groups expert in handling the complex schistosome life cycle and working with both finite yields of parasite ( adult parasites must be harvested from mammalian hosts ) and long screen timelines ( it takes approximately 30 days for S . mansoni infections to become patent in the mouse model [28] ) . Here , we have taken an alternative approach to phenotypic screening by designing a three-component screen workflow built upon juvenile parasites ( schistosomula ) that are easily obtainable from the vector snails and in far greater numbers than adult parasites . The screen is formatted to 96-well microtiter plates thus providing increased throughput and improved interfacing with similarly formatted small molecule libraries maintained in-house at the UCSF Small Molecule Discovery Center ( SMDC; http://smdc . ucsf . edu/ ) . Adult parasite screens in vitro form the second component of the workflow that is completed with compound efficacy tests in a murine model of patent schistosomiasis . Two GO/NO GO filter points are strategically placed in the screen workflow to prioritize compounds more likely to meet the target product profile ( TPP ) for treatment of schistosomiasis and its demand for short course oral chemotherapy [29] . As constructed , the entire process is intended to streamline and accelerate the identification of hit compounds and chemistries in vitro , and leads in vivo . The screen workflow was inaugurated using a library of commercially available and chemically diverse compounds . Approximately 41% of the library comprises drugs already approved for human use thereby opening the possibility for repositioning ( re-profiling or re-purposing ) [30] chemical entities as novel anti-schistosomals . The same collections have already provided a number of leads against other parasites [31] , [32] , [33] . Drug repositioning offers shortened development timelines and decreased risk with compounds having already passed regulatory clinical trials with full toxicological and pharmacokinetic profiles [7] , [30] . All of this adds to up to significant potential cost savings –important in the context of diseases afflicting the poor for which investment returns will be marginal . The results accrued from the inaugural screen are promising in that a number of potent anti-schistosomal single compounds and chemical classes have been identified in vitro , some of which elicit demonstrable anti-schistosomal effects in the murine model of disease . Importantly , these and future data arising from the screen workflow are made public via various online portals to allow those interested examine and mine the outputs and , hopefully , identify their own opportunities for NTD drug development .
A Puerto Rican isolate is maintained in the laboratory using the intermediate snail host , Biomphalaria glabrata and the Golden Syrian hamster Mesocricetus auratus ( 5–6 weeks old; Simonsen labs ) as the definitive host . Animals were maintained and experiments carried out in accordance with protocols approved by the Institutional Animal Care and Use Committee ( IACUC ) at UCSF . Infections with S . mansoni are initiated by subcutaneous injections of 800–1000 cercariae . At 6–7 weeks post-infection ( p . i . ) , hamsters are euthanized with peritoneal ( i . p . ) injections of 50 mg/kg sodium pentobarbital and adult worms harvested by reverse perfusion of the hepatic portal system [34] in RPMI 1640 medium ( Invitrogen , Carlsbad , CA ) . Upon exposure to light , 50–100 snails that are patent with S . mansoni infection , are induced to shed cercariae into the surrounding water . Cercariae are cleaned and concentrated over a series of sieves using distilled water and allowed to stand on ice in a 50 mL polystyrene tube for 1 h . During this time , cercariae clump , settle to the bottom and stick to the inside surface of the tube . The water is poured off and replaced with 9 mL ice-cold ‘Incomplete’ Medium 169 ( [35]; custom made at the UCSF Cell Culture Facility ) that contains 1× penicillin-streptomycin solution . Cercariae are mechanically transformed into schistosomula by passing back and forth between two 10 mL syringes attached via a 22-gauge double-headed needle ( adapted from [36] ) . After deposition into a 9 cm diameter Petri dish , cercarial heads are separated from tails by swirling in Incomplete Medium 169 and the lighter tails aspirated leaving the heads ( schistosomula ) settled in the center of the dish . Under sterile conditions , schistosomula are washed 3 times in Incomplete Medium 169 and allowed to settle over ice in a 1 . 5 mL microfuge tube . Parasites are kept on ice for up to 2 h prior to screening with compounds . As a note , Medium 169 is preferred over RPMI as a culture medium for schistosomula – worms survive with <10% mortality for up to 4 weeks whereas in RPMI , approximately 40–60% of the parasites die within 3 days with continued mortality out to two weeks ( Ruelas and Caffrey , unpublished ) . Adult worms , perfused from hamsters , are washed 5 times in RPMI 1640 containing 1× penicillin-streptomycin solution and 10 µg/mL amphotericin B ( both supplied by the UCSF cell culture facility ) . After 3 further washes in Incomplete Medium 169 , parasites are maintained in ‘Complete’ Medium 169 ( with the addition of 10% fetal bovine serum ( FBS; HyClone , Logan , Utah ) at 37°C and 5% CO2 for up to 24 h prior to screening with compounds . The ‘Spectrum’ and ‘Killer’ compound collections , together comprising 2 , 160 ( 1 , 992 unique ) compounds were purchased from Microsource Discovery Systems , Inc . ( Gaylordsville , CT , USA; http://www . msdiscovery . com/ ) . Information on both is available for download as . xls files from http://www . msdiscovery . com/spectrum . html and http://www . msdiscovery . com/killer . html , respectively . Together the library contains synthetic compounds , natural products and drugs of which 821 are FDA-approved [31] . The library is maintained as 1 and 5 mM stocks in 384-well plates and −80 C at the UCSF Small Molecule Discovery Center that is juxtaposed to the UCSF Sandler Center . For the first component of the screen workflow ( see Figure 1 for schematic ) involving primary screens of schistosomula , 96-well polystyrene dilution plates ( Corning , MA ) are prepared using a Matrix WellMate bulk dispenser and a Biomek FXp liquid handling system . To these plates , the FXp transfers 4 µL of 1 mM compound in neat DMSO from quadrants of the 384 well stock plates . Eighty compounds from each quadrant are transferred to each dilution plate leaving the outer two columns empty . The WellMate then dispenses 16 µL DMSO and 180 µL Incomplete Medium 169 to the dilution plates to yield 20 µM compound in 200 µL 10% DMSO . Finally , the FXp transfers 10 µL of diluted compounds to the 96-well screen plates followed by 180 µL of Complete Medium 169 . Under sterile conditions , 10 µL of schistosomula ( 200–300 worms ) maintained on ice are added manually so that the final concentrations of test compound and DMSO per well are 1 µM and 0 . 5% , respectively . The outer two columns ( 1 and 12 ) of each screen plate are kept empty for eventual manual addition of the anti-schistosomal compounds , PZQ and the cysteine protease inhibitor , K777 [16] , each at 1 and 5 µM . Plates are maintained at 37°C in a 5% CO2 atmosphere . For confirmatory screens of schistosomula ( Figure 1 ) , consensus hits ( details below ) from the primary screen are ‘cherry picked’ using the Matrix WellMate and Biomek FXp . A custom protocol in Pipeline Pilot software ( Accelrys , CA ) generates an Excel file that is read by the FXp . This file is designed to randomly distribute hit compounds among wells containing only DMSO ( ‘dummy wells’ ) in the dilution plate . To start the liquid-handling procedure , the WellMate transfers 48 µL of neat DMSO into the inner 80 wells of 96-well polystyrene dilution plates . From 5 mM stocks in neat DMSO , the Span 8 arm on the FXp then transfers 2 µL of consensus hit compounds ( or DMSO from a plate containing 100% DMSO ) into the dilution plates . To complete the dilution plates , 150 µL of Incomplete Medium 169 are added to each well . From this dilution plate , 4 µL are transferred into the 96-well screen plates followed by 186 µL of Complete Medium 169 . Under sterile conditions , 10 µL schistosomula ( 200–300 ) are then added manually to yield final concentrations of 1 µM and 0 . 5% for test compound and DMSO , respectively . For the second component of the screen workflow involving screening of adult S . mansoni ( Figure 1 ) , the FXp Span 8 transfers 4 µL 5 mM hit compounds in neat DMSO into 96-well polystyrene dilution plates . The hits are distributed randomly in the first few rows of these plates , but with fewer dummy DMSO wells to accommodate the smaller 24-well screen plates . Then , 96 µL of neat DMSO are added to this plate using the WellMate bulk dispenser . From this , 10 µL of the diluted compounds are added manually to the 24-well screen plates and immediately mixed with 0 . 99 mL of Complete Medium 169 to prevent evaporation of DMSO . Under sterile conditions , adult worms ( 4–8 pairs ) are manually added in 1 mL Complete Medium 169 . Final concentrations of test compound and DMSO are 1 µM and 0 . 5% , respectively . Two screen analysts spent approximately four weeks testing the Microsource ‘Killer Collection’ with both schistosomula and adult parasites in order to familiarize themselves with the types of phenotypes arising and their changes as a function of time . Phenotypes were scored using a Zeiss Axiovert 40 C inverted microscope and ×10 and ×2 . 5 objective lens for schistosomula and adults , respectively . Screening analysts are blind to the compound identities which are not disclosed until the conclusion of each of the schistosomular and adult components of the workflow . As a further precaution against subjective bias , each screen analyst visually scores and characterizes phenotype ‘hits’ in isolation . Both analysts then compile ‘consensus hits’ . In those cases for which a consensus cannot be reached the compounds in question are scheduled for re-screening . With repetition , the failure rate to identify consensus hits was decreased to less than 5% per plate for both the schistosomula and adult components of the workflow . It was also during the four week training period that a decision was reached on the compound concentration at which the Microsource collections would be screened . Initial testing of the Killer collection at 10 and 5 µM with schistosomula yielded too many hits ( average of 25 and 15% of the 80 compounds per plate , respectively ) to subsequently perform , in a reasonable time-frame , in vitro screening with the more limiting adult parasites . At 1 µM , however , an average 10% hit rate was achieved . For primary screens of schistosomula , phenotypes were monitored after 7 d a time frame considered long enough to record the development of any potentially relevant phenotype ( Figure 1 ) . For confirmatory screens with schistosomules , phenotypes were scored after 24 h and 7 days in order to identity fast-acting compounds and re-confirm the data from the primary screen , respectively . For adult screens , phenotypes were monitored after 7 and 24 h , and , thereafter , daily up to 4 days ( Figure 1 ) . Two GO/NO GO filters are positioned in the screen workflow in order to prioritize which compounds go forward ( Figure 1 ) based upon the TPP for schistosomiasis treatment and its demand for short course oral therapy [29] . The first filter , placed between the schistosomular and adult components of the workflow , prioritizes compounds yielding phenotypes by 24 h and removes compounds ( where data are available ) that are clearly toxic and/or unsuitable for oral administration . The second GO/NO-GO filter , upon completion of the adult screen component , prioritizes those hit compounds for tests in the murine model of schistosomiasis mansoni . This prioritization is more complex than the first filter as a number of parameters must be simultaneously considered . Primary emphasis is placed on the time to appearance of the phenotype plus the severity of that phenotype , e . g . , fast-acting ‘death’ phenotypes ( <24 h ) are most preferred . Other factors influencing the decision include clinical indication ( if known ) for the compound including undesirable side-effects ( e . g . , hormones disallowed; psychoactives less preferred ) ; oral bioavailability ( preferred over other routes of administration ) and data on acute toxicity ( e . g . , LD50 oral ( p . o . ) , i . p . , and/or intra-venous ( i . v . ) ) . Finally , where compounds with similar chemistries are represented more than once , a single example is initially considered for tests in the mouse model . The third and final component of the S . mansoni screen workflow ( Figure 1 ) entails infections of 4–6 week-old Swiss Webster mice ( Simonsen Laboratories ) with S . mansoni that are initiated by subcutaneous injection of 140 cercariae . Experiments were carried out in accordance with protocols approved by the Institutional Animal Care and Use Committee ( IACUC ) at UCSF . Groups of 4 or 5 mice are used per treatment . Commencing on day 42 p . i . when S . mansoni infections are patent ( i . e . , when parasite eggs are present in feces ) compound is administered once daily ( QD ) and/or twice daily ( BID ) for 4 days . This time period is considered sufficient to record any compound efficacy given that the desired TPP for any new-anti-schistosomal calls for short course therapy [15] , [29] . Compound is administered p . o . ( vehicle is 2 . 5% Cremophor EL unless otherwise stated ) or , i . p . , when data on oral bioavailability are not to hand . The amount of compound to administer is guided by available LD50 values for acute toxicity , in which case compound is given close to that value in order to determine whether a therapeutic window exists . Overt toxicity of compounds ( e . g . , death and behavioral changes ) is assessed daily during and after treatment until the time of euthanasia . At 55 days p . i . , mice are euthanized with an i . p . injection of 0 . 05 mg/g sodium pentobarbital , and adult worms perfused as described above for hamsters . Compound efficacy in vivo is measured as described [16] using a number of criteria and is compared to that of the anti-schistosomal drug , PZQ , as a ‘gold standard . ’ The criteria include the parasitological parameters of numbers of male and female worms recovered by perfusion , and hepatic egg burdens . Also , the amelioration of pathology as evidenced by decreased liver and spleen weights is recorded . Attention is also paid to worm size upon recovery . To recover eggs trapped in liver , whole livers ( or the caudate liver lobe , see below ) from individual mice are excised , weighed and digested in 0 . 7% porcine trypsin in PBS for 1 h at 37°C on an orbital shaker . Eggs are sedimented at 4°C and counted under a dissecting microscope as described previously [37] . Data for worm and hepatic egg counts , and organ weights , were compiled on a per mouse basis and median values calculated per treatment group . All data were subjected to the Mann-Whitney nonparametric test to determine any statistical differences in egg and worm burdens , and organ pathologies between treated and untreated control mice . As an expedient alternative in some experiments , particularly in those cases where worm burdens were not dramatically decreased , hepatic egg counts were calculated as an average per treatment group rather than per mouse . To do this , the caudate liver lobe from individual mice was excised , pooled per treatment group and weighed prior to trypsin digestion and counting of eggs . The single value arising was then calculated with respect to the total liver weight in the group and then divided by the number of mice in the group .
During initial testing of the Microsource ‘Killer Collection’ , it became clear that schistosomula display different and often multiple phenotypes that change over time . Eventually , we could consistently ascribe six phenotypes to worms under chemical insult relative to control worms exposed to 0 . 5% DMSO ( Figure 2; Table S1 ) . The phenotype terms we employed range from the obvious ( ‘dead’ , ‘overactive’ , and ‘rounded’ ) to the more sublime ( ‘dark’ , ‘slow’ ) , yet , nonetheless , clearly distinguishable from DMSO controls . An example of the overactive phenotype is shown in Video S1 which should be compared with normal worm movement displayed in Video S2 . The last phenotype , ‘degenerate but mobile’ , describes those cases in which the worms are clearly motile yet severely disrupted in morphology . An example of this phenotype is produced by PZQ , used as a ‘gold standard’ schistosomicide throughout the screen workflow . PZQ initially elicited an overactive phenotype ( observed within 10 mins ) that progressed to a combination of ‘overactive/degenerate but mobile’ by 7 days ( Figure 2C ) . In contrast , the cysteine protease inhibitor , K777 [16] , also used as a standard compound , had a more progressive effect; “slow/dark” by 3 days leading to ‘dark/slow/dead’ by 7 days ( Figure 2B ) . Similar to schistosomula , adult worms could manifest multiple and changing phenotypes in response to chemical insult ( Figure 3; Table S2 ) . There was some overlap in the phenotypes classified compared to schistosomula: ‘dead’ , ‘dark’ , ‘slow’ and ‘overactive’ remained relevant whereas ‘rounded’ and ‘degenerate but mobile’ did not . Additional , adult-specific phenotypes were: ‘tegumental blebbing’ ( teg . bleb . ) to document damage to the surface ( tegument ) of the adult; ‘sexes separated’ ( sex sep . ) , whereby the male and female worms become unpaired , and the self-evident ‘shrunken . ’ For example , PZQ , elicited ‘shrunken/dark/slow’ phenotypes ( observed within 2 min of addition of 1 µM PZQ ) that progressed to ‘shrunken/dark/sexes sep/teg . bleb/dead’ , by 4 days ( Figure 3C ) . In contrast , K777 had a more progressive effect; ‘slow/dark’ by 2 days leading to ‘dark/slow/sexes sep/on sides/dead’ by 4 days ( Figure 3B ) . Of the 1 , 992 unique compounds comprising the Microsource Spectrum and Killer collections , 118 yielded phenotypes ( termed ‘hits’ representing 5 . 9% of the total ) in the schistosomula primary screen component of the workflow after 7 days at a concentration of 1 µM . The compound names , structures , therapeutic uses and phenotypes identified are listed in Table S1 . The majority of these ( 105 ) were returned as hits in the confirmatory schistosomula screen component after 7 days and , of these , 61 ( 3 . 1% of all the compounds screened ) , were fast-acting , i . e . , phenotypes were recorded at the 24 h time-point ( Table S1 ) . When the Microsource Spectrum and Killer collections are broken down into their component drug classes ( Table 1 , Table S3 ) , and using the 7 d confirmatory screen data , the greatest percentage of hits per class , as might be expected , was for the anthelmintics ( 29% ) . Within this group the known anti-schistosomals PZQ and hycanthone were identified , as were other anthelmintics such as niclosamide , bithionol and pyrvinium pamoate ( Table S3 ) . Examples of other drug classes returning percentage hits greater than 10% are the antibiotics , fungicides , antineoplastics , dopaminergics and seratonergics . Upon completion of the first ( schistosomular ) component of the screen workflow and bearing in mind the target drug profile for schistosomiasis [29] , the first of two GO/NO-GO filters was enacted ( Figure 1; Table S1 ) . Fast-acting compounds ( phenotypes by 24 h ) were prioritized and those clearly toxic or otherwise inappropriate for oral use were removed . Accordingly , all but four ( phenylmercuric acetate , thimerosal , benzalkonium chloride and homidium bromide ) of the 61 fast-acting compounds were prioritized for in vitro tests against adult parasites During the first component of the screen workflow it became clear that certain tricyclic psychoactive compounds within the Microsource collections , notably the phenothiazines and dibenzazepines , elicited a striking “overactive” phenotype that lasted for the 7 day duration of the experiment . To verify the result , a mini-screen incorporating the tricyclic psychoactives and structurally related compounds was set up over three logs of concentration; 1 . 0 , 0 . 1 and 0 . 01 µM ( Table S4 ) . The overactive phenotype was confirmed at 1 . 0 µM for the same compounds and extended to 0 . 1 µM , but not 0 . 01 µM . Examination of structural-activity relationships ( SAR ) in the phenothiazine and dibenzazepine classes indicated the importance of an unsubstituted propyl side chain possessing a terminal dimethylamine function ( Figure 4; Table S4 ) . Even subtle alteration of this pharmacophore , such as the introduction of a branching methyl substituent , led to abrogation of the phenotype ( compare promazine and trimeprazine in Table S4 ) . More dramatic modification of the side chain ( e . g . , shortening , introduction of terminal piperazine or piperidine moieties ) similarly abrogated the overactive phenotype . The trend held whether carbon was substituted for nitrogen at position 10 of the phenothiazine ( or at position 11 of the dibenzazepine ) or whether the tricyclic core was altered internally or substituted . Of the 57 compounds passing the first GO/NO GO filter and pursued in the second ( adult ) component of the screen workflow , 30 were hits ( 1 . 5% of the 1 , 992 compound total ) after the maximal screen time of 4 days at 1 µM ( Table S2 ) . Seventeen and 10 compounds generated phenotypes by the 7 h and 24 h time points , respectively . Included in this set was the antibiotic , anisomycin , and the anthelmintics PZQ ( worms visibly contracted and shrank within seconds of adding 1 µM ) , niclosamide , pyrvinium pamoate and bithionol . Three compounds , including a former chemotherapy of schistosomiasis , hycanthone , generated phenotypes after 2 days . To prioritize candidates for tests in the murine model of schistosomiasis , a second GO/NO GO filter was implemented utilizing a greater number of considerations than for the first filter ( Figure 1; Table S2 ) . As for the first filter , due provision was made for the TPP for schistosomiasis therapy: fast-acting compounds ( phenotypes within 7 and 24 h ) were prioritized but with a preference now for those that generated the most severe phenotypes , e . g . , “teg . bleb” and “dead” . Prioritizations were counter-balanced by available knowledge of compound efficacy , toxicity , and side-effects . Thus , bithionol was deprioritized due to its lack of efficacy in schistosomiasis patients [38] and celastrol due to its toxicity in mice ( death within one day of a single 10 mg/kg i . p . dose ( R . Swenerton , unpublished data ) . Likewise , rhodomyrtoxin B displays low LC ( lethal concentration ) 50 values of between 2 and 20 µM against Hep-G2 ( human hepatocellular carcinoma ) and MDA-MB-231 ( human mammary adenocarcinoma ) cell lines [39] and was , therefore , not considered further . The tricyclic psychoactive compounds ( e . g . , chlorpromazine , imipramine ) were also deprioritized at this stage given that the less severe ‘overactive’ phenotype that appeared by 2 h was either transient ( lasting only until the 24 h time point ) or without progression to a more severe phenotype ( s ) . Also , their psychoactivity and associated side effects , e . g . , sedation , detracted from their immediate consideration . In all , therefore , five compounds were prioritized for efficacy tests in the model of murine schistosomiasis: two antibiotics , anisomycin and lasalocid sodium; two natural products , diffractaic acid and gambogic acid; and the helminthicide/molluscicide , niclosamide ( Figure 5A , Table S2 ) . For the third and final component of the screen workflow , all compounds were administered orally in 2 . 5% Cremophor EL ( unless otherwise stated ) either once ( QD ) and/or twice daily ( BID ) for 4 days to mice with patent S . mansoni infections ( 42 days p . i . ) . The antibiotic , anisomycin , at 100 mg/kg p . o . QD , had no significant effect on male or female worm burdens ( Figure 6A ) , yet decreased hepatic egg burdens by 36% ( calculated as an averaged single value for the treatment group ) . Neither liver nor spleen weights were significantly different from those of the untreated control group ( Figure 6B and C ) . Increasing to a BID administration resulted in toxicity; all mice died between 3 and 10 days after the commencement of treatment . The ionophoric antibiotic , lasalocid sodium , was better tolerated by mice . Significant decreases in male ( 44% ) and female ( 41% ) worm counts were measured at 100 mg/kg QD and BID , respectively ( Figure 6A ) . For egg burdens , reductions of 39 and 55% ( calculated as averaged single values per treatment group ) were measured QD and BID , respectively . Lasalocid sodium also significantly improved organ pathology compared to controls ( Figure 6B and C ) . For the Usnea lichen metabolite , diffractaic acid poor solubility in aqueous media or vehicle prevented the ability to accurately gavage mice . Therefore , i . p . administration in 50 µL 100% DMSO over a range of doses ( 10 , 40 and 100 mg/kg ) was performed to determine compound efficacy while also observing for overt toxicity . No decrease in worm or egg burdens was measured at the lower doses , whereas at 100 mg/kg , all mice died within 7 days of the cessation of treatment ( data not shown ) . Gambogic acid , a xanthone isolated from various species of the Garcinia tree , was recently shown to be non-toxic in rats after oral administration every other day for 13 weeks at 30 and 60 mg/kg in 2% carboxymethylcellulose-sodium [40] . Using the same vehicle at 100 mg/kg QD , no effects on worm or egg burdens were recorded ( data not shown ) . The final compound , niclosamide ( 2′5-dichloro-4′-nitrosalicylanilide ) , a molluscicide and intestinal helminthicide , is poorly absorbed across the intestinal wall . Therefore , we obtained from Bayer a wettable powder formulation of the compound ( marketed as Bayluscide WP 70 ) that is better absorbed ( in rats about a third of an oral dose [41] ) . However , in tests with both niclosamide formulations at 100 mg/kg BID , no effects on worm or egg burdens were noted ( data not shown ) . Further , niclosamide was ineffective at 100 mg/kg BID in two additional vehicles ( 2% Tween80/7% ethanol and 6% PEG 4000/2%Tween 80/7% ethanol ) . In a final attempt to demonstrate efficacy , i . p . administration of niclosamide at 100 mg/kg BID in 100 µL 25% DMSO was without effect – upon dissection of mice the compound was noted to adhere as a solid mass at the injection site on the inner side of peritoneal membrane , indicating that much of the compound had not been absorbed . Based on the strong in vitro efficacy measured for niclosamide against both schistosomula ( Table S1 ) and adults ( Table S2 ) , and notwithstanding its lack of in vivo efficacy , we searched for structurally related compounds that are commercially available and have demonstrated oral efficacy against helminths or protozoa . Three salicylanilides , closantel , oxyclozanide , and rafoxanide , and the nitrothiazolyl-salicylamide , nitazoxanide ( Figure 5B ) were purchased . The salicylanilides are well-established drugs used in the agribusiness sector as helminthicides , including against liver fluke disease caused by Fasciola hepatica [42] . They have also displayed variable efficacies in experiments with farm animals harboring agriculturally important Indian schistosome species such as Schistosoma incognitum and Schistosoma nasale ( [43] , [44] ) . Nitazoxanide ( marketed as Alinia ) is approved for the treatment of diarrhea caused by Cryptosporidium parvum and Giardia lamblia and has shown efficacy against human fascioliasis hepatica [45] . Accordingly , we judged there to be sufficient precedent and data available to move these compounds straight into our mouse disease model . The oral efficacy of these drugs were compared to the ‘gold standard’ drug , PZQ . When administered at 100 mg/kg p . o . QD for 4 days commencing at 42 days p . i . , PZQ significantly decreased male ( 91% ) and female ( 87% ) worm burdens ( Figure 7A ) and these were associated with a decreased hepatic egg load ( 60%; Figure 7B ) and improved organ pathology ( Figure 7C and D ) . The decrease in egg burden was not considered significant , however , due to the low load recorded for one of the control mice . The few worms surviving treatment and recovered by perfusion were the smallest seen in all of the in vivo experiments and some were physically damaged ( not shown ) . By comparison , BID administration of the salicylanilides , closantel and oxyclozanide , at 100 mg/kg yielded less pronounced effects on worm burdens ( only oxyclozanide significantly decreased female loads by 53%; Figure 7A ) and egg burdens were not affected ( Figure 7B ) . However , worms recovered after oxyclozanide treatment were smaller than controls ( not shown ) and organ pathology ( significantly so for the spleen ) was also improved ( Figure 7C and D ) . The third salicylanilide tested , rafoxanide , at either 100 mg/kg QD or BID , caused mouse mortality within 5 days of the cessation of treatment , however , this seemed not to be due to systemic toxicity per se but rather an accretion of drug in the stomach that caused gastric blockage . At 50 mg/kg QD ( i . e . , half the dose of PZQ ) all mice survived . The drug was the most effective of the niclosamide analogs tested significantly decreasing male ( 56% ) and female ( 50% ) worm loads ( Figure 7A ) . Also , worms recovered were smaller than controls ( not shown ) . Egg counts were decreased by 49% , but as noted for PZQ above , the value was not significant due to an outlier control mouse with a particularly low hepatic egg count . Rafoxanide was as effective as PZQ in improving organ pathology ( Figure 7C and D ) . The final niclosamide analog tested , nitazoxanide , was without effect on worm burdens at 100 mg/kg QD and BID ( Figure 8A ) but significantly improved organ pathology BID ( Figure 8B and C ) . Nitazoxanide also decreased egg outputs by 34% ( calculated as a single averaged value per treatment group ) .
Compared to the high profile activity supporting the development of novel anti-protozoal and anti-infective therapies [7] , [8] the pace of drug discovery for anti-schistosomals ( and anthelmintics in general ) is slow . For schistosomiasis , a number of mutually suppressive factors are responsible . Perhaps foremost is the success and clinical reliability of PZQ that have dampened investment in a dedicated drug development pipeline , a situation in stark contrast to malaria and protozoal NTDs for which drug toxicity and/or increasing drug resistance fuel a number of multinational PPP programs to identify new therapeutics [7] . Other contributing factors are; the relatively small number of groups involved in anti-schistosomal discovery , the need to maintain a complex life-cycle that generates finite parasite yields and the long identification and development time lines associated with phenotypic ( whole organism ) screens as traditionally prosecuted in animal models and/or in vitro with adult worms . Despite these drawbacks , phenotypic screening has successfully identified PZQ and other vital anthelmintics ( e . g . , albendazole and ivermectin ) that are in medical use today . Most often , the compounds originated in the animal health sector as part of its discovery programs to identify veterinary anti-parasitics [46] , [47] . For this report , we have designed a phenotypic screen process by introducing a three-component workflow ( Figure 1 ) that places S . mansoni schistosomula at its apex . The intent is to streamline and accelerate the identification of anti-schistosomal compounds by interfacing the helminth with the microtiter plate ( 96- and/or 384-well ) formatted compound libraries and associated robotic liquid handling systems now standard in industry and many academic institutions , and routinely employed to screen the more tractable protozoan parasites [7] . Given their small size ( ∼200×60 µm ) schistosomula are readily adaptable to the 96-well plate format and survive for 7 days with less than 10% mortality under the conditions described . Also , they are quickly and easily transformed from the invasive cercariae that are harvestable in their tens of thousands on at least a weekly basis from vector snails . Both points are immediately attractive and conducive to designing a higher throughput screen workflow . The alternative adult parasite is too large for 96-well plate formatting and can only be harvested from vertebrate hosts ( e . g . , mice and hamsters ) in more limiting numbers , and entails considerable expenditure associated with animal procurement and maintenance . That stated , adults are not omitted entirely from the screening process but are placed downstream of schistosomula when the number of compounds to be tested is more manageable – a consideration also of importance for the final component of the workflow involving the animal model of schistosomiasis . The Microsource collections of 2 , 160 compounds were prosecuted at a throughput of 640 compounds/month for the primary schistosomular screen component . With one full-time technician and an associate analyst it took 20 weeks to complete the in vitro screening of the collections against schistosomula and adults . Efforts to at least double the screening capacity of the first component of the screen workflow are being studied , for example , through the employment of additional staff and expansion of our in-house S . mansoni life-cycle . Also , screen formatting to 384-well plates is being considered together with the complete automation of both the liquid handling of the parasite and phenotype identification and categorization . Data accrued from each component of the screen workflow is available as a flat file online at the UCSF Sandler Center's ‘Low Hanging Fruit’ website http://www . sandler . ucsf . edu/fruit . html and at http://www . collaborativedrug . com/ , a database that can be mined across compounds and parasites to identify molecules and chemistries of interest . Both sites are continually updated as screening campaigns are concluded and it is hoped that the data will contribute to drug discovery efforts for schistosomiasis and other NTDs . The descriptive approach employed here to annotating the dynamic responses of this metazoan parasite to chemical stimuli differs necessarily from the single end-point fluorometric or colorimetric assays , now routine for high-throughput assays of single-celled organisms and with which a rigorous quantification of a live versus death ratio is relatively facile [7] . Given the traditionally slower compound throughput for schistosome screening , the demand for marker dyes or reagent-based kits has simply not been present with visual-based scoring systems being the norm [21] , [22] , [23] . Our attempts to incorporate nuclear dyes ( e . g . , propidium iodide and DAPI ) as a quantitative marker of cell death in schistosomula did not correlate with the clear deleterious action of some compounds observed under bright field microscopy ( Caffrey , unpublished data ) . Often dyes were simply excluded from crossing the schistosome tegument regardless of worm condition . Thus , our decision to visually classify phenotypes , though potentially prone to subjectivity , turned out to be a consistent semi-quantitative approach as employed , i . e . , using blind consensus determination of bioactivity by trained analysts familiar with the parasite's phenotypic manifestations ( see discussion below ) . Further , it might be argued that the workflow , because of its simplicity , and without the need for expensive kits or reagents , is more adaptable to a greater variety of discovery settings . Nevertheless , we are aware that any attempt to improve the quantitative rigor of hit identification and classification should be a primary goal . Accordingly , we are examining a number of automated time-lapse image capture platforms to improve efficiency and accuracy , including the ability to record phenotypes too subtle to be observed with the human eye . In addition to increased throughput and improved automation , the logistical decision to commence the screen workflow with the schistosomulum stage has both potentially advantageous and disadvantageous consequences . Of advantage is that the workflow may identify compounds that are active against both immature and adult stages of parasite , or , at least , against immature parasites . This is important in the context that the current chemotherapy , PZQ , is markedly less effective against the immature ( migratory and sub-adult ) parasite compared to mature egg-laying adults [10] , [11] , [48] . Thus , the identification and development of a small molecule prophylaxis for individuals harboring immature parasites , such as in areas of higher transmission , would be of considerable value . By extension , the opportunity to develop a combination ( possibly synergistic ) therapy with PZQ to decrease the threat of resistance to the latter may also be facilitated by the present screen workflow that commences with the schistosomular stage rather than adults . The concept of a PZQ-based combination therapy based on reciprocal drug efficacy against immature and mature parasites has already shown value with the artemisinin class of compounds ( [14] and references therein ) . A possible disadvantage of the current screening approach is the potential for missing compounds that are inactive against schistosomula , yet , nevertheless , might have yielded interesting phenotypes against adults worms . We accept this possibility as part of the overall goal to streamline and accelerate the identification of anti-schistosomal compounds . We would emphasize that , where smaller compound collections are concerned , the screen workflow can be conducted in a non-hierarchical manner whereby every compound is tested against both schistosomula and adults . Whether a compound is a hit or not or whether it passes or fails the GO/NO GO criteria as implemented here , all screen data are made publicly available for ( re ) interpretation . As to the choice of compound collections maintained at the UCSF SMDC ( http://smdc . ucsf . edu/ ) to initiate the screen workflow , the Microsource Spectrum and Killer collections seemed appropriate for a number of reasons . First , the collections comprise a tractable set of 1 , 992 unique compounds , so that with a modest throughput the first and second components of the workflow were complete within 20 weeks . Secondly , the collections have a track record of yielding novel leads against other parasites including Plasmodium falciparum [31] , [33] and Trypanosoma brucei [32] . Finally , the collections contain a chemically diverse set of natural and synthetic small molecules , 41% ( 821 compounds ) of which are drugs already FDA-approved . From a drug-repositioning standpoint , this is particularly attractive because of the existence of clinical data ( e . g . , adsorption , distribution , metabolism , excretion and toxicity ( ADMET ) ) that could contribute to fast-tracking these compounds as anti-schistosomals , especially as the compounds are off-patent and without intellectual property concerns . Of the 118 compounds identified as hits and phenotypically classified after 7 days of incubation in the primary schistosomular screen component , 105 were confirmed . Likewise , for the adult component of the workflow , repeated tests with compounds resulted , in most cases , in the same phenotypes . Thus , our blinded consensus approach to visually recording bioactivity provided reasonable reproducibility . In further support of the strategy , known schistosomicides , including PZQ and hycanthone , were , without fail , identified and consistently characterized , as were other anthelmintics , such as bithionol and niclosamide . Importantly , direct visual observation allowed us to identify and record the multiple and changing phenotypes that are possible with schistosomes and , not least , discover an apparent SAR for tricyclic psychoactive compounds primarily focused on the structure of the side chain . As yet the molecular target ( s ) of the dibenzazepine and phenothiazine drugs in question is unknown . It is possible that the ‘overactive’ phenotype is not neuroreceptor-mediated but perhaps a result of membrane interference ( depolarization ? ) . Nonetheless , it is interesting that , over a two log-fold concentration , compounds designed to interact with different ligand-gated receptors in humans nevertheless yield the same phenotype in the parasite , suggesting that a single parasite receptor or a discrete subset of receptors may be the target . By mining the available genome sequence information for S . mansoni [18] , [19] , one might envisage RNA interference of candidate cholinergic , dopaminergic or seratonergic receptors in an effort to modulate the overactive phenotype . This would prove the hypothesis that these compounds share a receptor and aid the development of an SAR-based drug discovery program . Both GO/NO GO filters in the workflow were designed in consideration of the TPP demanded for new anti-schistosomal drugs that employs the current therapeutic , PZQ , as a gold-standard . The bar is high - PZQ decreases worm burdens by between 60 and 90% in a single oral dose [3] , [48] . Criteria of speed of appearance , phenotype severity ( death preferred ) and oral suitability were balanced with clinical data on dosage and safety . As interpreted for this report , the second GO/NO GO filter removed a number of compounds and compound classes that elicited striking phenotypic effects . Among these were the tricyclic psychoactive compounds . The ‘overactivity’ they elicited may yet prove therapeutically significant , perhaps by disrupting the parasite's migratory program or its ability to remain in position within the host . Targeting neurotransmission ( if that indeed is the mechanism in schistosomes ) is a successful chemotherapeutic strategy for other helminths ( [49] ) . We will examine the efficacy of these compounds in the murine disease model at low doses either alone or in combination with PZQ . As implemented , the second GO/NO GO filter prioritized five ( niclosamide , anisomycin and lasalocid sodium , diffractaic acid and gambogic acid ) of the 30 hit compounds identified in the second ( adult ) component of the screen workflow for tests in the animal model of schistosomiasis . All five quickly kill ( within hours ) both schistosomula and adults at 1 µM in culture . Also , LD50 toxicity data are available against which a dosing regimen can be prepared and four of the five top hits ( excepting diffractaic acid ) can be administered orally . In the murine model of disease , anisomycin and lasalocid sodium demonstrated varying parasitological efficacies and amelioration of hepatic and splenic pathology . Though not as effective as PZQ , we consider the identification of these novel in vivo anti-schistosomal activities as proof that the screen as conceptualized ( w . r . t . drug-repositioning ) and implemented can identify potentially interesting and chemically diverse compounds . Such compounds might be employed for therapy as is or as leads for further derivatization ( e . g . , anisomycin is chemically relatively simple ) in order to improve bioactivity while reducing toxicity . The point is further underscored with the niclosamide analogs . Though niclosamide and its wettable powder formulation were ineffective in the mouse model of disease , niclosamide's rapid and severe in vitro bioactivity encouraged us to search for other salicylanilide analogs . We identified a number that are well-established in the veterinary sector and possess better oral bioavailability with systemic anthelmintic activity , including against related trematode parasites [42] . The often significant in vivo efficacy of these drugs in both the parasitological and pathological parameters measured , particularly with rafoxanide , encourage further study of the salicylanilides as a source of anti-schistosomal leads . These investigations are underway . In conclusion , as a central component in a pre-clinical drug discovery pipeline for schistosomiasis , we have developed a partially automated phenotypic screen workflow of increased throughput . All data arising are posted and updated online and work continues to improve automation , rigor and throughput . As currently performed , the workflow has already identified a diversity of hit compounds and chemistries in vitro , as well as lead compounds orally bioactive in a short time frame commensurate with the TPP for chemotherapy of schistosomiasis [29] . Critically , the drug-repositioning dimension with the availability of clinical data for many of these hits and leads can be leveraged to optimize further compound development . Accordingly , screens of other libraries containing known drugs are ongoing . Given the possibility of the emergence of resistance to the current PZQ monotherapy , and because any strategic planning for therapy of infectious diseases should incorporate provisions for drug combinations , our future studies will focus on the in vivo performance of the present and future lead compounds , either alone or in combination , including with PZQ .
|
The flatworm disease schistosomiasis infects over 200 million people with just one drug ( praziquantel ) available—a concern should drug resistance develop . Present drug discovery approaches for schistosomiasis are slow and not conducive to automation in a high-throughput format . Therefore , we designed a three-component screen workflow that positions the larval ( schistosomulum ) stage of S . mansoni at its apex followed by screens of adults in culture and , finally , efficacy tests in infected mice . Schistosomula are small enough and available in sufficient numbers to interface with automated liquid handling systems and prosecute thousands of compounds in short time frames . We inaugurated the workflow with a 2 , 160 compound library that includes known drugs in order to cost effectively ‘re-position’ drugs as new therapies for schistosomiasis and/or identify compounds that could be modified to that end . We identify a variety of ‘hit’ compounds ( antibiotics , psychoactives , antiparasitics , etc . ) that produce behavioral responses ( phenotypes ) in schistosomula and adults . Tests in infected mice of the most promising hits identified a number of ‘leads , ’ one of which compares reasonably well with praziquantel in killing worms , decreasing egg production by the parasite , and ameliorating disease pathology . Efforts continue to more fully automate the workflow . All screen data are posted online as a drug discovery resource .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"biotechnology",
"chemical",
"biology/small",
"molecule",
"chemistry",
"pathology",
"infectious",
"diseases/helminth",
"infections",
"biochemistry/drug",
"discovery"
] |
2009
|
Drug Discovery for Schistosomiasis: Hit and Lead Compounds Identified in a Library of Known Drugs by Medium-Throughput Phenotypic Screening
|
Primarily impacting poor , rural populations , the zoonotic malaria Plasmodium knowlesi is now the main cause of human malaria within Malaysian Borneo . While data is increasingly available on symptomatic cases , little is known about community-level patterns of exposure and infection . Understanding the true burden of disease and associated risk factors within endemic communities is critical for informing evidence-based control measures . We conducted comprehensive surveys in three areas where P . knowlesi transmission is reported: Limbuak , Pulau Banggi and Matunggung , Kudat , Sabah , Malaysia and Bacungan , Palawan , the Philippines . Infection prevalence was low with parasites detected by PCR in only 0 . 2% ( 4/2503 ) of the population . P . knowlesi PkSERA3 ag1 antibody responses were detected in 7 . 1% ( 95% CI: 6 . 2–8 . 2% ) of the population , compared with 16 . 1% ( 14 . 6–17 . 7% ) and 12 . 6% ( 11 . 2–14 . 1% ) for P . falciparum and P . vivax . Sero-prevalence was low in individuals <10 years old for P . falciparum and P . vivax consistent with decreased transmission of non-zoonotic malaria species . Results indicated marked heterogeneity in transmission intensity between sites and P . knowlesi exposure was associated with agricultural work ( OR 1 . 63; 95% CI 1 . 07–2 . 48 ) and higher levels of forest cover ( OR 2 . 40; 95% CI 1 . 29–4 . 46 ) and clearing ( OR 2 . 14; 95% CI 1 . 35–3 . 40 ) around houses . Spatial patterns of P . knowlesi exposure differed from exposure to non-zoonotic malaria and P . knowlesi exposed individuals were younger on average than individuals exposed to non-zoonotic malaria . This is the first study to describe serological exposure to P . knowlesi and associated risk factors within endemic communities . Results indicate community–level patterns of infection and exposure differ markedly from demographics of reported cases , with higher levels of exposure among women and children . Further work is needed to understand these variations in risk across a wider population and spatial scale .
After the initial recognition of a large number of human cases of the zoonotic malaria Plasmodium knowlesi in 2004 and advent of routine diagnosis of malaria cases by molecular methods , increasing numbers of human P . knowlesi cases have been reported in Southeast Asia and P . knowlesi is now the most common cause of human malaria in Malaysian Borneo [1–3] . Although regional control programmes have reduced the incidence of other malaria species in Malaysia and the Philippines , such as P . falciparum and vivax , the emergence of P . knowlesi presents a challenge to malaria elimination programmes . Despite increasing amounts of data available for symptomatic malaria cases presenting at hospital facilities , little is known about patterns of P . knowlesi exposure and infection at a community level [4] . Effectively targeting resources to identify and control P . knowlesi requires a detailed understanding of environmental and social risk factors . Carried by long and pig-tailed macaques ( Macaca fasicularus and M . nemestrina ) , environmental changes affecting contact between people , mosquito vectors and simian hosts are believed to contribute to this apparent emergence of P . knowlesi in people [5 , 6] . Anopheles balabacensis , the main knowlesi vector , has been associated with forest environments but is also found in peridomestic and agricultural areas [7 , 8] . Associations between deforestation and increases in village-level incidence have been shown for clinical cases but this may not fully reflect exposure in the wider community [9] . Additionally , multiple studies have reported asymptomatic P . knowlesi infections , including in women and children , demographic groups comprising a minority of cases reported to facilities [10–14] . Patterns of community-level exposure can be assessed by the prevalence of specific antibodies against malaria parasites; these antibodies reflect exposure to previous infection and can be used to characterise the level of transmission and identify areas or groups with higher transmission [15] . These serological markers may be particularly useful in low transmission settings , where the probability of detecting infections is low [16] . Seroconversion rates derived from age specific sero-prevalence have also been shown to be closely correlated with more traditional measures of malaria transmission intensity , such as entomological inoculation rates or parasite prevalence , and can be used to identify differences in spatial patterns in transmission [17 , 18] . Further , as these antibody responses represent exposure over time , longer term transmission patterns and temporal changes in transmission can be evaluated [19] . There are an increasing number of reagents for serological studies available for both P . falciparum and P . vivax e . g . [17 , 20 , 21]but antigens specific for P . knowlesi have only recently been described[22] . This study aimed to characterise these community level patterns of serological exposure to and prevalence of asymptomatic parasitemia of P . knowlesi and other malaria species in three case study communities where P . knowlesi transmission has been reported; a largely deforested and highly fragmented site at Matunggong , Kudat , an area with large patches of secondary forest bordering large scale clearing for an oil palm plantation in Limbuak , Pulau Banggi in Sabah , Malaysia and an area with intact secondary forest and some remaining primary forest in Bacungan , Palawan , The Philippines ( Fig 1 ) . These areas were selected as areas representative of locations were P . knowlesi transmission is occurring based on district hospital reports and were the sites of integrated entomology , primatology and social science studies within a wider research programme on risk factors for P . knowlesi ( http://malaria . lshtm . ac . uk/MONKEYBAR ) . P . knowlesi is the main cause of reported human malaria in both the Matunggong and Limbuak sites while only few sporadic P . knowlesi cases have been reported from Bacungan [23–25] . Based on reporting of symptomatic cases to the national malaria programmes , the annual parasite incidence per 1000 people for P . knowlesi in 2014 was 12 for Matunggong , 2 for Limbuak and 0 for Bacungan .
This study was approved by the Medical Research Sub-Committee of the Malaysian Ministry of Health ( NMRR-14-713-21117 ) , the Institutional Review Board of the Research Institute for Tropical Medicine , Philippines and the Research Ethics Committee of the London School of Hygiene and Tropical Medicine ( 8340 ) . Written informed consent was obtained from all participants or parents or guardians and assent obtained from children under 18 in this study and all methods were performed in accordance with relevant guidelines and regulations . This study involved comprehensive sampling of all individuals residing within the study areas . Study sites were selected based on the locations of previously reported clinical P . knowlesi cases and all households within these communities were enumerated and geo-located . All individuals were asked to participate in the study and consenting individuals were interviewed on demographic characteristics , movement patterns , malaria prevention methods and land use practices . Individuals were excluded if they were less than 3 months old , had not primarily resided in the area for the past month or could not be reached after three attempts to contact them , including during evenings and weekends . Finger-prick blood samples were collected to test for malaria infection and exposure; these included blood smears to detect malaria parasites by microscopy and approximately 200μl whole-blood specimens collected in a tube containing EDTA ( Becton-Dickinson , Franklin Lakes , New Jersey ) and three 20μl spots stored on filter paper ( 3MM , Whatman , Maidstone , United Kingdom ) . Filter paper was dried and stored with desiccant at 4°C . All blood smears were examined by trained malaria microscopists . DNA was extracted from filter paper or 10 μl blood pellets using the Chelex-100 boiling method and a nested polymerase chain reaction ( PCR ) method targeting the Plasmodium small subunit ribosomal RNA ( ssRNA ) was used to identify malaria infected individuals , as described by [10 , 26] . This assay used the genus-specific primers rPLU1 ( 5’-TCA AAG ATT AAG CCA TGC AAG TGA-3’ ) and rPLU5 ( 5’-CCT GTT GTT GCC TTA AAC TTC-3’ ) for nest 1 and rPLU3 ( 5’-TTT TTA TAA GGA TAA CTA CGG AAA AGC TGT-3’ ) and rPLU4 ( 5’-TAC CCG TCA TAG CCA TGT TAG GCC AAT ACC-3’ ) for nest 2 . Thermal cycling conditions for primary and nested PCRs were 35 cycles at 94°C , 60°C and 72°C . Samples positive for the Plasmodium genus were then screened using species specific primers targeting the ssRNA region; for P . knowlesi these included PkF1140 ( 5’-GATTCATCTATTAAAAATTTGCTTC-3’ ) and PkR1150 ( 5' GAGTTCTAATCTCCGGAGAGAAAAGA 3' ) for 35 cycles at 50°C , 72°C and 94°C . All products were visualised on a 2% agarose gel . PCR for malaria infection was performed at laboratories at the Universiti Sabah Malaysia in Malaysia and Research Institute for Tropical Medicine in the Philippines , with PCR validation of a subset of samples at the London School of Hygiene and Tropical Medicine in the UK . Enzyme-linked immunosorbent assays ( ELISA ) were performed as previously described [27] . Briefly , 3 mm disc was excised from each dried blood spot and incubated in reconstitution buffer ( PBS/tween with sodium azide ) overnight at 4°C . Antibodies were eluted from the blood spots equivalent to a 1:100 dilution of whole blood or a 1:200 dilution of serum [16] . Antibody responses were measured against apical membrane antigen-1 or the 19 kDa fragment of merozoite surface protein-1 for P . vivax ( PvAMA-1 and PvMSP-119 , respectively ) , P . falciparum ( PfAMA-1 ( PMID: 17192270; PMID: 19165323 ) and PfMSP-119 ( PMID: 8078519 ) and P . knowlesi SERA3 antigen 2[22] . The Pk serine repeat antigen ( SERA ) 3 antigen 2 ( PKNH_0413400; chromosome 4 ) is a novel recombinant protein , N-terminally located between positions 826–998 aa , inclusive . SERA3 ( 1079 aa ) belongs to a multigene family whose members encode a papain-like cysteine protease domain ( ref: PMID: 21423628 ) . In P falciparum , the N-terminal domain of SERA 5 is showing promise as a potential vaccine candidate ( ref: PMID: 24886718 , PMID: 27343834 ) . The recombinant protein was expressed in Escherichia coli and affinity purified by a GST tag . Knowlesi -exposed hospital clinical case control samples showed antigen specific reactivity to the SERA3 antigen 2 recombinant when compared to responses from European malaria naïve and Ethiopian vivax-exposed serum samples ( Herman et al . submitted ) Eluates were tested in duplicate at a final concentration of 1:1000 for all antigens except 1:2000 for PfAMA-1 . In addition , blank wells and a dilution series of the appropriate positive plasma pool were added per plate . Positive controls based on a hyper-immune endemic adult Tanzanian pool ( PMID: 15792998 ) , a lyophilised anti-malaria patient sample ( NIBSC , UK; 72/96 ) and pooled Pk-exposed hospital serum samples were used to assay for P . falciparum , P . vivax and P . knowlesi antigens , respectively . Polyclonal rabbit anti-human IgG-HRP ( Dako , Denmark ) was used at 1/15 , 000 dilution and plates were developed using TMB ( One component HRP microwell substrate , Tebu-bio ) . Optical density ( OD ) values were measured at 450 nm with a microplate reader . Values in excess of 1 . 5 CV between duplicates were considered fails and re-ran . OD values were corrected by subtracting the background of the blank well per plate . For P . falciparum and P . vivax OD readings , values were normalised between plates using a standardised control . Normalisation was not done for P . knowlesi results due to the lack of standard control . All serological analysis was performed at the Universiti Malaysia Sabah and the London School of Hygiene and Tropical Medicine . All households and roads within these areas were geo-located using a hand-held GPS ( global positioning system ) . Land cover maps were derived from LANDSAT 8 30m resolution satellite images [28] and supervised classification was performed using random forests [29 , 30] . In order to generate training data , high- resolution aerial images of areas within study sites were produced using the Sensefly eBee unmanned aerial vehicle flown at 400 metres above ground level ( UAV or drone; Sensefly , Cheseaux-sur-Lausanne , Switzerland ) and processed using Postflight Terra 3D ( Pix4D SA , Lausanne , Switzerland ) as described by [31] . These data were manually digitised and classified as forest , agricultural land ( including cropland and agroforestry such as rubber and palm oil ) , open areas and water bodies . Additional data on elevation , aspect and slope was extracted from the ASTER global digital elevation model [32] . All data were resampled to 30m per pixel and datasets including topographic variables , distance from roads and houses , normalised differential vegetation indices ( NDVI ) and Landsat satellite data were included in the initial model . Random forest models were run using 10 , 000 trees to ensure stability and were run iteratively with least predictive variables excluded at every run [33] . A random subset of the training data for each site was withheld to independently validate the classification; estimated classification accuracy was 88% , 97% and 85% for Matunggung , Limbuak and Palawan respectively ( Fig 2 ) . These classified land cover maps were used to calculate distance from the household to the forest edge . The proportions of different land types surrounding all households were evaluated for 100m , 500m and 1000m buffer radii . Additionally , the level of forest fragmentation was assessed within 500m and 1000m of each household; this was represented as the ratio of forest perimeter to forest area as described by [34] . All geographic data was stored and visualised in a Geographic Information System using ArcGIS ( ArcGIS , Redlands , USA ) and all other analysis was performed using R statistical software ( R Foundation for Statistical Computing , Vienna , Austria , http://www . R-project . org ) . Questionnaire data was collected electronically using Pendragon Forms VI ( Pendragon Software Corporation , Chicago , USA ) and analysed using R statistical software . To define sero-positive individuals , mixture models were fit for normalised optical densities ( ODs ) , with the distribution of ODs modelled as two Gaussian distributions . Cut off values to define sero-prevalence for each antigen were defined as the mean OD of the sero-negative population plus 3 standard deviations for P . falciparum and P . vivax as described by [16] . For the P . knowlesi antigen a more parsimonious cutoff value was defined as the mean OD plus 5 standard deviations due to a lack of prior data . Because the assays were run in different laboratories , cut off values were defined separately for each antigen , malaria species and location ( Palawan and Sabah ) . For P . falciparum and P . vivax , individuals were considered positive if they were positive for either MSP-1 and/or AMA-1 . Reversible catalytic models were fit to age sero-prevalence data using maximum likelihood methods; these models were then used to generate age sero-prevalence curves and estimate the seroconversion rate ( SCR ) [17] . Evidence of historical changes in transmission was explored by using profile likelihood plots . Models with two SCR were assessed by likelihood ratio tests and used if the fit was significantly better ( p < 0 . 05 ) than models with a constant seroconversion rate [19] . Models were fit separately for each parasite species and site . Risk factors associated with P . knowlesi sero-positivity were evaluated using multivariate logistic regression , with household included as a random effect to account for correlation between individuals from the same household . An additional model was developed to compare individuals sero-positive for P . knowlesi with those sero-positive for non-zoonotic malaria species . Explanatory variables included age , gender , site , individual and household level farming activities , residence in the area , elevation and distance to forest . Additionally , the proportions and configuration of different land types were extracted for each household at 100m , 500m and 1000m radii and categorised as greater or less than 30% coverage within a specific radius in the final model . Univariate analysis was conducted for all explanatory variables and variables with p < 0 . 2 were included in multivariate analyses . For highly correlated variables ( such as land cover proportions at different radii ) , single variables were selected based on marginal increases in Akaike Information Criterion ( AIC ) . The final adjusted models were developed by retaining all variables significant at a 0 . 05 level and variables were added in a forward stepwise fashion to check for interactions . Potential residual spatial autocorrelation of exposure to P . knowlesi was assessed separately for all sites using Moran’s I . Correlation between spatial patterns of exposure to P . knowlesi and nonzoonotic malaria species was explored through correlograms , plots of spatial autocorrelation with lag distances . First , ODs were log-transformed and adjusted for age by linear regression as described by [18] . For each site , cross-correlograms of antibody responses to P . knowlesi and each other antigen were plotted . Correlation ranges were determined by significance values ( p < 0 . 05 ) of individual bins of lag distances of 500m . Pairwise correlation between antibody responses was determined using a simple Mantel test to test the significance of associations [35 , 36] .
Two microscopy positive individuals were identified from the Matunggong , Kudat site; these were both subsequently identified as P . knowlesi mono-infections by PCR . All PCR infections were re-confirmed at the laboratory in the U . K . Both of these individuals were male plantation workers ( ages 21 and 25 ) residing in the same household . An additional two individuals in Matunggong were microscopy negative but identified as P . knowlesi infected by molecular methods; these included a three year old girl and 33 year old woman residing in different villages within the study site . Only one out of these four infected individuals identified self-reported history of fever . None of the survey participants in either the Limbuak or Bacungan sites were positive by microscopy or PCR and no infections with any other malaria species were identified in any participants . Overall , 7 . 1% ( 178/2503 ) of the population surveyed was seropositive to P . knowlesi ( Fig 3 ) . Exposure varied substantially between study sites , with the highest P . knowlesi antibody prevalence detected in Limbuak , Pulau Banggi ( 11 . 7%; 93/795 ) followed by 6 . 8% ( 79/1162 ) in Matunggong Kudat . Bacungan , Palawan had the lowest sero-prevalence ( 1 . 1%; 6/546 ) . Similar reactivity to P . knowlesi was observed in men ( optical density ( OD ) : med: 0 . 035 , IQR: 0 . 006–0 . 094 ) and women ( OD: median: 0 . 035 , IQR: 0 . 007–0 . 089 ) and gender was not significantly associated with P . knowlesi sero-positivity ( OR: 0 . 99 , 95% CI: 0 . 71–1 . 37 , p = 0 . 95 ) . Antibody prevalences to P . falciparum and P . vivax were higher in all sites , with 16 . 1% ( 364/2266 ) of individuals sero-positive to one or both P . falciparum antigens and 12 . 6% ( 270/2141 ) positive for one or more P . vivax antigens . Sero-prevalence to P . falciparum was 16 . 9% ( 196/1162 ) in Matunggong , 13 . 5% ( 107/795 ) in Limbuak and 10 . 4% ( 61/587 ) in Bacungan . In contrast , reactivity to P . vivax was highest in Limbuak ( 16 . 7%; 133/795 ) with sero-prevalences of 6 . 9% ( 80/1162 ) and 9 . 7% ( 57/587 ) in Matunggong and Bacungan respectively . Due to insufficient samples and non-systematic errors in labelling , results for all antigens were not available for all individuals . Out of individuals with complete test results for all antigens , 25 . 7% ( 499/1941 ) of participants were sero-positive to at least one species of malaria and 7 . 9% ( 54/1941 ) were sero-positive for two or more malaria species . Of individuals exposed to P . knowlesi , 29 . 7% ( 53/ 178 ) were also positive for P . falciparum or P . vivax antigens . There was no evidence of correlation between P . knowlesi and other antigens tested ( S1 Fig ) . Sero-prevalence was positively associated with increases in age for all antigens tested . However , despite this , seroreactivity , including individuals with high antibody titres ( S2 Fig ) , was still detected in the youngest age groups and 4 . 2% ( 39/921 ) individuals under 15 years had antibodies to P . knowlesi , 3 . 5% ( 29/821 ) had antibodies to P . falciparum and 2 . 9% ( 23/792 ) to P . vivax . Changes in age sero-prevalence were more pronounced for P . falciparum and P . vivax , with 32 . 9% ( 78/237 ) and 28 . 1% ( 64/228 ) reactivity to P . falciparum and P . vivax in individuals over the age of 60 years . In contrast , antibodies for P . knowlesi were detected in 9 . 4% ( 25/265 ) of individuals over 60 years old and the highest sero-prevalence was detected in adults from 45–60 years old ( 11 . 6%; 43/370 ) . As reactivity to P . knowlesi was low and not evenly distributed through the population , seroconversion rates ( SCR ) for P . knowlesi could not be calculated . Historical changes in falciparum transmission intensity were apparent in all sites and SCR models fitted with two forces of infection suggest substantial reductions in P . falciparum transmission occurred 18–30 years ago ( p < 0 . 05 ) ( Fig 4 ) . Strong evidence of decreased transmission intensity for P . vivax was only seen in Limbuak , where transmission decreased over 25-fold in the past 20 years . Demographic and environmental characteristics of survey participants are summarised in ( Table 1 ) . In addition to age and site , reporting farm or plantation work as a primary occupation was positively associated with P . knowlesi sero-positivity ( Table 2 ) . Higher proportions of forest cover within 1km of the household and cleared areas within 500m of the house were both associated with increased odds of P . knowlesi positivity . While forest fragmentation , elevation and agricultural land around the house were significant within the univariate analysis , none of these variables were significant in the final multivariate model ( Supplementary information , S1 Table ) . Similar proportions of men and women reacted to P . knowlesi in all sites and gender was not associated with sero-positivity . Individuals reacting to P . knowlesi were more likely to be younger than individuals sero-positive for only non-zoonotic malaria species ( Table 3 ) . Forest cover was not associated with exposure to non-zoonotic malaria and malaria positive individuals residing in areas with high forest cover around the house had 4 . 86 ( 95% CI: 2 . 30–11 . 37 ) the odds of being positive for P . knowlesi . Similarly , cleared areas around the house were also positively associated with P . knowlesi cases compared to other malaria species . Based on Moran’s I , there was no evidence of residual spatial autocorrelation for P . knowlesi antibody responses ( Moran’s I p > 0 . 2 for all sites ) . There was no significant spatial correlation detected between age-adjusted antibody responses for P . knowlesi and other malaria species for either Matunggong or Limbuak ( p > 0 . 30 for all pairwise comparisons ) . Comparisons between P . knowlesi and other malaria species could not be evaluated for Bacungan due to the low prevalence of P . knowlesi sero-positivity .
This is the first study to describe exposure to P . knowlesi through antigen specific antibody responses and associated risk factors and is one of few studies to assess P . knowlesi carriage prevalence at a community level . Results indicate spatial and temporal patterns of P . knowlesi transmission differ markedly from other non-zoonotic malaria species within the region . Although P . knowlesi sero-positivity was associated with some landscape attributes within these communities , extensive cross sectional surveys are needed to identify ecological risk factors across a broader geographic scale . Sero-prevalence data indicate distinct heterogeneities in P . knowlesi transmission intensity between sites . Although formal comparisons between P . knowlesi infection and exposure could not be undertaken due to the low prevalence of parasite carriage , these geographical differences in transmission mirror hospital-based reporting rates in the study sites at Kudat , Pulau Banggi and Palawan [23–25] . These results also highlight the utility of serological techniques to identify differences in transmission intensity in settings where the sensitivity of parasite prevalence surveys is limited by the scarcity of infected individuals and suboptimal diagnostics . This is the first time these knowlesi-specific antigens have been used at a population level to assess species-specific exposure to malaria . Although high levels of homology between P . knowlesi and P . vivax indicate the possibility of cross reactivity between antigens , relatively low numbers of individuals were identified as sero-positive for both knowlesi and vivax ( 2%; 43/2102 individuals with results for both assays ) and individuals could have been plausibly exposed to both species due to the co-endemicity of these species within this region . Additional work has been done to characterise response to P . knowlesi in vivax-exposed individuals and validate these antigens for population-based studies[22] . Changes in seroconversion rates can also reflect temporal changes in malaria transmission . In Sabah , state-wide malaria notification records describe dramatic decreases in clinical P . falciparum and P . vivax cases within the past 20 years following the scale up of malaria control and elimination programmes [2] . The Philippines has also reported a substantial decline in the number of malaria cases reported in the past few decades , most notably for P . falciparum[37] . These changes are evident in seroconversion rates to non-zoonotic malaria species from the 3 sites with over 5-fold difference between current and previous SCRs . P . knowlesi exposure was identified in children under 5 in all sites , suggesting recent or on-going transmission albeit at a low level . Further work is needed to refine P . knowlesi serological analysis to allow for antigenic variation , identify further antigenic targets and assess the differential responsiveness of individuals and longevity of antibody responses [38] . Despite these similarities between existing case data and community-level exposure to P . knowlesi , levels of exposure between different demographic groups varied markedly from clinical case reports . While clinical P . knowlesi has been commonly reported in adult men , men and women had similar antibody reactivity to P . knowlesi antigens in all sites [23 , 39] . Within Kudat district , wide age distributions and family clusters of knowlesi cases have previously been described; however , from 2012–2015 , 73% ( 84/115 ) and 77% ( 27/35 ) of all clinical cases reported from Kudat and Pulau Banggi respectively were men [23] . Asymptomatic knowlesi carriage has been detected in higher proportions of women by this study and other studies; however these results are extremely limited by sampling design and the small numbers of infected individuals detected [10 , 12] . As forest and agricultural activities have been identified as risk factors for clinical P . knowlesi infection , more men could develop clinical infections due to higher exposure or number of bites; however , this requires further research to assess [40] . Larger scale population-based cross sectional surveys are needed to determine if these patterns occur in the wider community and if P . knowlesi affects groups which may be underrepresented by current passive surveillance systems . P . knowlesi exposure was also associated with landscape factors . Both the proportion of forest cover and cleared areas around the household were positively associated with knowlesi sero-positivity , potentially reflecting the higher vectorial capacity and sporozoite rates reported in secondary forest within these study sites [7] . Although plantation or farm work as a primary occupation was associated with increased exposure and previous reports have described associations between P . knowlesi and forest activities , data on movement into different environments was not available for all survey participants [39 , 41] . Instead , to explore the potential range of spatial interactions between people and mosquito vectors , proportions of habitat were evaluated at different buffer distances around houses . The significance of both clearing and forest areas at different radii suggests the importance of edge effects , transition areas between habitats where increased overlap of human , macaque and mosquito populations may occur [9 , 42 , 43] . Despite this , no associations were identified between metrics of fragmentation or distance to forest edges; this may reflect the limited environmental variation within these small spatial scales . Future studies could assess the importance of these variables across different ecotypes as well as collect more detailed data on the human movement into different environments , particularly during peak mosquito biting times . These spatial patterns differed markedly from exposure to other non-zoonotic malaria species . Individuals with antibodies to P . knowlesi were more likely to reside in areas with higher proportions of forest cover; this may reflect differences in disease dynamics between species or temporal changes in transmission . Because of the longevity of antibody responses and the rapid rates of land use change within these areas , seroreactivity to non-zoonotic species is probably more likely to be associated with past rather than current environmental factors . The main mosquito vector species of P . knowlesi , Anopheles balabacensis , was historically incriminated as the main vector of other human malaria species within these same areas [44 , 45] . While these vectors have been primarily associated with forest habitats , high vector densities have also been reported in small scale farms and other habitat types [7 , 42 , 46] . Deforestation and increased application of vector control measures may have triggered changes in vector composition and biting preferences; similarly , habitat changes and encroachment of human settlements into forest areas may have also led to changes in macaque population densities and closer contact between macaques , people and mosquito vectors [6 , 47] . The main limitations of this study are the non-randomised population sampling approach and limited geographical scale . While this study describes fine scale patterns of malaria exposure and infection within these three case study communities , these results cannot be generalised to extrapolate P . knowlesi risks across wider populations . As this study surveyed three relatively homogenous populations , there was minimal variation in environment , ethnicity , socioeconomic status and access to healthcare within each site . Identifying environmental and population-level risk factors will require randomised sampling across a wider ecological gradient; community level data on presence and absence of exposure and infection are required to understand spatial heterogeneity of disease transmission and develop and refine predictions of disease risk [48] . Additionally , extensive surveys of parasite prevalence may allow the application of genetic approaches to track parasite diversity and transmission and explore the roles of host and parasite genetic factors . Despite these limitations , this study describes P . knowlesi infection and exposure within these communities and illustrates how serologic markers can be used to describe differences in transmission intensity between malaria species in low transmission settings . Results from these surveys indicate patterns of P . knowlesi exposure and infection within the community may be substantially different from cases detected by passive surveillance systems . Cross sectional surveys across a broader geographical scale are needed to describe spatial variation in transmission intensity and identify associated environmental and population-based risk factors . Integration of serology into these surveys would provide vital information on rare infections for control programmes [49] .
|
Plasmodium knowlesi is a species of malaria parasite found in wild macaque populations which is now the main cause of human malaria in Malaysian Borneo . Spread from macaques to people through infected mosquitoes , human P . knowlesi malaria cases have primarily been reported in adult men working in forests or plantations . However , little data is available on the extent of asymptomatic infections or people exposed to P . knowlesi not reporting to clinics . We conducted comprehensive surveys of three case study communities in Malaysian Borneo and Palawan , the Philippines with varying numbers of P . knowlesi cases reported . In addition to testing for infection , we measured species-specific antibody responses to P . knowlesi and other malaria species to identify exposed individuals . Few asymptomatic infections were detected and varying levels of P . knowlesi exposure was detected between sites . P . knowlesi exposure was identified in both men and women and associated with farm work and forest and clearing around the house . Spatial patterns and risk factors for P . knowlesi differed from other malaria species , highlighting the need for knowlesi specific disease control measures . Results suggest more people are exposed to P . knowlesi than are identified at clinics and exposure to P . knowlesi may occur in different demographic groups and geographic areas than previously reported .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"parasite",
"groups",
"ecology",
"and",
"environmental",
"sciences",
"plasmodium",
"immunology",
"tropical",
"diseases",
"vertebrates",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"animals",
"parasitology",
"mammals",
"primates",
"apicomplexa",
"protozoans",
"infectious",
"disease",
"control",
"antibody",
"response",
"old",
"world",
"monkeys",
"infectious",
"diseases",
"monkeys",
"malarial",
"parasites",
"ecosystems",
"immune",
"response",
"macaque",
"eukaryota",
"ecology",
"forests",
"biology",
"and",
"life",
"sciences",
"malaria",
"amniotes",
"organisms",
"terrestrial",
"environments"
] |
2018
|
Exposure and infection to Plasmodium knowlesi in case study communities in Northern Sabah, Malaysia and Palawan, The Philippines
|
Over the recent years , several homologues with varying degrees of genetic relatedness to hepatitis C virus ( HCV ) have been identified in a wide range of mammalian species . HCV infectious life cycle relies on a first critical proteolytic event of its single polyprotein , which is carried out by nonstructural protein 2 ( NS2 ) and allows replicase assembly and genome replication . In this study , we characterized and evaluated the conservation of the proteolytic mode of action and regulatory mechanisms of NS2 across HCV and animal hepaciviruses . We first demonstrated that NS2 from equine , bat , rodent , New and Old World primate hepaciviruses also are cysteine proteases . Using tagged viral protein precursors and catalytic triad mutants , NS2 of equine NPHV and simian GBV-B , which are the most closely and distantly related viruses to HCV , respectively , were shown to function , like HCV NS2 as dimeric proteases with two composite active sites . Consistent with the reported essential role for NS3 N-terminal domain ( NS3N ) as HCV NS2 protease cofactor via NS3N key hydrophobic surface patch , we showed by gain/loss of function mutagenesis studies that some heterologous hepacivirus NS3N may act as cofactors for HCV NS2 provided that HCV-like hydrophobic residues are conserved . Unprecedently , however , we also observed efficient intrinsic proteolytic activity of NS2 protease in the absence of NS3 moiety in the context of C-terminal tag fusions via flexible linkers both in transiently transfected cells for all hepaciviruses studied and in the context of HCV dicistronic full-length genomes . These findings suggest that NS3N acts as a regulatory rather than essential cofactor for hepacivirus NS2 protease . Overall , unique features of NS2 including enzymatic function as dimers with two composite active sites and additional NS3-independent proteolytic activity are conserved across hepaciviruses regardless of their genetic distances , highlighting their functional significance in hepacivirus life cycle .
Approximately 63–79 million individuals were estimated to be chronically infected by hepatitis C virus ( HCV ) worldwide in 2015 and are at risk of developing severe liver disease including fibrosis , cirrhosis and hepatocellular carcinoma . While the total number of HCV infections is expected to decline or remain flat in many countries thanks to the new era of direct acting antiviral agents ( DAAs ) , HCV-related mortality and morbidity is expected to increase as the infected population ages and progresses to more advanced liver diseases [1] . There is no vaccine available and major challenges in basic , translational and clinical research remain . HCV is a single-stranded positive sense RNA virus belonging to the Hepacivirus genus of the Flaviviridae family . Until recently , the only known members of the Hepacivirus genus were HCV and GB virus B ( GBV-B ) , a hepatotropic virus of unknown origin identified in experimentally infected New World primates [2] . GBV-B infected tamarins ( Saguinus species ) and marmosets ( Callithrix species ) generally develop acute self-resolving hepatitis [3] , although several cases of chronic infections have also been reported , highlighting the value of this animal model [4–6] . Over the recent years , a growing number of phylogenetically-related HCV homologues have been identified in the wild in a wide range of mammalian species , including horses [7 , 8] , cattle [9 , 10] , rodents [11–13] , bats [14] and Old World primates [15] . Recent studies demonstrated that equine hepaciviruses cause mild hepatic disorders and may establish protracted infections [16–18] . Further epidemiological , pathogenesis and molecular studies are awaited to assess whether these recently identified hepaciviruses have a potential for zoonosis . GBV-B and the newly discovered hepaciviruses share between 25 and 50% nucleotide sequence identity with HCV . The equine nonprimate hepacivirus ( NPHV ) , the best characterized of the novel hepaciviruses , is HCV genetically closest known relative , whereas GBV-B is one of the hepaciviruses most distantly related to HCV ( Fig 1 and S1 Fig ) . The HCV genome consists of a 9 . 6 kb positive-strand RNA molecule and encodes a polyprotein precursor that is cleaved co- and post-translationally by cellular and viral proteases to yield the capsid protein ( C ) and the two envelope glycoproteins ( E1 and E2 ) , as well as seven nonstructural ( NS ) proteins ( p7 , NS2 , NS3 , NS4A , NS4B , NS5A and NS5B ) . Several studies demonstrated that , despite limited sequence homology , GBV-B and HCV share a common genomic organization including enzymatic functions [19–22] . Although experimental studies are not as extensive , the recently identified hepaciviruses and particularly the equine NPHV are predicted to have a genomic organization similar to HCV and GBV-B [23] , including an internal ribosome entry site ( IRES ) in their 5' nontranslated regions ( 5’NTRs ) [24–26] and binding sites for miR-122 , an essential host factor for viral RNA translation and/or replication [18 , 27 , 28] . In addition , the NS3-4A serine proteases of several hepaciviruses have been reported to disrupt host innate immune responses through the cleavage of mitochondrial antiviral signaling protein ( MAVS ) [29–33] . Mechanisms involved in HCV particle morphogenesis and release , such as capsid protein maturation and lipid droplet targeting [34–36] , as well as p7 ion channel activity [37–39] are also shared by GBV-B and NPHV , although GBV-B appears to be uniquely endowed of a more complex and larger ( p13 ) ion channel protein [40] . Through the comparison of several mammalian hepaciviruses , the study undertaken here aimed at identifying conserved or divergent features of the hepaciviral life cycle that could provide new insights into the molecular mechanisms of HCV entry , replication and particle morphogenesis , as well as build bases towards the establishment of immunocompetent surrogate models that would ideally rely on closely-related rodent hepaciviruses . More precisely , the present study focuses on NS2 , which in HCV is a 217 amino acid ( aa ) transmembrane protein that carries a cysteine protease activity responsible for the cleavage of the viral polyprotein at the NS2/NS3 junction [41 , 42] and is essential for particle morphogenesis through mechanisms that remain incompletely understood [43 , 44] . NS2 proteolytic activity is a key step in HCV life cycle since NS2/NS3 cleavage is prerequisite for genome replication , NS2 function in particle morphogenesis and NS5A hyperphosphorylation [45–47] . However , two features of NS2 proteolytic mode of action remain intriguing . On the one hand , the crystal structure of HCV NS2 C-terminal catalytic domain revealed an unexpected dimeric protease with two composite functional active sites , highlighting a unique mode of action of this protease [48] . On the other hand , although HCV NS2 C-terminal domain contains the cysteine-based catalytic triad , NS2 was reported to exhibit very low intrinsic protease activity and to require NS3 N-terminal domain for efficient processing at the NS2/NS3 junction [41 , 42 , 49] . To gain insight into the properties of hepacivirus NS2 proteases , we carried out a comparative study of the proteolytic modes of action and regulatory mechanisms of NS2 from a selection of genetically divergent hepaciviruses . Following our recent work showing that GBV-B NS2 is a cysteine protease that shares common topological organization with HCV NS2 [22] , we report here that NS2 from equine , bat , rodent and Old World primate hepaciviruses are also cysteine autoproteases . By using an approach based on the coexpression of proteolytically inactive NS2-NS3 precursors in mammalian cells , we demonstrated that GBV-B and NPHV NS2 are also dimeric proteases with two composite active sites , highlighting the conservation and the functional significance of this peculiar mode of action among distantly related members of the Hepacivirus genus . In addition , we studied the role and virus specificity of NS3 N-terminal domain as NS2 protease cofactor in both transiently transfected and infected cells . Our data revealed unexpected substrate specificity and efficient intrinsic proteolytic activity of HCV NS2 that we further found to be conserved across hepaciviruses .
In order to determine and characterize the function of NS2 from various hepaciviruses , we initially carried out comparative sequence analyses of NS2 and of the N-terminal region of NS3 ( NS3N ) from selected strains of HCV , GBV-B and related viruses recently identified in horses ( nonprimate hepaciviruses , NPHV ) , bats ( bat hepaciviruses , BHV ) , rodents ( rodent hepaciviruses , RHV ) and Old World primates ( Guereza hepaciviruses , GHV ) . The phylogenetic analyses of these genomic regions illustrate the wide diversity of the Hepacivirus genus ( Fig 1A ) and demonstrate virus clustering that does not necessarily follow species-specific virus segregation . Virus clustering was similar whether NS2-NS3N ( Fig 1A ) , NS3 helicase ( S1A Fig ) , or NS5B RNA-dependent RNA polymerase ( S1B Fig ) aa sequences were considered . NS3 helicase and NS5B are the most conserved proteins across HCV genotypes and are generally used to address phylogenetic links among HCV strains [50] . Interestingly , the currently known strains of NPHV present a restricted genetic diversity , which is notably lower than the divergence between HCV genotypes . In contrast , wide sequence heterogeneity was observed among BHV and RHV strains , which appears unrelated to host tropism [11 , 14] . NPHV is HCV closest relative . BHV strains are also phylogenetically close to HCV with the exception of BHV PDB-112 isolate ( Fig 1A and S1 Fig ) . More precisely , NPHV NS2 shares approximately 39% aa identities and 71% aa similarities with HCV NS2 , while BHV NS2 shares 31% aa identities and 59% aa similarities with HCV NS2 ( S2 Fig ) . GHV , RHV and GBV-B NS2 are more distantly related to HCV NS2 , exhibiting only 20–24% aa identities and 47–52% aa similarities ( S2 Fig ) . Based on these phylogenetic analyses , we selected one representative hepacivirus strain for each mammalian species to perform a comparative experimental study of NS2 properties: HCV ( JFH1 ) , NPHV ( H3-011 ) , BHV ( PDB-452 ) , GHV ( BWC08 ) and RHV ( NLR07-oct70 ) ( Fig 1A ) . HCV and GBV-B NS2 contain 217 and 208 residues , respectively [22 , 40 , 41] , and the selected strains of NPHV , BHV , GHV and RHV are predicted to contain 217 , 217 , 212 and 199 residues , respectively , based on both predicted signal peptidase cleavage sites and sequence homology with HCV and GBV-B . The alignment of NS2 aa sequences from the selected hepaciviruses showed that NS2 N-terminal domains ( corresponding to HCV NS2 aa 1–93 ) are poorly conserved , whereas a greater degree of aa homology is found within the C-terminal domains ( Fig 1B and S2 Fig ) . In particular , His 143 and Cys 184 residues of HCV NS2 protease catalytic triad are fully conserved within hepaciviral sequences and the third negatively charged catalytic residue ( HCV Glu 163 ) is either identical ( NPHV , BHV and GBV-B ) or highly similar ( Asp for GHV and RHV , Fig 1B ) . In addition , a proline residue with a cis-peptide conformation at position 164 of HCV NS2 , which is suspected to have a critical role in NS2 catalytic activity by establishing the correct geometry of the Glu 163 side chain [48] , is fully conserved within hepaciviral sequences ( Fig 1B ) . NS3 N-terminal region ( NS3N , aa 1–213 in HCV ) presents a higher degree of homology across hepaciviruses than NS2 does , including the conservation of His , Asp and Ser residues forming the catalytic triad of NS3 serine protease in HCV ( Fig 1C and S2 Fig ) . NS3N and in particular a Zn2+ binding site within HCV NS3 protease domain that stabilizes NS3N structure was previously demonstrated to be critical for an efficient processing at the NS2/NS3 junction [49] . The HCV NS3 residues that coordinate Zn2+ atom ( Cys 97 , Cys 99 , Cys 145 and His 149 ) are either conserved among other hepacivirus sequences or possibly substituted by cysteine residues in the immediate vicinity for NPHV and RHV ( Fig 1C ) . Altogether , these observations prompted us to determine whether despite limited sequence homology , NS2 of the recently described hepaciviruses displayed proteolytic activity , as previously demonstrated for HCV and GBV-B NS2 [22 , 41 , 42] . Toward this aim , we designed a series of plasmid DNAs that allowed the transient expression of NS2-NS3N precursors from the 6 selected hepaciviruses . These polypeptide precursors were expressed downstream of the CD5 signal peptide in order to mimic ER translocation of NS2 N-terminus and their C-termini were fused to a twin Strep-tag ( ST ) for detection purposes ( Fig 2A ) . Proteins extracted from DNA-transfected 293T cells were separated by SDS-PAGE and ST-reactive products were detected by immunoblotting . For all recently discovered hepaciviruses , like for HCV and GBV-B , wild-type NS2 ( wt ) -NS3N-ST precursors were totally or partially cleaved to generate 23- to 28-kDa products , corresponding to NS3N-ST ( Fig 2B ) . The contribution of NS3 serine protease to these cleavage events could be ruled out since the conserved , presumed catalytic Ser residue of NS3 was mutated into Ala in all precursors ( Fig 2A ) . Moreover , the proteolytic events observed were abrogated when precursors containing mutated NS2 in which the Cys residue of putative catalytic triads was substituted for Ala ( CA ) were used . This led to the detection of products with higher molecular masses corresponding to uncleaved NS2-NS3N-ST precursors ( Fig 2B ) . For HCV , NPHV , GHV , RHV and GBV-B , the apparent molecular masses of NS2-NS3N uncleaved precursors were as expected according to coding sequences used ( S1 Table ) and polypeptide calculated molecular masses ( S2 Table ) . In addition , for NPHV , GHV and to a lesser extent for HCV , polypeptides with lower molecular masses were also detected . For BHV , the full-length NS2-NS3N uncleaved precursor was not detected , while a truncated polypeptide of approximately 40 kDa was identified ( Fig 2B ) . It is unlikely that these truncated products resulted from aberrant internal translation initiation since codon-optimized genes were used . These products may rather result from cleavage events by unidentified cellular proteases acting at cryptic sites located within NS2 N-terminal domains , as previously reported for the JFH1 strain of HCV [22 , 44] . It is , however , important to note that these additional cleavage events did not interfere with the analysis of NS2 proteolytic activity . The electrophoretic profiles of all NS3N-ST cleaved products were in agreement with calculated molecular masses ( S2 Table ) . Altogether , these data demonstrated that NPHV , BHV , GHV and RHV NS2 proteins , like HCV and GBV-B , exhibit cysteine protease activities responsible for the cleavage at the NS2/NS3 junction . HCV NS2 contains 3 transmembrane segments located within its N-terminal region and a C-terminal globular cytosolic domain [22 , 51] . Previous studies demonstrated that for HCV subtypes 1a and 1b , NS2 proteolytic activity is carried by its C-terminal subdomain , whereas the N-terminal transmembrane region is dispensable for NS2/NS3 junction processing [41 , 42 , 52] . Since HCV NS2 topological organization was demonstrated or predicted to be shared by GBV-B and recently discovered hepaciviruses , respectively , notably with respect to three N-terminal transmembrane segments [22] , we investigated whether NS2 N-terminal domain from HCV-related viruses was required for an efficient catalytic activity . Hepaciviral truncated precursors in which NS2 N-terminal region was deleted [ΔN ( NS2 ) -NS3N-ST] were generated ( Fig 3A ) . Immunoblotting of transfected cell extracts using ST-specific antibodies led to the detection of the 6 hepacivirus ΔN ( NS2 ) -NS3N-ST truncated precursors with apparent molecular masses of approximately 37 to 41 kDa , as expected ( Fig 3B and S2 Table ) . HCV JFH1 truncated precursor was efficiently cleaved ( Fig 3B ) , extending previous observations to HCV subtype 2a . Similarly , NPHV truncated precursor was cleaved at the NS2/NS3 junction ( Fig 3B ) with comparable efficiency as full-length precursor ( Fig 2B ) , indicating that NS2 N-terminal region is also fully dispensable for NPHV NS2 catalytic activity . GHV truncated precursor was cleaved , yet with decreased efficiency compared to the full-length precursor ( compare Figs 3B and 2B ) . In contrast , for BHV , RHV and GBV-B , the deletion of NS2 N-terminal region completely abrogated NS2/NS3 cleavage , indicating that NS2 N-terminal hydrophobic domain is required for the catalytic activity of these latter three viral NS2 ( Fig 3B ) . To address whether proper folding of the membrane-associated N-terminal segments may directly impact NS2 catalytic activity in some hepaciviruses , the proteolytic activity of NS2 from the most distantly related virus , GBV-B , was next compared to that of HCV NS2 in a cell-free expression system using wheat germ extracts [53] . Consistent with the absence of membrane-associated segment requirement ( Fig 3B ) , both full-length and truncated HCV precursors were autoprocessed in this acellular system as in cells , although with overall decreased efficiency ( compare Fig 3C with Figs 2B and 3B ) . While GBV-B truncated precursor was expectedly not cleaved ( Fig 3B and 3C ) , the full-length GBV-B precursor remained totally unprocessed in the cell-free system ( Fig 3C ) , in contrast to complete cleavage in transfected cells ( Fig 2B ) . We recently reported that the addition of 0 . 1% lauryl maltose neopentyl glycol ( MNG-3 ) detergent to wheat germ extracts greatly improved NS2 solubility and allowed the purification of functional HCV NS2 [53] , indicating that this detergent mimicks a membranous environment and supports proper folding of this transmembrane protein . By using wheat germ extracts supplemented with MNG-3 , we observed a cleavage of GBV-B full-length precursor , but not of GBV-B truncated precursor ( Fig 3D ) , demonstrating that GBV-B NS2 hydrophobic N-terminal region is required for the proper proteolytic activity of the C-terminal domain of NS2 , likely by contributing to critical authentic folding of the protein in a membranous environment . These data highlight structural differences for NS2 protease requirements among hepaciviruses . The crystal structure of NS2 protease domain from the H77 strain of HCV genotype 1a revealed NS2 dimerization with two composite active sites in which the catalytic His143 and Glu163 residues are contributed by one monomer and the third catalytic partner ( Cys184 ) by the other monomer [48] . These findings were surprising in the context of the early , presumably cis-acting cleavage event that NS2 has to carry out during the course of HCV life cycle . We therefore investigated whether NS2 protease dimerization was a unique feature of HCV NS2 or whether this peculiar mode of action was conserved among distantly related hepaciviruses . For this study , we selected three hepaciviruses , the JFH1 strain of HCV genotype 2a , as well as closely and distantly related hepaciviruses NPHV and GBV-B , respectively . Experiments were based on the coexpression of two mutated NS2-NS3N precursors containing an Ala substitution of either the Cys ( CA ) or the His ( HA ) residue of NS2 catalytic triads and C-terminally fused to a tag ( ST or V5 epitope ) ( Fig 4A ) . As expected , when expressed alone in cells , all ST- and V5-tagged wt precursors were efficiently cleaved at the NS2/NS3 junction ( Fig 4B–4D , lanes 1–2 ) , whereas NS2 proteolytic activity was abolished by the introduction of either CA or HA substitution ( Fig 4B–4D , lanes 3–6 ) . These results confirm the predicted catalytic role of NPHV NS2 residues His 143 and Cys 184 , as well as GBV-B NS2 residues His 138 and Cys 177 . Remarkably , the coexpression of CA- and HA-mutated precursors , both tagged with either ST or V5 resulted in partial NS2/NS3 cleavage ( Fig 4B–4D , lanes 7–8 ) . These results could only be explained by NS2 protease dimerization and the formation of two composite catalytic triads , with the His and Cys residues of a same active site being contributed by two different monomers . Using the crystallographic 3D structural NS2pro dimeric model described by Lorenz et al . [48] , the coexpression of CA- and HA-mutated precursors is indeed expected to lead to the formation of two catalytically inactive CA/CA and HA/HA homodimers , as well as a CA/HA heterodimer ( Fig 4A ) . Based on the putative composite nature of active sites , the CA/HA heterodimer is expected to form one doubly-mutated active site ( CA/HA ) and one native catalytic site capable of proteolytic activity ( Fig 4A , middle structure ) . To corroborate our data , we determined which of the CA- and/or HA-mutated NS2-NS3N precursor ( s ) was ( were ) cleaved by using one mutated precursor in C-terminal fusion with ST and the other in fusion with V5 . Interestingly , only HA-mutated precursors were partially cleaved , whereas CA-mutated precursors remained unprocessed , as shown by immunoblotting using ST- or V5-specific antibodies ( Fig 4B–4D , lanes 9–10 ) . These results indicate that for the three hepaciviruses studied , both NS2 C-terminal Leu residue and catalytic Cys residue originate from the same monomer , whereas catalytic His residue is contributed by the second monomer ( Fig 4A ) . Altogether , these data demonstrate that NS2 from the JFH1 strain of HCV genotype 2a , NPHV and GBV-B are dimeric proteases with composite active sites , fully supporting the model previously established for the H77 strain of HCV genotype 1a [48] . The conservation of NS2 dimeric mode of action between distantly related hepaciviruses , although unexpected for a cis-acting protease , underscores its functional relevance in the infectious hepaciviral cycle . Consistently , the homology models generated using the crystal structure of HCV 1a NS2 protease [48] revealed overall similar three-dimensional dimeric folds for NS2 protease domains of the various hepaciviruses , including the catalytic pockets ( S3 Fig ) . However , local variations , notably in the length of alpha helices that were shown to associate with membranes in HCV [54] , and in the length and orientation of the connecting loops were predicted , with RHV NS2 protease appearing to harbor the widest structural local divergence with respect to HCV and NPHV NS2 proteases . It is worth mentioning that the presence of N-terminal NS2 domains may affect the folding of NS2 C-terminal protease domains and be necessary to regulate or to allow their proteolytic activity in the diverse hepaciviruses ( Figs 2 and 3 ) . In order to further characterize the role of NS3N as HCV NS2 protease cofactor , we investigated whether NS3N from divergent hepaciviruses could functionally substitute for HCV NS3N or whether the stimulation of NS2/NS3 cleavage by NS3N was virus-specific . We thus generated chimeric precursors comprising HCV JFH1 NS2 fused to NS3N from NPHV , BHV , GHV , RHV or GBV-B and C-terminally tagged with ST ( Fig 5A ) . As shown in Fig 5B , the chimeric precursor containing NS3N from NPHV was fully cleaved ( lanes 3–4 ) , like parental HCV precursor ( lanes 1–2 ) . We detected a significant albeit partial cleavage of precursors containing heterologous NS3N from BHV and RHV ( lanes 5–6 and 9–10 ) . These data suggested that NS3N from NPHV , BHV and RHV appeared able to stimulate HCV NS2 protease with varying efficiencies . In contrast , NS3N from GHV and GBV-B were not able to significantly activate cleavage at the heterologous NS2/NS3 junction ( lanes 7–8 and 11–12 ) . Our data thus indicated that the activation of HCV NS2-mediated cleavage at the NS2/NS3 junction by heterologous hepacivirus NS3N could not be simply explained by overall genetic proximity with HCV , but might depend on the nature of specific residues within heterologous NS3N sequences . Interestingly , a recent study by Isken et al . [47] identified a hydrophobic surface patch in HCV NS3N that promotes HCV NS2 protease stimulation and is potentially critically involved in the coordination of replicase assembly . This hydrophobic patch involves aa residues at positions 3 ( Ile ) , 105 ( Tyr ) , 115 ( Pro ) and 127 ( Leu ) of NS3 ( Fig 6A ) . We generated homology models of NS3N from the various hepaciviruses considered in our study using the crystal structure of HCV genotype 1b NS3 protease previously reported [55] . Regardless of the genetic distances , structural models showed overall strikingly similar NS3N folds with relatively minor local structural changes ( Fig 6A and S3 Fig ) . With respect to HCV NS3N hydrophobic surface patch , the examination of hepacivirus NS3N sequence alignment revealed that residues equivalent to HCV Tyr105 are either Tyr or an aromatic residue ( Phe ) in other hepaciviruses considered in this study ( Fig 1C ) . Interestingly , HCV residue Pro115 is conserved in NPHV , whereas it aligns with the negatively charged residue Glu in GHV and GBV-B ( Fig 1C ) , i . e . in the two hepacivirus NS3N that were shown to be unable to stimulate HCV NS2 protease ( Fig 5B ) . In NPHV and GHV , the hydrophobic residue Ile at position 127 is similar to HCV Leu127 , whereas a cysteine residue is found at the equivalent position in GBV-B and a threonine residue in BHV and RHV ( Fig 1C ) . In order to evaluate the importance of the conservation of NS3N hydrophobic patch for NS3 cofactor activity , we engineered substitutions of residues at positions 105 , 115 , and/or 127 ( or equivalent positions ) in NS2 ( HCV ) -NS3N ( hepaci ) -ST precursors and examined autoprocessing of these mutated chimeric precursors using infrared fluorescent revelation of anti-ST immunoblots and quantification of NS2/NS3 junction cleavage efficiency . Dual substitutions of Tyr/Phe105 and Pro115 into Ala residues or single substitutions of Pro115 into Ala or Glu ( the latter as found in GHV/GBV-B ) introduced in HCV or in chimeric HCV-NPHV precursors resulted in statistically-significant , decreased cleavage efficiency ( 35–80% cleaved products , as compared to >90% in corresponding parental precursors , Fig 6B and 6C ) . These data fully support the results obtained by Isken et al . [47] , confirming that altering NS3N hydrophobic patch led to a loss in NS3N activating function in HCV . In addition , our results revealed that a similar region in NPHV NS3N is also important for NPHV NS3N cofactor activity . Remarkably , the converse substitution of Glu116 into either Pro , as found in HCV and NPHV , or Ala resulted in a substantial gain of function of GHV NS3N as HCV NS2 stimulating cofactor ( 55% and 25% cleaved products , respectively , as compared to <5% in corresponding chimeric GHV/HCV precursor , Fig 6D ) . It should be noted , however , that none of the substitutions designed to introduce HCV-like hydrophobic residues into GBV-B NS3N was sufficient to result in HCV NS2 protease stimulation ( Fig 6E ) . This highlights the greater genetic distance between GBV-B and HCV . Altogether , these data indicate that NS3N of some hepaciviruses are able to activate HCV NS2 protease and that this heterologous stimulation is dependent on NS3N hydrophobic surface patch . We investigated whether similar surface residues in NS3N of NPHV , GHV and GBV-B were important for the stimulation of cognate NS2 proteases . For this , we engineered substitutions of residues at positions 105 and/or 115 ( or equivalent positions ) of NS3N in NS2-NS3N-ST precursors of NPHV , GHV and GBV-B and examined autoprocessing of these mutated precursors as described above . Cleavage efficiencies were scored by quantification of product/precursor ratios using infrared fluorescent revelation of anti-ST immunoblots ( Fig 6F–6H ) . All mutations resulted in significant loss of NS2 protease activity to varying degrees for the three hepaciviruses . Importantly , these residues lie at a similar surface location in NS3N ( Fig 6A and S3 Fig ) as the hydrophobic patch previously defined for HCV [47] , strengthening the conserved role of this NS3N region for NS2 protease stimulation within the hepacivirus genus . Interestingly , sequence alignment of hepacivirus NS3 N-termini showed that the N-terminal five residues of HCV NS3 are highly conserved in NPHV and BHV sequences , and that the N-terminal two residues are identical in HCV , RHV and GBV-B NS3 ( Fig 5A ) . Since HCV NS2 fused to 2 aa of HCV NS3 was previously reported to support only basal proteolytic activity [49] , we undertook to determine to which extent the conservation of NS3N N-terminal residues contributed to cleavage efficiency at heterologous NS2HCV/NS3hepaci junctions . We generated precursors containing HCV NS2 followed by NPHV/ BHV NS3 aa residues 1–5 ( SPITA ) or by RHV/ GBV-B/ HCV NS3 aa residues 1–2 ( AP ) and C-terminally fused to green fluorescent protein ( GFP , 238 aa ) ( Fig 7A ) , and we quantified their autoprocessing using anti-GFP antibodies . Surprisingly , these 2 precursors were efficiently cleaved ( ~30–50% ) , although not to completion ( Fig 7B , lanes 1 and 3 ) . Cleavage was abrogated by inactivation of NS2 catalytic Cys residue ( Fig 7B , lanes 2 and 4 ) . This indicated that the fusion of the 5 SPITA or the 2 AP residues downstream of HCV NS2 was sufficient to permit NS2 autoproteolytic activity in the absence of NS3 moiety . Such an efficient cleavage was not anticipated since only an extremely basal proteolytic activity was previously reported for HCV NS2 followed by these few NS3 aa in a different expression system , whereas the N-terminal domain of NS3 was required for a productive processing at the NS2/NS3 junction [49] . Based on the unexpected efficient processing of NS2HCV-SPITA-GFP and NS2HCV-AP-GFP ( Fig 7B ) , we further examined the nature of residues located immediately downstream of HCV NS2 that may be permissive for NS2 proteolytic activity in the context of GFP C-terminal fusion . A precursor containing HCV NS2 directly fused to GFP remained unprocessed ( Fig 7C , lane 1 and Fig 7D , lane 1 ) . In contrast , the fusion of NS2 to GFP via a 5 aa flexible linker comprised of 4 Gly and a Ser ( NS2-4GS-GFP ) yielded substantial amounts ( ~50% ) of a GFP-reactive polypeptide with a molecular mass close to that of GFP ( Fig 7C , lane 2 and Fig 7D , lane 3 ) . This product was not observed upon mutation of NS2 catalytic Cys residue ( CA , Fig 7D , lane 4 ) . Such a productive processing was unexpected since the 4GS linker and the downstream GFP heterologous sequences do not present sequence homology with any hepacivirus NS3 N-terminal sequence ( Figs 5A and 7A ) . In order to determine the sequence requirements for NS2 catalytic activity in the absence of NS3N , we next examined the autoproteolytic activity of NS2 precursors fused to GFP via various short linkers selected to introduce flexibility [GGGGSGGGGS ( 2x4GS ) , GSAGS and SKSTS] or rigidity ( EAAAK and PAPAP ) to the polypeptide backbone [56] . The 10-aa ( 2x4GS ) and the 5-aa GSAGS flexible linkers both allowed efficient ( ~50% ) cleavage at NS2 C-terminus ( Fig 7C , lanes 3–4 ) . This cleavage was demonstrated to be mediated by NS2 catalytic activity since cleaved products were abrogated upon NS2 CA mutation ( S4 Fig ) . In contrast , the flexible SKSTS or the rigid EAAAK or PAPAP sequences were not appropriate substrates for NS2 catalytic pocket ( Fig 7C , lanes 5–7 ) . These data suggested that the composition and the flexibility of the sequence located immediately downstream of NS2 , but not necessarily its length were critical for NS2 proteolytic activity in the absence of NS3 . In particular , this set of data suggested that the presence of a glycine residue downstream of NS2 C-terminus may promote NS3-independent NS2 catalytic function . Such a glycine residue is not naturally found as the P'1 residue of the NS2/NS3 cleavage site in any reported sequence of hepacivirus . Interestingly , a proline residue is fully conserved as the P'2 residue of the authentic cleavage sites in all hepaciviruses examined . To obtain further insight into the importance of the first 2 residues downstream of NS2 for autoproteolytic cleavage , we engineered HCV NS2-linker-GFP precursors bearing GPGGS , APGGS , NPGGS or SPGGS as linkers . While the former two precursors were cleaved to similar efficiency as NS2-4GS-GFP , cleavage of the latter two precursors was significantly reduced ( S5 Fig ) . These data indicate that a small residue ( Gly/Ala ) is preferred at the P'1 position , whereas a proline at the P2' position has no effect in a linker-GFP fusion context . We subsequently analyzed the effect of the sequence located downstream of the 5aa linker on NS2 catalytic activity . For this , we used precursors containing HCV NS2 C-terminally fused to the 4GS linker followed by various tags , GFP ( 238 aa ) , Strep tag ( ST , 28 aa ) , or V5 epitope ( 14 aa ) ( Fig 7A ) . In order to be able to compare the cleavage efficiencies of these precursors , a HA epitope ( HAep ) was introduced at NS2 N-terminus and demonstrated to have no effect on NS2 protease activity ( S6 Fig ) . Immunoblotting using HAep-specific antibodies showed that the 3 precursors were cleaved with similar efficiencies in an NS2 protease-mediated manner , regardless of the nature of the C-terminal tag sequence ( Fig 7D , lanes 3–4 , 7–8 and 11–12 ) . Of note , the NS2-4GS-GFP precursor demonstrated consistent cleavage efficiencies , whether HAep-NS2 or GFP counterparts of cleaved products were quantified using anti-HAep or anti-GFP antibodies , respectively ( Fig 7C , lane 2 and Fig 7D , lane 3 ) . Furthermore , whereas precursors containing NS2 directly fused to GFP or ST remained unprocessed ( Fig 7D , lanes 1–2 and 5–6 ) , NS2-V5 precursor was efficiently cleaved ( Fig 7D , lanes 9–10 ) . This proteolytic processing in the absence of a linker sequence was probably due to the aa composition of the V5 epitope which N-terminal residue is a glycine ( see Fig 7A ) . Altogether , these results demonstrate for the first time an efficient catalytic activity of HCV NS2 protease in the absence of any NS3 sequence , revealing an unexpected intrinsic proteolytic activity of NS2 . We next assayed NS2-linker-tag fusion polypeptides in a composite protease assay using GFP and ST as tags , as previously performed with NS2-NS3-ST/V5 substrates ( Fig 4 ) . We found that co-expression of catalytically inactive NS2 ( HA ) -4GS-GFP fusion precursor together with catalytically active NS2 ( WT ) -4GS-ST or co-expression of two catalytically inactive linker-tag fusion precursors , e . g . NS2 ( CA ) -4GS-GFP and NS2 ( HA ) -4GS-GFP ( same tag ) or NS2 ( CA ) -4GS-ST and NS2 ( HA ) -4GS-GFP ( different tags ) led to partial cleavage on the monomer harboring the catalytic cysteine residue ( HA ) ( Fig 8A , lines 4–6 ) . Interestingly , we also documented partial cleavage of catalytically inactive NS2 ( HA ) -4GS-GFP precursor by native NS2-NS3-ST or by catalytically inactive NS2 ( CA ) -NS3-ST ( Fig 8B , lines 4–5 ) . Conversely , partial cleavage of NS2 ( HA ) -NS3-ST was also observed in the presence of catalytically inactive NS2 ( CA ) -4GS-GFP ( Fig 8B , line 6 ) . These data thus indicate that NS2 dimerization can also lead to NS2-mediated cleavage in the context of NS2 linker-tag fusion polypeptides . We investigated whether the NS3-independent substrate specificity of NS2 protease was conserved across HCV genotypes as well as more distantly related hepaciviruses . First , we compared the autoproteolytic properties of NS2-4GS-GFP precursors containing NS2 from the JFH1 or the J6 strains of HCV genotype 2a , the H77 strain of HCV genotype 1a or the Con1 strain of HCV genotype 1b ( Fig 9A ) . Anti-GFP immunoblots presented in Fig 9B showed that all wt precursors exhibited substantial to very efficient autoproteolytic activities carried out by NS2 cysteine protease . Similarly , NS2-4GS-GFP precursors containing NS2 from NPHV , BHV , GHV , RHV and GBV-B were all specifically cleaved by NS2 protease in the absence of NS3 ( Fig 9C ) . Altogether , these data indicate that NS3-independent NS2 protease activity is conserved across HCV genotypes and distantly related hepaciviruses , suggesting that this intrinsic activity is likely to have functional significance in the hepaciviral life cycle and/or hepacivirus/host interactions . HCV NS2 is not only an essential protease , but is also required for virion assembly and this function was shown to depend on both N-terminal transmembrane and C-terminal protease domains of NS2 , including its C-terminal leucine residue [46] . To further study the relevance of NS3-independent NS2 protease activity in HCV infectious life cycle , we generated a genome-length HCV cDNA derived from a cell culture adapted JFH1 variant [57] , in which the heterologous IRES sequence from the murine encephalomyocarditis virus ( EMCV ) was inserted between NS2 and NS3 coding sequences . This resulted in two translational units , therefore uncoupling NS2/NS3 processing from RNA replication as previously shown [44] and allowing the study of the proteolytic activity of NS2 upon its C-terminal fusion to heterologous sequences . In addition , the Firefly luciferase reporter sequence was introduced downstream of the EMCV IRES followed by the foot and mouth disease virus ( FMDV ) 2A peptide and the ubiquitin coding sequences in order to facilitate the monitoring of genome replication and infectivity ( Jad-2EIL3 , Fig 10A ) . Several cDNAs were engineered in the Jad-2EIL3 backbone and designed to encode NS2 with a HAep at its N-terminus ( between NS2 aa 1 and 2 ) and either C-terminally fused to a 4GS linker followed by GFP , ST or V5 ( Jad-2EIL3/HAep-NS2-4GS-tag ) or with no C-terminal fusion ( Jad-2EIL3/HAep-NS2 ) . Additionally , these cDNAs were created to encode NS2 protease with either a native ( wt ) or a mutated ( CA ) catalytic site designed to abrogate NS2-mediated cleavages . Human hepatoma Huh7 . 5 cells were transfected with Jad-2EIL3-derived synthetic RNAs . Proteins extracted at 72h post-transfection were analyzed by immunoblotting with NS2-specific antibodies . The three NS2 ( wt ) -4GS-GFP/ST/V5 precursors were specifically cleaved by NS2 protease at comparably high levels regardless of the nature of the C-terminal tag ( Fig 10B ) . Cleavage efficiencies in RNA-transfected cells were of the same magnitude as those observed in the transient expression system ( Fig 7D ) , indicating the relevance of NS2 intrinsic protease activity in both experimental systems . We further analyzed the effect of these fusions on HCV replication and particle assembly by quantifying luciferase reporter activities at 72h post-transfection in cells transfected with Jad-2EIL3-derived RNAs and at 72h post-infection in cells infected with supernatants collected at 72h post-transfection , respectively . All RNAs replicated efficiently , similarly to Jad-2EIL3 parent and in contrast to a replication-deficient Jad-2EIL3/GAA RNA encoding inactivating mutations in the polymerase active site ( Fig 10C ) . This was in agreement with the fact that the EMCV IRES insertion between NS2 and NS3 coding sequences rendered bicistronic RNA replication independent of NS2 proteolytic activity [44] . In contrast to an assembly-deficient RNA devoid of E1-E2-p7 coding sequences ( Jad-2EIL3/ΔEp7 ) , Jad-2EIL3 and Jad-2EIL3/HAep-NS2 RNAs encoding wt or CA-mutated NS2 proved equally infectious ( S7 Fig ) . This indicated that both the insertion of a HAep downstream of NS2 N-terminal residue and the Ala substitution of NS2 catalytic Cys residue did not alter NS2 function in particle morphogenesis , extending previous observations [43 , 44 , 51] . Remarkably , whereas the transfection of RNAs encoding NS2 ( wt ) -4GS-tag polypeptides led to high-level infectious virus production , the transfection of corresponding RNAs encoding CA-mutated NS2 did not yield TCID50-measurable infectious production ( Fig 10D ) . Thus , the lack of infectious particle production from RNAs encoding inactive NS2 could conclusively be attributed to the C-terminal fusion of NS2 to 4GS-tags , which impaired NS2 function in particle assembly . Conversely , wt NS2 autoproteolytic activity leading to the release of functional , C-terminally untagged HAep-NS2 was corroborated by efficient infectious virus production . Accordingly , the ~ 1 log unit decrease in infectivity observed with Jad-2EIL3/HAep-NS2 ( wt ) -4GS-tag RNAs compared to Jad-2EIL3/HAep-NS2 ( wt ) ( Fig 10D ) may be explained by the lower intracellular levels of assembly-competent HAep-NS2 , as HAep-NS2 ( wt ) -4GS-tag polypeptides were not cleaved to completion ( Fig 10B ) . In conclusion , these results concur to demonstrate that HCV NS2 substrate specificity initially unveiled in a transient expression system also operates in infected hepatoma cells , and that NS3N is not a mandatory cofactor for NS2 protease activity .
We recently experimentally demonstrated that GBV-B and HCV NS2 exhibit similar membrane topological organizations with three N-terminal transmembrane segments , which we modeled to be also shared by other recently identified mammalian hepaciviruses [22] . In this work , we show that NS2 from hepaciviruses of equine , bat , Old World monkey and rodent origins are cysteine proteases responsible for cleavage at the NS2/NS3 junction ( Fig 2 ) , as previously reported by us and others for GBV-B and HCV [22 , 41 , 42] . These results thus demonstrate that despite limited aa sequence similarity , NS2 from HCV and related mammalian hepaciviruses share both structural and proteolytic properties . We further show that the N-terminal hydrophobic domain of NS2 is required for the catalytic activity of BHV , RHV and GBV-B NS2 proteases , but not for that of HCV , NPHV and GHV NS2 ( Fig 3B ) . The characterization of GBV-B NS2 proteolytic activity in a cell-free expression system confirmed the critical role of GBV-B NS2 N-terminal domain interaction with a membrane-mimicking detergent ( Fig 3C and 3D ) . These data indicate that intramolecular interactions of NS2 N-terminal membrane-integrated domain and C-terminal protease domain contribute to the proper folding into a catalytically active domain , at least for a subset of hepaciviruses ( BHV , RHV , GBV-B; see NS2 C-terminal homology models in S3 Fig ) . Interestingly , although not required for NS2 proteolytic activity , interactions between NS2 N- and C-terminal regions have been shown to be critical for HCV NS2 function in particle morphogenesis [22 , 58] . Accordingly , for HCV and all related mammalian hepaciviruses examined here , interactions between NS2 N-terminal transmembrane region and C-terminal protease domain most likely play a critical role for NS2 proteolytic activity and/or NS2 function in viral particle assembly . An intriguing feature of HCV NS2 protease is that it is catalytically active as a dimer with two composite active sites , as previously suggested by size-exclusion chromatography analysis of re-folded HCV 2b-J8 NS2 [59] and further substantiated by HCV 1a-H77 NS2 protease domain crystal structure [48] . Importantly , the results presented here indicate that NPHV and GBV-B NS2 , like HCV 2a-JFH1 NS2 , also form homodimers with two composite active sites , with each catalytic triad composed of the His and Glu residues from one monomer and the Cys residue from the other ( Fig 4 ) . Although other virus-encoded or cellular dimeric proteases have been described , the formation of homodimers with two composite active sites is presently a unique feature of hepacivirus NS2 among currently characterized proteases . Indeed , herpes virus , HIV or caspase dimeric proteases form either two active sites with each site contributed by one monomer [60] or only one composite catalytic site [61 , 62] . The conservation of NS2 particular mode of action between HCV and both closely- ( NPHV ) and distantly- ( GBV-B ) related hepaciviruses suggests a critical role for NS2 dimerization in hepacivirus life cycle . An attractive hypothesis is that NS2 protease dimerization could regulate the kinetics of genomic replication initiation , which relies on NS2/NS3 cleavage for the release of nonstructural proteins ( NS3-NS5B ) comprising the active replicase . Consistent with NS2 dimerization , the rate of HCV NS2/NS3 cleavage is dependent on the concentration of NS2-NS3 precursor until reaching a plateau [59] . As a result , the requirement for minimal amounts of NS2-NS3 precursor could delay hepacivirus genomic replication . Of note , it was shown that rate-limiting NS2-mediated polyprotein cleavage translated into reduced replication competence at least for some HCV isolates , possibly as the result of kinetic differences in membranous replication complex biogenesis [63] . NS2/NS3 cleavage kinetics may also impact the intracellular levels of active NS3/4A protease , modulating the inhibition of innate immune sensing . Indeed , it is known that HCV NS3/4A protease plays a central role in innate immune evasion by cleaving MAVS and Toll-like receptor 3 adaptor protein ( TRIF ) , hence disrupting signaling pathways leading to type I interferon induction [31 , 32 , 64] . Interestingly , we and others showed that at least MAVS cleavage is conserved across NS3-4A proteases of various hepaciviruses [29 , 30 , 33] , suggesting that all hepaciviruses may have evolved similar kinetic regulation of innate immunity evasion strategies , potentially contributing to hepacivirus persistence and pathogenesis . The current dogma is that efficient NS2-mediated processing at the NS2/NS3 junction is dependent on the stimulation by NS3 N-terminal domain [49] . Conserved hydrophobic residues located at the surface of NS3N were recently shown to play an essential role in this process , as well as in replicase assembly [47] . We demonstrate here that NS3N domains of some but not all mammalian hepaciviruses are able to stimulate HCV NS2 protease ( Fig 5 ) and that this cross-species activation of HCV NS2 protease depends on the conservation or engineering of HCV-like hydrophobic residues in heterologous NS3N surface patches ( Fig 6 , S3 Fig ) . In addition , NS3N residues lying at a similar surface location also appeared important for NS3N-mediated NS2 protease activation in other nonhuman hepaciviruses ( Fig 6 ) . These results strengthen the role of NS3 surface patch for NS3 cofactor activity , indicating that interactions between specific NS3 residues and a yet unidentified region within NS2 could contribute to NS2 protease active site conformation in all hepaciviruses . In addition to this NS3-dependent proteolytic activity , a major finding of our study is the first-time report of HCV NS2 efficient intrinsic activity in the absence of any NS3 sequence . NS2 substrate specificity was characterized in the context of heterologous fusion of a linker-tag polypeptide to HCV NS2 C-terminus , in the absence of any NS3 sequence . We show that HCV NS2 protease tolerates small aa flexible linkers but not non-flexible linkers or larger P'1 residues ( Fig 7 , S5 Fig ) , indicating that the nature of the residues fused immediately downstream of NS2 C-terminus is critical for NS3-independent NS2 proteolytic activity . Of note , some of the permissive 5-aa linkers ( GGGGS , GSAGS , GPGGS , APGGS ) do not present any sequence homology with NS3 N-terminal residues ( APITA ) . A previous mutagenesis study showed that , although few nonconservative single aa substitutions severely impaired cleavage efficiency , the NS2/NS3 cleavage site of HCV genotype 1a H strain was remarkably resistant to mutation [65] . Our data ( Fig 7C and 7D , S5 Fig ) support these results and further indicate that NS2 is even more tolerant than expected since NS2 proteolytic activity is resistant to some favorable substitutions of all 5 residues directly fused to its C-terminus ( residues P1’ to P5’ ) . In addition , the nature and length of the tag fused downstream of the linker sequence did not impact NS2/linker-tag cleavage efficiency ( Fig 7D ) . This feature contrasts with the stimulatory effect of NS3 N-terminal domain on NS2/NS3 cleavage and with the critical role of NS3 surface patch discussed above . To conciliate these results , one hypothesis would be that NS3 N-terminus may rather have a negative impact on NS2/NS3 cleavage , which can be compensated by the positive effect of selected residues in NS3 protease domain , whose translation and folding may promote NS2 protease stimulation . In this context , the most conserved residue across hepaciviruses on the P' side of NS2/NS3 cleavage site ( a proline residue at the P'2 position ) did not appear to restrict NS2 protease in a linker-tag fusion context ( S5 Fig ) . Fig 11 summarizes our findings with respect to NS2 protease characteristics across various mammalian hepaciviruses , including ( i ) its activity responsible for NS2/NS3 cleavage , regulated by NS3 N-terminal domain ( particularly NS3 surface patch ) and dependent on NS2 dimerization , and ( ii ) its NS3-independent proteolytic activity . Importantly , the NS3-independent activity described here occurred efficiently in human hepatoma cells supporting virus production and was conserved among all mammalian hepaciviruses studied ( Figs 9 and 10 ) . This NS2 intrinsic activity may be exerted through protease dimerization involving precursors with different NS2 C-terminal fusions ( Fig 8 ) . It is tempting to speculate that the intrinsic NS2 proteolytic activity may have implications for HCV replication cycle and/or interference of HCV with host regulations . One attractive hypothesis is that HCV NS2 by virtue of its wider substrate specificity may be able to cleave cellular factors . Multifunctionality is a common feature among viral proteins including viral proteases such as HCV NS3 or poliovirus 2A that not only target viral polyproteins but also host cell factors such as MAVS and eIF4G , respectively [31 , 32 , 66] . However , HCV NS2 protease inactivation through the mutation of catalytic residues in the context of HCV bicistronic genomes in which an IRES was inserted between NS2 and NS3 sequences had no effect on viral particle production ( Fig 10D and S7 Fig ) , as previously reported [43 , 44] . This indicates that should they occur , such NS2-mediated additional proteolytic events are not required for the completion of HCV infectious cycle in vitro . In addition , although NS2 cleavage of host cell factors is an appealing hypothesis , it should be stressed that in the reported crystal structure of HCV NS2 protease domain [48] , NS2 C-terminal residue ( Leu 217 ) remains located within the active site , suggesting that NS2 catalytic site is locked after NS2/NS3 cleavage , which would prevent subsequent proteolytic events . Whether a so far unrecognized alternative post-cleavage conformation co-exists for HCV NS2 following displacement of the C-terminal residues would remain to be shown . Interestingly , a cellular J-domain protein ( Jiv ) , a member of the DNAJ chaperone family , has been shown to act as an activating cofactor of NS2 protease from a related pestivirus , bovine viral diarrhea virus ( BVDV ) . This interaction operates by promoting positioning of NS2 active site and substrate peptide into cleavage competent conformations for cis- and trans-cleavages [67] . A similar mechanism may exist for hepacivirus NS2 proteases . A high-throughput proteomic approach may allow to address whether NS2 could cleave nonviral substrates in trans and seek putative cellular substrates of NS2 protease , as was successfully done for HCV NS3/4A [68] . In conclusion , our results highlight important conserved features among mammalian hepacivirus NS2 proteases , some of which are , to our knowledge described for the first time: ( i ) hepacivirus NS2 are cysteine proteases acting as dimers forming two composite active sites responsible for the cleavage of the viral polyprotein at the NS2/NS3 junction , and ( ii ) NS2 proteases from HCV and related hepaciviruses exhibit an NS3-independent efficient proteolytic activity with defined substrate specificity . Our study also underlines a conserved , finely tuned regulation of the viral polyprotein proteolytic cleavage , which may have functional importance in hepacivirus infectious life cycle in vivo . Together with other studies pointing to important common characteristics across HCV and recently described hepaciviruses circulating in several mammalian species [11 , 13 , 16 , 18 , 30 , 35 , 37] , our data provide further support for the value of these hepaciviruses to be developed as surrogate , immunocompetent animal models of HCV infection [69] . Such models would be instrumental to investigate aspects of HCV-associated pathogenesis and to assess vaccine strategies toward global HCV eradication .
Viral strains considered in this study and the respective GenBenk accession numbers ( in parentheses ) are as follows: bat hepacivirus ( BHV ) PDB-112 ( KC796077 ) , BHV PDB-445 ( KC796091 ) , BHV PDB-452 ( KC796090 ) , BHV PDB-829 ( KC796074 ) , GBV-B ( AY243572 ) , Guereza hepacivirus ( GHV ) GHV-1 BWC08 ( KC551800 ) , GHV-2 BWC04 ( KC551802 ) , HCV 1a-H77 ( NC_004102 ) , HCV 1b-Con1 ( AJ238799 ) , HCV 2a-JFH1 ( AB047639 ) , HCV 2a-J6 ( AF177036 ) , HCV 2b-MD2b-1 ( AF238486 ) , HCV 3a-CB ( AF046866 ) , HCV 3b-TrKj ( D49374 ) , HCV 4a-ED43 ( NC_009825 ) , HCV 5a-SA13 ( AF064490 ) , HCV 6a-EUHK ( Y12083 ) , HCV 6b-Th580 ( NC_009827 ) , HCV 7a-QC69 ( EF108306 ) , nonprimate/equine hepacivirus ( NPHV ) B10-022 ( JQ434004 ) , NPHV G1-073 ( JQ434002 ) , NPHV H3-011 ( JQ434008 ) , NPHV NZP-1 ( JQ434001 ) , rodent hepacivirus ( RHV ) RHV- 339 ( KC815310 ) , RHV NLR07-oct70 ( KC411784 ) , RHV RMU10-3382 ( KC411777 ) , and RHV SAR-46 ( KC411807 ) . GBV-B NS2 sequences were amplified from pGBV-B/2 , which contains genome-length GBV-B cDNA [4] . HCV 1b-Con1 , HCV 2a-JFH1 and 2a-J6 NS2 sequences were amplified from plasmids pCMVNS2-GFP [51] kindly provided by D . Moradpour ( Centre Hospitalier Universitaire Vaudois , Lausanne , Switzerland ) , pJFH1 [70] kindly provided by T . Wakita ( National Institute of Infectious Diseases , Tokyo , Japan ) , and FL-J6/JFH-5'C19Rluc2AUbi [71] kindly provided by C . Rice ( The Rockefeller University , New York , NY ) , respectively . NS2-NS3N coding sequences from BHV PDB-452 , GBV-B , GHV-1 BWC08 , HCV 2a-JFH1 , NPHV H3-011 and RHV NLR07-oct70 ( S1 Table ) in which the Ser codon of NS3 established / putative catalytic triad was mutated into an Ala codon were sequence-optimized for expression in human cell lines and de novo synthesized ( GenArt , Life Technologies ) . Plasmids pCMV/NS2-NS3N-ST and pCMV/NS2-NS3N-V5 were modified from pCMV-KEB-GFP [72] by an overlapping-PCR mutagenesis strategy designed to introduce ( i ) the signal peptide sequence derived from the CD5 cellular gene , ( ii ) NS2-NS3N native or codon-optimized sequences from the selected hepacivirus , ( iii ) twin-strep-tag ( ST ) [WSHPQFEK- ( GGGS ) 3-WSHPQFEK] or V5 epitope [GKPIPNPLLGLDST] coding sequence and eliminate enhanced green fluorescent protein ( GFP ) coding sequence . Plasmids pCMV/NS2HCV-NS3Nhepaci-ST were similarly constructed such as to introduce HCV NS2 codon optimized sequence followed by NS3N codon-optimized sequences from the selected hepacivirus at step ( ii ) and ST at step ( iii ) . Derivatives of plasmids pCMV/NS2HCV-NS3Nhepaci-ST and pCMV/NS2-NS3N-ST ( NPHV , GHV , GBV-B ) encoding point mutations of NS3Nhepaci aa 105 , 115/116 , and/or 127 were generated by PCR-based site-directed mutagenesis of corresponding codons . Plasmids pCMV/ΔNS2-NS3N-ST were modified from the respective pCMV/NS2-NS3N-ST by site-directed mutagenesis in order to delete the following sequences ( nucleotide positions within NS2 sequence , using NS2 N-terminal boundaries defined in Fig 1B ) : ( 1–279 ) for HCV and NPHV , ( 1–264 ) for BHV and GHV , ( 1–246 ) for RHV and ( 1–258 ) for GBV-B . Plasmids pCMV/NS2HCV±linker-GFP and pCMV/NS2hepaci-4GS-GFP were modified from pCMV-KEB-GFP [72] by overlapping-PCR mutagenesis to introduce ( i ) CD5 signal peptide sequence , ( ii ) native ( GBV-B , HCV 2a JFH1 , HCV 2a J6 , HCV 1a H77 , HCV 1b Con1 ) or codon-optimized ( BHV , GHV , HCV , NPHV and RHV ) NS2 sequences and ( iii ) in appropriate cases , the sequence coding for a linker ( SPITA , AP , 4GS [GGGGS] , 2x4GS [GGGGSGGGGS] , GSAGS , SKSTS , EAAAK or PAPAP ) inserted immediately upstream of the GFP coding sequence . Plasmids pCMV/NS2HCV-4GS-ST and pCMV/NS2HCV-4GS-V5 were generated similarly from pCMV-KEB-GFP to introduce CD5 signal peptide sequence , HCV NS2 codon-optimized sequence , 4GS linker , and ST or V5 tag sequences in place of GFP coding sequence . Plasmids pCMV/NS2HCV-linker-GFP , in which the encoded linker is GPGGS , APGGS , NPGGS or SPGGS , were derived from pCMV/NS2HCV-4GS-GFP by overlapping , PCR-based site-directed mutagenesis . To generate pEU plasmids used for cell-free translation in the wheat germ expression system , JFH1 and GBV-B native sequences coding for NS2-NS3N-ST or ΔNS2-NS3N-ST were PCR amplified and cloned into the pEU-E01-MCS vector ( CellFree Sciences ) . In order to generate pJad-2EIL3 plasmid , pJad [22 , 57] was modified such as to insert by overlapping PCR the EMCV IRES cDNA followed by the Firefly luciferase reporter sequence , the FMDV 2A and the ubiquitin coding sequences between NS2 and NS3 coding sequences . Primer-based mutagenesis by overlapping PCR was next used to insert the HA epitope ( HAep ) sequence [YPYDVPDYA] downstream of HCV NS2 first N-terminal residue ( pJad-2EIL3/HAepNS2 ) and to generate derivatives of pJad-2EIL3/HAepNS2 encoding NS2 C-terminally fused to GFP , ST , or V5 sequences via a 4GS linker ( pJad-2EIL3/HAep-NS2-4GS-tag ) . For all plasmids described above , derivatives were generated by PCR-based site-directed mutagenesis in order to encode mutated NS2 catalytic triads in which the Cys or His codons were substituted by an Ala codon . PCR amplified DNA fragments from selected plasmid clones were checked by automated nucleotide sequencing using capillary electrophoresis ( Applied Biosystems ) . Further details of the cloning procedures can be provided upon request . N- and C-terminal boundaries of NS2 and NS3N domains from RHV , BHV , GHV and NPHV were predicted by determining signal peptidase cleavage sites using the SignalP 4 . 0 server ( http://www . cbs . dtu . dk/services/SignalP/ ) and examining sequence homology with HCV and GBV-B polyproteins . Protein sequence analyses were performed by using the Mobyle portal for bioinformatics analyses ( http://mobyle . pasteur . fr/ ) [73] . Multiple sequence alignments were performed with the T-coffee multiple sequence alignment program [74] using the corresponding web server facility ( http://www . tcoffee . org ) and aa identity or similarity percentages were calculated according to CLUSTAL W conventions . Phylogenetic trees were constructed using the neighbor joining method under the Jones-Thornton-Taylor model of aa substitution implemented in the MEGA6 program [75] and bootstrap resampling from 2 , 000 replicates was performed . Three-dimensional homology models of NS2 and NS3 proteases were constructed by the Swiss-Model automated protein structure homology modeling server [76] ( http://www . expasy . org/spdbv/ ) by using the crystal structures of HCV NS2 protease domain and NS3 as templates ( PDB entries: 2HD0 [48] and 1CU1[55] , respectively ) . Figures were generated from structure coordinates by using VMD [77] ( http://www . ks . uiuc . edu/Research/vmd/ ) and rendered with POV-Ray ( http://www . povray . org/ ) . Human embryonic kidney cells HEK 293T ( American Type Culture Collection ) were cultured in Dulbecco's modified Eagle's medium ( Invitrogen ) supplemented with 10% fetal calf serum , 100U/ml penicillin and 100 μg/ml streptomycin ( DMEM-10% ) , at 37°C in a 5% CO2 environment . Huh-7 . 5 human hepatocellular carcinoma cells ( Apath , LLC ) [78] , kindly provided by C . M . Rice , were cultured in DMEM-10% supplemented with nonessential aa and 1mM sodium pyruvate ( complete DMEM ) . Rabbit polyclonal antibodies NS2-1519 raised against HCV JFH1 NS2 synthetic peptides ( anti-NS2JFH1 ) [44] were kindly provided by R . Bartenschlager . Monoclonal antibodies specific for HCV-JFH1 NS3 , GFP ( JL-8 ) , V5 , ST , and HA were purchased from BioFront , Clontech , Invitrogen , Qiagen , and Sigma-Aldrich , respectively . HEK 293T cells were seeded at a concentration of 2x105 cells per well of 24-well plates and transfected 24 h later with plasmid DNA ( 0 . 1–0 . 8 μg ) using FuGENE 6 Transfection Reagent ( Promega ) , as recommended by the manufacturer . Genome-length pJad-2EIL3-based plasmids were linearized with MluI and served as templates for in vitro transcription using T7 RiboMAX Express Large Scale RNA Production System ( Promega ) according to the manufacturer's instructions . Synthetic RNAs were then purified by phenol-chloroform extractions , precipitated with isopropanol , resuspended in RNAse-free water , and stored at -80°C until use for cell transfection . RNA quality and quantity were monitored following 1% agarose gel electrophoresis and measurement of the absorbance at 260 nm . Huh-7 . 5 cells ( 2x106 cells ) were transfected by electroporation with 5 μg of in vitro transcribed genome-length RNA . Cells were electroporated in 4mm-gap width cuvettes by applying one pulse at 240 V , 900 μF ( EasyjecT Plus , Equibio ) , then immediately resuspended in complete DMEM and seeded at 2x105 cells per well in 6-well plates . Lysates were prepared using Reporter Lysis Buffer ( Promega ) at 72 h post-transfection from 6-well plates ( 0 . 2 mL ) or at 72 h post-infection from 12-well plates ( 0 . 1 mL ) that have been infected with supernatants collected from transfected cells at 72 h post-transfection . Ten-μL aliquots of lysates were assayed for firefly luciferase activity using 50 μL of Luciferase Assay Reagent ( Luciferase Assay System , Promega ) and a Centro LB 960 plate luminometer ( Berthold Technologies ) . Huh-7 . 5 transfected cell supernatants harvested at 72 h post-transfection were processed for infectivity endpoint dilution titration , essentially as described previously [22] with the following minor modification: 2 . 5x103 cells were seeded per well of 96-well plates and incubated for 5 days after infection prior to processing for infected foci revelation . Wheat germ cell-free protein expression and sample preparation were performed according to the small-scale bilayer method described previously [53 , 79] . Translation reactions were incubated at 22°C for 16h in 96-well plates . Expression in the presence of detergent was carried out by adding 0 . 1% MNG-3 in both the reaction mix and the buffered substrate solution . After protein synthesis , the translation reactions were incubated with benzonase on a rolling wheel for 30 min at room temperature ( RT ) , centrifuged at 20 , 000 g during 30 min at 4°C and the supernatant was purified using Strep-Tactin coated magnetic beads ( MagStrep type 2HC beads , IBA Lifesciences ) . Purified ST-containing proteins were eluted in 1X Laemmli sample buffer and further diluted in NuPAGE LDS sample buffer ( Invitrogen ) containing 0 . 71 M 2-mercaptoethanol for SDS-PAGE and immunoblot analysis . RNA or DNA transfected cells were lysed at 48 h and 32–40 h post-transfection , respectively , in NuPAGE LDS sample buffer ( Invitrogen ) containing 0 . 71 M 2-mercaptoethanol and heated for 10 min at 95°C . Proteins were loaded onto NuPAGE 10% or 12% Bis-Tris gels ( Invitrogen ) , separated by SDS-PAGE in MOPS SDS running buffer ( Invitrogen ) and transferred to PVDF or nitrocellulose membranes . Membranes were saturated in DPBS ( Invitrogen ) containing 0 . 1% Tween-20 ( PBS-T ) and 5% dry skimmed milk at RT for 1 h , prior to incubation for 1 h at RT or overnight at 4°C with one of the following monoclonal antibodies or antisera diluted in PBS-T containing 1% dry skimmed milk: anti-NS2JFH1 ( 1:2000 ) , anti-V5 ( 1:4000 ) , anti-ST ( 0 . 05 μg/ml ) , anti-GFP ( 0 . 125 μg/ml ) , or anti-HAep ( 0 . 25 μg/ml ) . Following incubation for 1 h at RT with peroxidase-conjugated ( chemiluminescence , GE Healthcare ) or Dylight 680 or 800nm-conjugated ( fluorescence , Thermo Scientific ) anti-mouse or anti-rabbit antibodies , proteins were visualized using ECL Prime Western Blotting Detection Reagent followed by exposure to Hyperfilm MP ( GE Healthcare ) or an infrared scanner imager ( Odyssey CLx , Li-Cor ) , respectively . Quantifications were performed on images acquired with the latter system using Image Studio Lite software .
|
Despite remarkable progress in the development of therapeutic options , more than 70 million individuals are chronically infected by hepatitis C virus ( HCV ) worldwide and major challenges in basic and translational research remain . Phylogenetically-related HCV homologues have recently been identified in the wild in several mammalian species , whose host restriction and potential for zoonosis remain largely unknown . We comparatively characterized the functions and properties of nonstructural proteins 2 ( NS2 ) from several animal hepaciviruses and HCV . We demonstrated that NS2 from animal hepaciviruses , like HCV NS2 , are cysteine proteases , which function as dimers with two composite active sites to ensure a key proteolytic event of the single viral polyprotein at the NS2/NS3 junction . In addition to the activation of HCV NS2 protease by NS3 N-terminal domain , our data revealed a novel NS3-independent substrate specificity and efficient intrinsic proteolytic activity of NS2 . The conservation of its properties and peculiar mode of action among distantly related hepaciviruses supports an important regulatory role for NS2 protein in the life cycle of these viruses . It also strengthens the value of animal , notably rodent hepaciviruses for the development of surrogate , immunocompetent models of HCV infection to address HCV-associated pathogenesis and vaccine strategies .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"crystal",
"structure",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"hepacivirus",
"pathogens",
"condensed",
"matter",
"physics",
"enzymology",
"microbiology",
"precursor",
"cells",
"viruses",
"rna",
"viruses",
"crystallography",
"enzyme",
"chemistry",
"polypeptides",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"solid",
"state",
"physics",
"sequence",
"alignment",
"bioinformatics",
"proteins",
"medical",
"microbiology",
"animal",
"cells",
"microbial",
"pathogens",
"hepatitis",
"c",
"virus",
"hepatitis",
"viruses",
"physics",
"biochemistry",
"peptides",
"cell",
"biology",
"flaviviruses",
"viral",
"pathogens",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"proteases",
"cellular",
"types",
"cofactors",
"(biochemistry)",
"physical",
"sciences",
"organisms"
] |
2018
|
NS2 proteases from hepatitis C virus and related hepaciviruses share composite active sites and previously unrecognized intrinsic proteolytic activities
|
The mechanism by which mice , exposed to the cold , mobilize endogenous or exogenous fuel sources for heat production is unknown . To address this issue we carried out experiments using 3 models of obesity in mice: C57BL/6J+/+ ( wild-type B6 ) mice with variable susceptibility to obesity in response to being fed a high-fat diet ( HFD ) , B6 . Ucp1-/- mice with variable diet-induced obesity ( DIO ) and a deficiency in brown fat thermogenesis and B6 . Lep-/- with defects in thermogenesis , fat mobilization and hyperphagia . Mice were exposed to the cold and monitored for changes in food intake and body composition to determine their energy balance phenotype . Upon cold exposure wild-type B6 and Ucp1-/- mice with diet-induced obesity burned endogenous fat in direct proportion to their fat reserves and changes in food intake were inversely related to fat mass , whereas leptin-deficient and lean wild-type B6 mice fed a chow diet depended on increased food intake to fuel thermogenesis . Analysis of gene expression in the hypothalamus to uncover a central regulatory mechanism revealed suppression of the Npvf gene in a manner that depends on the reduced ambient temperature and degree of exposure to the cold , but not on adiposity , leptin levels , food intake or functional brown fat .
Reduced ambient temperature will increase thermogenesis and reduce obesity . However its long-term effectiveness as a strategy to reduce obesity has been questioned because of the expectation that increased energy expenditure for the cold environment will increase food intake , thereby neutralizing the weight reducing effects of the cool environment [1] , a skepticism also associated with the effectiveness of physical activity as an anti-obesity strategy [2] . This skepticism emerges from the adipostat hypothesis itself , which predicts that reductions in fat mass by cold stimulation will be compensated by increased food intake to maintain its adiposity index [3] . On the other hand , studies on loss of fat mass by increasing thermogenesis with the chemical uncoupler dinitrophenol ( DNP ) showed that increased food intake does not necessarily occur [4] . Therefore compensation as predicted by the adipostat model may also not occur in association with BAT thermogenesis . Since chemical uncoupling by DNP , or even activation of thermogenesis by adrenergic receptor agonists [5] , are unregulated inductions of thermogenesis , compared to normal physiological mechanisms regulating body temperature , the problem of predicting the effectiveness of achieving energy homeostasis from food intake and endogenous energy reserves during cold exposure remains . Specifically , when an individual is exposed to a cold environment how the physiological decision is made to use endogenous energy reserves or to increase food intake and how this decision is influenced by the obese state of the individual is unknown . Although significant recent research progress has enhanced our understanding of the central control of BAT thermogenesis and energy expenditure in cold-exposed mammals , some areas are yet not well understood . In cold-exposed animals increased thermogenesis is associated with increased feeding , but is not accompanied by a gain of weight [6] . Coordinated increases in thermogenesis and food intake during cold exposure are controlled by signaling events in hypothalamus that are undefined . Within the hypothalamus , only a few genes are known to be differentially regulated in response to reduced ambient temperature [7–12] , but one cannot identify a clear pattern of neuropeptide expression characteristic for the hypothalamic response to the cold . The contribution of the selective neuro-hormone systems such as NPY or TRH in the regulation of cold-activated thermogenesis and feeding behavior has been extensively studied using pharmacologic approaches [13 , 14] or animal knockout models [15–17] . However , neither of these approaches identifies a critical molecule or describes signaling events that account for central mechanisms controlling energy availability and utilization under cold conditions . In this study , using wild-type B6 and brown fat deficient Ucp1-/- mice with DIO and genetically obese ( Lep-/- ) mice , we first determined that cold-induced thermogenesis is preferentially fueled by oxidation of fat reserves in individuals with environmental obesity and by food intake in lean individuals . We then analyzed global gene expression in the hypothalamus of cold-exposed mice and found that suppression of Npvf neuropeptide precursor mRNA levels occurred in the three models of obesity . To our knowledge Npvf is the only transcriptional target in hypothalamus known to be selectively regulated by changes in ambient temperature .
A wide range of body weight in genetically identical B6 mice results from their high natural variation in susceptibility to DIO [18] . We utilized this variation together with feeding mice a HFD for different lengths of time to generate mice with a range of adiposity . After 8 weeks or 1 week of feeding a HFD a cohort of mice was produced in which body weight ranged between 32 . 4 and 43 . 8g ( greater obese mice ) and between 24 . 4 and 32 . 7g ( lesser obese mice ) ( Fig 1A ) . Reducing the ambient temperature from 24 to 4°C resulted in an immediate lowering in body weight that was highest on day one and gradually diminished during the succeeding days ( Figs 1B and S1A ) . Although lesser and greater ( range of body weight ) obese mice showed the same response , the weight loss was larger in the greater obese group than the lesser obese group ( Figs 1B , 1C , S1A and S1B ) . Fat mass was the major endogenous substrate fueling thermogenesis ( Figs 1D and S1B ) . In the greater obese group , after 4 days at 4°C 97 . 5 kJ of energy came from fat mass and 33 . 8 kJ from fat free mass . For the lesser obese group , 30 kJ came from fat mass and 22 . 9 kJ from fat-free mass . Thus , 4 days of cold exposure resulted in total use of endogenous energy that equaled 131 . 3 kJ for the greater obese mice and only 52 . 9 kJ for the lesser obese mice ( Fig 1D and 1E ) . After one day at 4°C both groups of mice experienced a slight decline in body temperature ( 1–2°C ) , however , by the 2nd day at 4°C all mice were able to thermoregulate and maintain their body temperature at the level at which they started ( 36 ± 1°C ) . If the lesser obese group utilized less of their endogenous energy reserves during cold exposure than the greater obese , then where did the energy for thermogenesis come from ? For this we measured food intake . After 16 weeks on the dietary regime at 24°C , as described in the Methods , food intake was 56 . 5 ± 3 . 64 kJ/day for the lesser obese and 53 . 6 ± 2 . 29 kJ/day for the greater obese ( Fig 1F ) . When mice were transferred to 4°C , food intake immediately increased in the lesser obese mice to 75 kJ/day ( 35% increase ) and to 84 kJ/day ( 50% increase ) after 1 and 4 days , respectively; the increase in food intake was smaller in the greater obese mice going to 55 kJ/day ( 2% increase ) and 67 kJ/day ( 25% increase ) , respectively , after 1 and 4 days in the cold . With increasing time at 4°C the difference in food intake between the lesser and greater obese groups was reduced ( S1C Fig ) , consistent with the diminishing difference in fat mass . After the first day of cold exposure the difference in food consumption between mice from 2 cohorts equaled 20 . 06 ± 4 . 53 kJ , after 4 days at 4°C it was 14 . 44 ± 4 . 23 kJ ( Fig 1F ) and only 9 . 53 ± 2 . 87 kJ after 7 days at 4°C ( S1C Fig ) . Cold-induced thermogenesis is associated with increased consumption of fuel reserves and , as evident in Figs 1F and S1C , mice with lower endogenous fuel reserves compensate by increasing food intake , a process that apparently increases with time as endogenous fuel reserves become depleted . After 4 days at 4°C , regardless of the level of obesity present in the animals before cold exposure , cumulative energy coming from feeding and mobilized endogenous energy stores was comparable in the greater and lesser obese mice ( Fig 1G ) . For both lesser and greater obese animals linear regression analysis revealed a strong negative correlation between energy reserves ( fat and fat free mass ) mobilized per day and daily energy consumed during time spent in the cold ( R2 = 0 . 62 for greater obese mice and R2 = 0 . 75 for lesser obese after 4 days in the cold ) ( Fig 1H ) . An equally strong negative relationship was observed when values of adiposity index calculated for each mouse before cold exposure were plotted against daily food intake during 4 days at 4°C ( R2 = 0 . 74 for both greater and lesser obese mice ) ( S1D Fig ) . An important observation is that mice with robust DIO after 8 weeks on a high fat diet at 24°C will concurrently increase food intake and reduce body weight when transferred to an ambient temperature of 6°C ( S2A and S2B Fig ) . They will stabilize both body weight and food intake to a new state of energy balance to maintain body temperature . When they are returned to 24°C food intake returns to the level observed before the cold exposure and they resume the increase in adiposity characteristic of B6 mice . At 24°C there were no differences in the level of plasma free fatty acids ( FFAs ) and insulin between the groups ( Fig 2A and 2B ) . After 4 days at 4°C , greater obese mice had significantly elevated levels of circulating FFAs in comparison to lesser obese , consistent with increased fat mobilization in the greater obese animals . Substantial fat mass loss after 7 days of cold exposure resulted in reduced plasma FFAs in both groups of mice . Similarly , after 7 days at 4°C , circulating insulin was decreased in greater and lesser obese mice compared to 24°C ( Fig 2B ) . At 24°C leptin levels were positively correlated with adiposity ( Fig 2C ) . Leptin levels did not drop during the first 4 days at 4°C , only after 7 days in the cold did highly significant reductions in leptin levels occur ( Fig 2D ) . We evaluated the effects of leptin administration to DIO B6 mice fed HFD or lean B6 mice fed chow diet on the food intake and utilization of endogenous energy substrates before and after cold challenge . Leptin administration at 24°C decreased average daily food intake from 49 . 12±1 . 68 to 39 . 15±1 . 69 kJ in DIO mice and from 45 . 40±1 . 55 to 34 . 30±1 . 21 kJ , in chow fed lean mice ( Fig 2E ) . There was no effect of leptin administration on either body weight or body composition of mice at 24°C ( Fig 2E ) . When the ambient temperature was reduced from 24 to 4°C lean mice receiving leptin immediately increased food intake , whereas their body weight and fat mass did not change . On the other hand , cold-exposed and leptin-administered DIO mice immediately utilized endogenous reserves , then as these reserves diminished , they increased food intake ( Fig 2E ) . These results on food intake and fat utilization with leptin administration are not different from the phenotypes in the absence of exogenous leptin ( Fig 1B and 1F ) . Mice deficient in either leptin or the leptin receptor are cold intolerant when acutely exposed to 4°C; however , they are able to adapt to a lower temperature if the exposure is gradual [19–21] , thereby enabling an analysis of energy utilization during a cold challenge . Although there were large differences in body mass and composition between Lep+/ ? and Lep-/- mice fed a low fat chow diet at 24°C , after 9 days in the cold neither genotype showed significant changes in body weight mass nor composition ( Fig 3A–3C ) . With no reduction in endogenous energy reserves we looked to an increase in food intake . At 24°C average daily food intake was about 40% higher in leptin-deficient than in the Lep+/ ? control mice , as previously observed by Coleman [20] ( Fig 3D ) . One would anticipate that this source of energy would be used to fuel thermogenesis , however , reducing the ambient temperature by 3°C per day resulted in an increase in food intake in both control and Lep-/- mice . This food intake curve is displaced upward by an amount corresponding to the difference in food intake between control Lep+/ ? and Lep-/- mice at 24°C ( Fig 3D ) . Therefore , the rate of increase in food intake per degree Celsius reduction in ambient temperature by the control mice and Lep-/- mice was essentially indistinguishable ( Fig 3E ) . The most striking observation was that Lep-/- mice , already hyperphagic at 24°C , further increased their food intake under a cold challenge . After correcting for the slight changes in body composition that occurred in mice upon cold exposure , the total energy used for cold-induced thermogenesis was equal in leptin-deficient and control mice ( Fig 3F ) . Accordingly , there were no significant correlations either for control lean Lep+/ ? or for obese Lep-/- mice between the daily increase in food intake and endogenous body fuel reserves mobilized per day in the cold ( Fig 3G ) , in contrast to the significant correlations in DIO mice ( Fig 1H ) . It is assumed that non-shivering thermogenesis of brown fat is essential for providing the heat to protect the animal from the cold . Indeed Ucp1-/- newborn mice on either the B6 and 129 genetic backgrounds cannot survive the first days of birth in a breeding room maintained at ~23°C and Ucp1-/- adult mice acutely exposed to the cold at 4°C will succumb within 5 hours [22 , 23] . However , similar to Lep-/- mice , Ucp1-/- mice can adapt to the cold [24] . Ucp1-/- and Ucp1+/ ? mice were exposed to the cold using the same protocol as that used for Lep-/- mice , except that DIO was first induced at 24°C as with the greater and lesser obese mice ( Fig 1A ) . The level of obesity for the Ucp1+/ ? resembled that of the greater obese B6 . +/+ mice , whereas the Ucp1-/- mice resembled the lesser obese mice ( Fig 4A–4C ) , even though they were fed the HFD for the full 8 weeks . This is expected , since at 24°C Ucp1-/- mice are resistant to DIO [23] . At 24°C food intake was similar for mutant and control mice , whereas the daily energy intake during cold adaptation was higher for Ucp1-/- mice ( Fig 4D ) . Similar to the results of the initial experiment with wild type B6 mice , Ucp1+/ ? mice which had the greater obese phenotype preferentially lost fat mass during cold adaptation , whereas the Ucp1-/- mice which had the lesser obese phenotype preferentially increased food intake ( Fig 4E ) . Thus , energy balance and substrate utilization in DIO Ucp1-/- mice during cold exposure resembles that of lesser obese wild-type mice . In summary , UCP1-dependent brown fat thermogenesis is not required to derive the weight reducing benefits of adapting to the cold and there is no mechanism associated with thermogenesis that will increase food intake of the greater obese to preserve the obese state . There is a mechanism , however , to preserve a minimal adiposity index typified by young adult C57BL/6J mice fed a low fat chow diet . Total energy consumption as shown by 6 experimental groups ( Figs 1G , 3F and 4E ) indicates that energy expenditure during cold exposure is generally similar , except that Ucp1-/- mice are metabolically inefficient and have higher O2 consumption per mouse [25] . The difference among groups describes source of energy for the induction of thermogenesis , endogenous reserves vs food intake , and it is this difference which is the focus of this study . At 24°C Lep-/- mice are hyperphagic compared to the Lep+/+ or Lep+/- mice ( Fig 3D ) . Reducing the ambient temperature from 24 to 6°C was accompanied by a graded parallel increase in food intake , corresponding to approximately 50 kJ of energy for both control and mutant mice ( Fig 3D ) . Consequently , the same leptin-independent increase in food intake was observed during the transition from 24 to 6°C in both Lep+/+ and Lep-/- . Since the energy content of Lep+/+ and Lep-/- mice was unchanged during cold exposure , thermogenesis is fueled solely by food intake . Accordingly , we predicted that the same changes in gene expression associated with the central regulation of thermogenesis by the hypothalamus must occur in both Lep+/+ and Lep-/- mice during the transition from 24 to 6°C . Microarray analysis of gene expression was performed on hypothalamic tissue dissected from Lep-/- and Lep+/+ mice kept at different temperature conditions , that is , in mice maintained at 24°C ( point A , Fig 3D ) and in mice in which the ambient temperature had been reduced to 6°C ( point B , Fig 3D ) . We identified a small subset of genes in Lep-/- in common with Lep+/+ mice during the transition from 24 to 6°C ( Fig 5A ) . Among these genes , neuropeptide VF precursor ( Npvf ) , showed a robust down-regulated expression of 4 . 0 and 3 . 5 fold in the hypothalamus of cold-exposed Lep-/- and Lep+/+ , respectively . A group of genes encoding for G protein-coupled receptors ( GPCRs ) including the dopamine receptor D1 ( Drd1a ) , adenosine receptor 2A ( Adora2a ) , GABA ( A ) receptor subunit delta ( Gabdr ) and Gpr88 as well as some of their downstream targets including cAMP-regulated phosphprotein 21 ( Arpp21 ) and protein phosphatase 1 regulatory subunit 1B ( Ppp1r1b ) were up-regulated 1 . 4 to 3 fold in both Lep+/+ and Lep-/- mice following cold exposure . Cold exposure also increased the expression of antidiuretic hormone arginine vasopressin ( Avp ) gene in both mutant and wild-type animals by 1 . 8 and 1 . 4 fold , respectively . Each of the genes expressed in parallel in Lep+/+ and Lep-/- mice were validated by qRT-PCR ( Fig 5B ) . To further investigate a potential role for Npvf in food intake as a function of cold , we determined its expression in the hypothalamus of mice with different levels of dietary-induced obesity following cold exposure ( Figs 1A and S1A ) . Similar to the experiment with Lep+/+ and Lep-/- mice , Npvf expression was suppressed in both greater and lesser obese mice after the temperature shift from 24 to 4°C , but its expression was not associated with either adiposity or food intake ( Fig 6A ) . Increasing the duration of cold exposure at 4°C from 1 to 7 days gradually amplifies the reduction in Npvf mRNA levels . In an independent experiment DIO mice that were maintained at 4°C for 14 days and then returned to 24°C for 25 days restored their levels of Npvf mRNA to that initially observed at 24°C ( Fig 6A ) . Npvf mRNA expression in hypothalamic tissue showed a positive correlation with ambient temperature . Mice kept for 14 days at thermoneutrality ( 29°C ) had higher expression of Npvf mRNA in hypothalamus than mice maintained at 24°C . Similarly , 2 weeks at 17°C resulted in a reduction of mRNA expression to levels below that observed at 24°C ( Fig 6B ) . Although modulation of Npvf precursor mRNA occurs during cold-stimulated thermogenesis , an involvement of Npvf in the regulation of non-shivering thermogenesis in brown fat is unlikely . Down-regulation of Npvf mRNA was not influenced by the absence of UCP1 protein ( Fig 6C ) . Acute exposure to the cold requires an immediate response for heat generation and leads to immediate UCP1 production in BAT and WAT . A separate experiment performed to illustrate time-course of changes in the expression of Npvf under low temperature conditions showed that significant suppression in the amount of Npvf mRNA does not occur before 12h at 4°C; a significant decrease in the accumulation of Npvf mRNA in hypothalamus is found after 24h at 4°C compared to 29°C ( Fig 6D ) . Moreover , one week administration of β3-adrenergic agonist CL 316 , 243 ( 1mg/kg of body weight ) did not result in the suppression of Npvf mRNA in hypothalamus compared to saline-treated control mice ( Fig 6E ) , providing evidence that changes in expression of the Npvf gene are not linked to heat production or brown adipocyte induction in peripheral β3-AR-expressing tissue targets . Genes associated with food intake in the hypothalamus , thermogenesis in the adipose tissue , and lipid metabolism in the liver and adipose tissues were analyzed by qRT-PCR . No patterns in gene expression could illuminate mechanisms associated with the phenotypes described above ( see Supplement; S3A and S3B Fig for thermogenic genes in iBAT and iWAT , S4A and S4B Fig for neuropepetides of feeding behavior , and S5A Fig for genes of fatty acid metabolism in the liver , S5B Fig in iBAT and S5C Fig in iWAT ) .
We show that the total energy expended by a mouse from food intake and endogenous energy reserves to sustain thermogenesis during cold exposure is independent of the degree of obesity in the animals . This is true in genetically obese Lep-/- mice , chow-fed wild-type mice ( Fig 3F ) and in wild-type mice and B6 . Ucp1-/- with variable levels of DIO ( Figs 1G and 4E , respectively ) . However , in chow-fed mice the energy that is necessary to sustain a thermogenic program to maintain body temperature in the cold comes exclusively from feeding , as observed by others [20 , 26]; whereas in a wild-type mouse with diet-induced obesity induced by a high-fat diet , the fuel to support thermogenesis is obtained from endogenous energy reserves ( mostly fat ) and food intake . In DIO mice the source of energy required to maintain body temperature during cold exposure is determined by the degree of obesity . In DIO mice the energy reserves in fat mass are not privileged or restricted as those in a normal wild-type mouse maintained on a low-fat chow diet , rather they are utilized in proportion to their absolute levels . DIO mice with the highest levels of stored fat immediately mobilize fat , subsequently as these reserves become depleted , food intake becomes progressively a larger contributor to the fuel mix . In contrast , those mice that are at the other end of the DIO spectrum , the lesser obese mice , will preferentially increase food intake and use less of their endogenous fuel reserves to support thermogenesis . An important finding is that wild type mice with high levels of adiposity behave in response to cold exposure by the utilization of available energy sources in a manner that is independent of hormonal status , i . e . leptin and insulin . Serum leptin levels measured before cold exposure indicated that leptin resistance should have been higher in the greater obese mice than in the lesser obese , predicting a defense of adipose stores and higher food intake in greater obese mice . However , from the very beginning of cold exposure a defense of the fat status in mice with leptin levels predictive of leptin resistance was not observed . In fact the opposite was observed , with food intake reduced and fat utilization increased in the greater obese mice compared to the lesser obese . Interestingly , our observations on body composition-dependent differential fuel selection occurring during cold exposure in DIO mice parallels findings in exercising human subjects ( 32 ) . Moderate to intense physical activity performed regularly and on a long-term basis by lean individuals is compensated for by a corresponding change in food intake while body mass is maintained . On the other hand , obese individuals with excess fat storage do not significantly increase food intake and loss of body fat occurs as a consequence [27] . A return of mice fed a HFD from 6 to 24°C leads to a decrease in food intake and increase in adiposity characteristic of their phenotype on a high fat diet ( S2A and S2B Fig ) . We previously observed the same response of mice fed a high fat diet when energy balance was interrupted with food restriction [18] . Accordingly , mice do not assume increased levels of food intake transiently acquired when they are in the cold , rather food intake is set by the requirements for heat production as originally hypothesized by Brobeck [28] . The long-term defense of body weight in humans and mice has been described and discussed as a consequence of under- and over-feeding [29]; however , mechanisms associated with a negative energy balance resulting from reduced energy intake during dieting may be different from increased energy expenditure in response to cold-induced energy expenditure , since the latter condition is supported by increased food intake and the neutralization of insulin and leptin resistance [2 , 30] . Wild-type B6 ( Lep+/- or Lep+/+ ) mice fed a low fat chow diet exhibited almost no change in endogenous energy reserves , that is , lean mass or fat mass when the ambient temperature was reduced from 24 to 6°C , but they increased food intake . This observation fits with the thermostatic theory proposed by John Brobeck in the late 1940s , which relates the regulation of body temperature to the control of feeding behavior [28] . Brobeck summed up his theory by saying: “…animals eat to keep warm and stop eating to prevent hyperthermia” . In the present study , the wild-type B6 mouse maintained energy balance and body composition on a normal diet , when exposed to the cold , by increasing calorie intake . Importantly , the Lep-/- mouse behaved in the same manner , it adapted to the cold by increasing food intake in a manner quantitatively indistinguishable from the normal B6 mouse and it preserved its endogenous energy reserves . The β-oxidation of fat stores of Lep-/- mice is not an option for fuel to maintain body temperature [31] and this is a major factor in cold intolerance of leptin-deficient mice during acute exposure [32] . Lep-/- mice sensed that existing fat stores were unavailable and compensated by increasing food intake in a leptin-independent manner . This feeding behavior in the cold underscores the inability of mice with leptin-deficiency to utilize endogenous fat reserves; furthermore it also shows that in the face of a cold challenge fuel for thermogenesis must come from food intake . On the other hand , the wild-type mouse on a low fat chow diet can access its energy reserves in an acute situation , but quickly turns to increased food intake to maintain energy balance . This similarity in the metabolic response to the cold environment between normal and leptin-deficient mouse suggests that leptin is not important for the acute thermogenic phenotype in the Lep-/- mouse , nor for the regulation of food intake during cold exposure by normal wild-type mice fed a chow diet . Mice with mutations to leptin and the leptin receptor have a thermogenic phenotype in which body temperature drops about 10°C in about 4 hours at an ambient temperature of 4°C [21]; however , as illustrated in Fig 3D they can adapt to the cold when it is gradually reduced . A key feature of cold-induced thermogenesis in normal animals is the increase in food intake that occurs over and above the increase in food intake necessary to support nutrition [33–36]; as exemplified by the remarkable boost in food intake that occurs in lactating females exposed to the cold [26] . This suggests that central mechanisms controlling food intake , as related to nutrition , growth and body composition , may be independent of those associated with cold-induced thermogenesis . A similar idea has been put forth by Speakman and Krol [37] , but with a necessary role for leptin in the cold-induced food intake , which we did not see , nor was a role for leptin proposed by Melnyk and Himms-Hagen [6] . We had observed previously , as did Coleman [20] , that Lep-/- mice exposed to the cold further increased food intake above that normally occurring in these mice [19] . This preliminary observation has been extended in this study to show that this hyperphagia , which is above that normally occurring in Lep-/- mice fed a chow diet , is very similar in magnitude and kinetics to that occurring in Lep+/+ mice . Accordingly , mechanisms controlling cold-associated food intake in Lep-/- mice are independent of leptin-based regulation of food intake . We tested further the role of leptin in regulating thermogenesis during cold exposure in wild-type DIO mice . Plasma leptin and insulin levels in DIO mice of this study are remarkably similar to mice described in previous studies that were leptin resistant [38] . If the mobilization of fuels for cold-induced thermogenesis in DIO mice is controlled by the leptin resistance at the time of cold exposure , then one would predict that food intake would be high and mobilization of endogenous fat stores would be low . However , within one day of exposure to the cold the opposite phenotype was observed in DIO mice: food intake was low and fat mobilization was high . Even 4 days after cold exposure plasma levels of leptin were not significantly different from those at 24°C; only after 7 days in the cold were the levels of leptin significantly reduced ( Fig 2D ) . Additional leptin administered intraperitoneally to DIO and lean mice did not affect the observed pattern of energy substrate utilization in the cold ( Fig 2E ) . This data additionally suggests that the mechanism controlling food intake during acute cold exposure is independent of leptin signaling . Chronic cold adaptation may involve leptin by another mechanism [19] . The primary motive driving this study was to explore the feasibility of using cold exposure as an anti-obesity strategy . Human studies on brown fat show dramatic inter-individual differences in brown adipocyte content and BAT activity [39 , 40] . Thus , it is important to assess how the cold-stimulated effect of body weight reduction is influenced when the capacity for thermogenesis in brown fat is variable . The extent to which BAT-mediated adaptive thermogenesis could account for variability in substrate utilization in the reduced ambient temperature is also not known . For these reasons we evaluated the phenotype of DIO Ucp1-/- mice lacking functional brown fat . Ucp1-/- mice are sensitive to the cold; however , they can adapt to the cold if the ambient temperature is gradually reduced [24 , 41] . Therefore , if UCP1 is essential to the thermogenic process , then in its absence the capacity for heat production from brown fat would be severely suppressed and we could expect effects on food intake and or the utilization of endogenous fuels that would differ from the wild-type mouse . As expected , when the ambient temperature was reduced average food intake was higher in the Ucp1-/- mice than in control mice , because these mice are less obese when fed a high-fat diet and they burned less of their endogenous reserves compared to normal Ucp1+/ ? mice with the greater obese phenotype ( Fig 4B and 4D ) . Thus , there does not seem to be any difference in the pattern of utilization of endogenous food reserves or food intake between UCP1-deficient and wild type mice , provided that these mice have similar adiposity phenotypes as occurs with the lesser and greater obese mice . QRT-PCR analysis of the expression of several genes in the hypothalamus encoding neuropeptides implicated with food intake did not provide evidence for the involvement of any of the neuropeptides associated with food intake with the possible exception of CART and POMC which have expression reduced by 30% and 50% in Lep+/+ only . However , a microarray analysis of gene expression in Lep-/- and Lep+/+ mice at 24 and 6°C showed that the expression of neuropeptide VF precursor was decreased 4-fold during cold exposure , and a similar level of down-regulation for this gene was observed for all three of the genetic models we have studied . In rodent brain , the sequence of the Npvf precursor gene predicts two–RFamide peptides: RFRP-1 and RFRP-3 , also named NPSF and NPVF [42 , 43] . There is an expanding body of evidence for a role of various–RFamide peptides in the modulation of nociception , hormone secretion , reproduction or blood pressure [44–46] . Finally , although little is known of the functional significance of this particular biological effect , various–RFamide peptides were able to illicit a transient 10–300% induction or suppression of food intake in chicks , rats or mice after i . c . v . injection [44 , 47–50] . Moreover , food restriction or deprivation , both stimulating hunger and food hoarding , have been shown to be positively correlated with activation of RFRP-3 cells in the DMH of Syrian hamsters [51] . Effects on thermogenesis are unknown . From a functional viewpoint , specific expression of Npvf mRNA in the rodent central nervous system is restricted to a population of neurons localized between dorsomedial hypothalamic ( DMH ) and ventromedial hypothalamic ( VMH ) nucleus [42 , 43 , 52 , 53] , which is consistent with a putative role in feeding or thermogenic processes [54] . Our observations on lack of association between leptin status and regulation of Npvf mRNA in the cold demonstrate that Npvf system in hypothalamus is unlikely to be leptin responsive , which is in agreement with a recent study , where no evidence for leptin signaling after leptin injection or detection of leptin receptors in RFRP3 expressing neurons was found in mice hypothalamus [55] . Thus , the reduction in Npvf expression at lower temperature when a higher level of energy expenditure ( EE ) is required suggests that reduction of Npvf releases a brake on EE . Although increased food intake provides the fuel for the increase in EE in lesser obese mice , endogenous fat provides the fuel in greater obese mice . Since Npvf is similarly suppressed in both lesser and greater obese mice , neither endogenous substrate or food intake per se are the signals associated with Npvf expression levels . Lack of association between body energy reserves and hypothalamic Npvf expression was also shown in the recent study in which no significant difference in Npvf mRNA was detected between mice fed high-fat and low-fat diet for 20 weeks [55] . A motive for conducting our experiment was to establish in a mouse model the effects of cold on substrate utilization and long-term effects of cold exposure on food intake after a return to ambient temperature . This study clearly showed that upon cold exposure obese mice fuel their increase in energy expenditure with endogenous fat supplies , whereas lean mice increase food intake . An analysis of three mouse models of obesity suggests that reduced ambient temperature is effective in reducing diet-induced obesity without long-term compensatory increases in food intake . Whether humans will behave in a similar manner needs to be determined . The second part of our study uncovered evidence for a new hypothalamic signaling pathway , involving the Npvf gene , that is regulated in cold-activated thermogenesis . We will work towards determining whether a similar signaling pathway is present in humans .
Breeding pairs of C57BL/6J . +/+ , C57BL/6J . Ucp1+/- and C57BL/6J . Lep+/- mice were obtained through the generosity of Dr . Martin Klingenspor of the Technical University of Munich , Germany . All procedures concerned with breeding , housing , maintenance and experimental treatment of the mice were approved by the Local Animal Care and Use Committee for University of Warmia and Mazury , Olsztyn . Guidelines for animal experiments followed EU Directive 2010/63/EU . The goal of this protocol was to generate a series of mice with a range of adiposities by use of a high-fat diet to determine the effects of cold exposure on changes in food intake and endogenous energy stores . Breeding pairs of C57BL/6J+/+ mice were housed at standard temperature ( 24±1°C ) and maintained in ventilated rooms under a standard-day photoperiod ( 12:12-h light-dark period , lights on from 0700 to 1900 h ) with free access to low-fat diet ( PicoLab Rodent Diet 20 , LabDiet 5053 , 11 . 9 kcal % fat ) and water . At 21 days of age male progeny were weaned and housed in groups of 3–5 in plastic cages with fresh sawdust bedding . Body weight and body composition by NMR ( Bruker , BioSpin , Germany ) were monitored until mice were 8 weeks of age , at which time mice were individually housed and divided into two nutritional groups matched for similar mean body mass and body fat content to form the lesser and greater obese groups . By 8 weeks of age , when mice were still on a low-fat diet , their body weights ranged from 19 . 2–25 . 5g ( Fig 1A ) . Based on the NMR analysis , the distribution in body weights was mainly caused by differences in fat mass , with only a small contribution from fat free mass The greater obese group was fed a high-fat diet ( AIN-76A with 33% hydrogenated coconut oil , 58 kcal % fat ) from the 8th to 16th week to establish a range of mice with a higher adiposity index . The lesser obese would continue to be fed the low-fat diet ( PicoLab Rodent Diet 20 , LabDiet 5053 , 11 . 9 kcal % fat ) for 7 weeks and then the high fat diet for the 16th week . Mice with body weights and fat mass ranging from 24 . 4 to 43 . 8g and 3 . 7 to 18 . 8g , respectively , were formed ( Fig 1A–1D ) . The aim of such a dietary intervention was to establish as broad a range in adiposity as possible between two groups of mice and at the same time to induce metabolic adaptations associated with high-fat feeding in the lesser obese mice at the time of cold exposure . In addition to having mice with a variation in adiposity , food intake of the lesser and greater obese mice was measured for the 16th week , just prior to being exposed to the cold . Food intake , initially varied in the lesser obese mice when presented with a highly palatable high-fat diet for the first time; however , on the last day before cold exposure , there was no significant difference in food intake between the lesser and greater obese mice ( S1E Fig ) . In order to determine the relative contribution of food intake and endogenous energy reserves to fuel cold-induced thermogenesis , individually housed mice in the lesser and greater obese groups of mice were transferred to a cold room at 4°C for either 4 or 7 days . Food intake ( high-fat; AIN-76A ) and body weight were measured daily and body composition was analyzed by NMR at the end of the cold exposure . To calculate daily energy expenditure in the cold coming from endogenous and exogenous energy sources , fat mass and fat free mass measured after cold exposure were subtracted from fat mass and fat free mass measured before cold exposure and divided by number of days spent in the cold; average daily food intake measured at 24°C was subtracted from average daily food intake measured at 4°C . Energy values in kJ for g of fat mass or fat free mass were calculated as follows: 4 . 18 kJ/kcal × ( 9 or 4 kcal/g , respectively ) . Energy values in kJ for g of low-fat chow diet or high-fat diet were calculated as follows: 4 . 18 kJ/kcal × ( 3 . 07 or 5 . 44 kcal/g , respectively ) . To observe the effects of cold-induced hyperphagia on DIO mice that were returned to an ambient temperature of 24°C , adult C57BL6/J+/+ mice were fed a high-fat diet ( AIN-76A ) from 8 to 16 weeks of age then transferred to 4°C until food intake stabilized over a course of 16 days . Mice were returned to an ambient temperature of 24°C and the suppression of food intake was monitored for an additional month . To assess the effects of leptin treatment ( 1μg/g BW twice a day ) on utilization of energy fuel coming from endogenous reserves or food intake , we measured daily changes in food intake , body weight and composition in 8 week-old lean B6 male mice fed chow diet and 16 week-old DIO B6 male mice fed HFD . Leptin was administered for 4 days at 24°C and for an additional 4 days at 4°C . All mice used in the following studies were adult wild-type C57BL/6J+/+ mice . Mice were placed in individual cages with free access to food ( LabDiet 5053 ) and water . At the end of each experiment mice were sacrificed and hypothalamus was dissected in order to measure the level of Npvf mRNA expression . To establish the influence of different ambient temperature on Npvf expression mice were kept in climate-controlled rodent incubators set to 29 and 17°C for the period of weeks prior to sacrifice . Additional experiment was performed to observe the kinetics of changes in the level of Npvf mRNA in the cold . All mice used in this study were first allowed to acclimate to 29°C for 2 weeks before cold challenge . Temperature of the housing unit was then transitioned from 29 to 6°C and mice were cold-challenged for 6 , 12 or 24h . To evaluate the effects of the β3 adrenergic receptor agonist on Ucp1 and Npvf thermoneutrally acclimated mice were injected subcutaneously with 1 mg/kg BW/day CL 316 , 243 or saline for 7 days . Mice were anesthetized by the solution of ketamine , xylopan and chlorpromazine ( 26 . 6 mg/ml , 1 . 67 mg/ml and 0 . 53 mg/ml , respectively , 40μl/10g body weight ) and the blood was collected through heart puncture to EDTA coated tubes . After decapitation , interscapular brown adipose tissue depot ( iBAT ) , inguinal white adipose tissue depot ( iWAT ) and the liver were removed , rapidly frozen in liquid nitrogen and stored at -80°C for subsequent preparation of total RNA . To isolate the whole hypothalamus , the brain was removed and placed on an ice-cooled glass plate with the cortex facing down . The hypothalamus was dissected along the following boundaries: laterally 2 mm either side of the third ventricle , 2 mm dorsally from the base of the brain and rostrocaudally from the optic chiasm to the posterior border of the mammillary bodies . The dissected hypothalami were stored at -80°C until further analysis . The blood was centrifuged for 10 min at 3 , 000 g , 4°C . Plasma was removed and stored at -80°C until assayed . Plasma insulin and leptin were measured by enzyme-linked immunosorbent assay with commercial kits ( Wide range mouse insulin immunoassay kit , Biorbyt Ltd . , Cambridge , UK; Mouse/rat leptin ELISA kit , Phoenix Pharmacuticals , Inc . , Burlingame , CA , United States , respectively ) . Assessment of FFA in plasma was performed with plasma non-esterified free fatty acid detection kit ( Zenbio , Inc . , Research Triangle Park , NC , United States ) . Total RNA was isolated from adipose tissue , liver and hypothalamus using TRI Reagent and BCP phase separation reagent ( Molecular Research Center Inc . Cincinnati , OH , United States ) . RNA was further purified by using the RNAeasy minikit ( QIAGEN , Valencia , CA , United States ) and stored at -80°C in RNase-free H2O with addition of SUPERase-In ( Ambion , Austin , TX , United States ) for RNase protection . Quality and quantity of RNA was determined using UV spectrophotometry ( Nanodrop ) and agarose gel visualization of intact RNA . Quantitative RT-PCR using TaqMan probes and primers ( Applied Biosystems , Foster City , CA , United States ) was performed with standard curves generated using pooled RNA isolated from corresponding tissues collected from eight 8 week old C57BL/6J . +/+ mice . Probe and primer sequences used to perform the analyses are available upon request . All the gene expression data were normalized to the level of cyclophilin b . Total RNA was isolated from the hypothalamus of 8 Lep-/- and 8 Lep+/+ mice maintained at 24 and 6°C , as described above . RNA with RNA Integrity number higher than 8 . 5 ( Agilent 2100 Bioanalyzer , Agilent Technologies , Santa Clara , CA ) was used for microarray analysis of each individual mouse . RNA was amplified , labeled and hybridized onto chips containing over 56 , 000 probes of mouse genes ( Agilent Single Color SurePrint G3 Mouse GE 8x60K Microarray Kit , G4852A , Agilent Technologies ) according to manufacturer’s guidelines . Agilent Feature Extraction software was used for array image analysis . Absolute and comparative analyses were performed using the GeneSpring GX 10 ( Agilent Technologies ) . Quality control filtering after quantile normalization resulted in approximately 33 , 000 probes . Probes that were not above microarray background signal or whose sequences could not be mapped to Ensembl transcripts were discarded . Fold change of gene expression was calculated based on the normalized signal values . Genes were considered significantly down-regulated or up-regulated if the fold-change was less than -1 . 4 or greater than 1 . 4 , respectively , and the FDR-corrected P-value was less than 0 . 05 . To validate the reliability of the results obtained from the microarray analysis , we performed qRT-PCR for all genes of interest . Graphs were created with the GraphPad Prism Software ( Version 6 . 0 , GraphPad Software , Inc . ; La Jolla , USA ) . All data sets were analyzed using Student’s test for groups ( GraphPad Prism Software ) . Data are presented as means ± SEM . Differences between the means for all tests were considered statistically significant if P < 0 . 05 .
|
Current knowledge does not provide a clear , definite view of central mechanisms controlling energy balance upon cold-activated thermogenesis . Here we show that upon cold exposure lean mice maintain body composition but increase food intake to fuel thermogenesis , whereas cold-exposed mice with DIO utilize endogenous fat stores and then transition to increased food intake as body composition approaches that of the lean controls . Using knockout mice with leptin and Ucp1 gene deficiency our study indicates that the relative energy utilization from food intake and endogenous energy reserves to maintain body temperature during cold exposure is independent of both leptin action and brown fat-linked thermogenesis . Using a combination of genetic and biological approaches , we demonstrate that Npvf gene expression in the hypothalamus is regulated by changes in ambient temperature in a manner independent of the nutritional status of the mouse .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Npvf: Hypothalamic Biomarker of Ambient Temperature Independent of Nutritional Status
|
Pandemic 2009 H1N1 influenza A virus ( 2009 H1N1 ) differs from H1N1 strains that circulated in the past 50 years , but resembles the A/New Jersey/1976 H1N1 strain used in the 1976 swine influenza vaccine . We investigated whether sera from persons immunized with the 1976 swine influenza or recent seasonal influenza vaccines , or both , neutralize 2009 H1N1 . Using retroviral pseudovirions bearing hemagglutinins on their surface ( HA-pseudotypes ) , we found that 77% of the sera collected in 1976 after immunization with the A/New Jersey/1976 H1N1 swine influenza vaccine neutralized 2009 H1N1 . Forty five percent also neutralized A/New Caledonia/20/1999 H1N1 , a strain used in seasonal influenza vaccines during the 2000/01–2006/07 seasons . Among adults aged 48–64 who received the swine influenza vaccine in 1976 and recent seasonal influenza vaccines during the 2004/05–2008/09 seasons , 83% had sera that neutralized 2009 H1N1 . However , 68% of age-matched subjects who received the same seasonal influenza vaccines , but did not receive the 1976 swine influenza vaccine , also had sera that neutralized 2009 H1N1 . Sera from both 1976 and contemporary cohorts frequently had cross-neutralizing antibodies to 2009 H1N1 and A/New Caledonia/20/1999 that mapped to hemagglutinin subunit 2 ( HA2 ) . A conservative mutation in HA2 corresponding to a residue in the A/Solomon Islands/3/2006 and A/Brisbane/59/2007 H1N1 strains that circulated in the 2006/07 and 2007/08 influenza seasons , respectively , abrogated this neutralization . These findings highlight a cross-neutralization determinant influenced by a point mutation in HA2 and suggest that HA2 may be evolving under direct or indirect immune pressure .
In June 2009 the World Health Organization declared a new influenza pandemic due to sustained human to human transmission in several geographic regions of the novel swine-origin influenza A H1N1 virus , which was first identified in April by the Centers for Disease Control and Prevention ( CDC ) of the United States of America [1] . This novel H1N1 virus , referred to as pandemic 2009 H1N1 virus ( 2009 H1N1 ) , has a hemagglutinin ( HA ) of classical swine lineage viruses that have circulated in the swine population for decades with little change in HA antigenicity [2] . The 2009 H1N1 HA is antigenically different from those of recent human seasonal influenza H1N1 viruses , but is closely related to A/New Jersey/1976 ( NJ/76 ) influenza virus ( Figure 1 ) , a strain used in 1976 to immunize approximately 45 million people in the US during the swine influenza vaccination campaign after a localized outbreak [3] . However , NJ/76 influenza virus did not circulate . Emergence of the novel pandemic 2009 H1N1 virus raised questions about whether immunization with the 1976 swine or recent seasonal influenza vaccines could confer any protection . Several groups have reported that older persons may have substantial cross-immunity to the 2009 H1N1 , though the literature is mixed on the degree of cross-immunity induced by prior seasonal influenza vaccines [4]–[9] . Influenza virus surface glycoprotein HA mediates virus entry and is the most important target of antibody-mediated protection . Cellular proteases cleave the HA precursor ( HA0 ) to generate the HA1 surface subunit that mediates the binding to cell surface sialic acid receptors and the HA2 transmembrane subunit that mediates membrane fusion between viral and endosomal membranes after endocytosis ( reviewed in [10] , [11] ) . During infection and vaccination , HA elicits neutralizing antibodies . Antigenic maps of HA show that HA1 is the major target of neutralizing antibodies that inhibit virus binding to target cells [12] , [13] and are classically detected by the hemagglutination inhibition ( HI ) assay . However , HA2 is more conserved than HA1 . Neutralizing antibodies that bind to the stalk region of HA2 have been shown to confer broadly cross-neutralizing activity against several subtypes of viruses across clades but within a group [14]–[20] and to provide protection in animal models [16] , [18]–[20] . These antibodies typically do not have HI activity and appear to neutralize virus by interfering with HA-mediated conformational changes required for virus entry [14] , [17] , [18] . Using lentiviral pseudovirions bearing HA on their surface ( HA-pseudotypes ) [21] , we investigated whether persons immunized in 1976 with the NJ/76 swine influenza vaccine or more recently with seasonal influenza vaccines produced neutralizing antibodies to 2009 H1N1 . Both sera from the 1976 swine influenza vaccine trials and contemporary sera from a cohort of subjects who received recent seasonal influenza vaccines , regardless of whether they received the 1976 swine influenza vaccine or not , often contained cross-neutralizing activity to 2009 H1N1 . Some of this cross-neutralizing activity was dependent on the HA2 subunit and surprisingly was sensitive to a naturally-occurring variant at position 89 in HA2 that emerged in recent years . The implications of these findings for potential immune escape are discussed .
Because HA from A/New Jersey/1976 ( NJ/76 ) and 2009 H1N1 influenza viruses are highly related ( Figure 1 ) , we first asked whether immunization in 1976 with the NJ/76 swine influenza vaccine could provide any immunity against the 2009 H1N1 influenza virus . Sixty five pre- and post-vaccination sera archived from the NJ/76 swine influenza vaccine trials conducted in 1976 [22] were evaluated for neutralizing activity to either NJ/76 or 2009 H1N1 A/Mexico/4108/2009 ( Mex/4108/09 ) using HA-pseudotypes . Previously , we showed that HA-pseudotype neutralization titers using 95% inhibitory concentration ( IC95 ) correlate well with conventional microneutralization titers using replicating influenza virus [23] and that HA-pseudotype neutralization is specific [21] , [24] . Microneutralization titers >160 and a 4-fold increase after vaccination in assays involving replicating influenza virus have been proposed as correlates of seroprotection [4] , but protective titers for HA-pseudotype neutralization have not yet been established . Positive control sera from 2009 H1N1 influenza virus infected ferrets typically have titers >10 , 000 [24] . Sera from the NJ/76 swine influenza vaccine trial were then tested and showed that NJ/76 vaccination generated neutralizing antibodies ( titers >160 and a 4-fold increase after vaccination ) in 85% and 77% of subjects against NJ/76 and Mex/4108/09 HA-pseudotypes , respectively ( Table 1 and Figure 2A ) , consistent with the high degree of relatedness between the viruses and other recent reports [4] , [8] . The neutralizing antibody titers to NJ/76 ( GMT 597 ) and Mex/4108/09 ( GMT 573 ) were also similar and correlated ( Figure 2B ) . Most importantly , all sera with neutralization activity to NJ/76 showed significant neutralization activity to Mex/4108/09 ( Figure 2B ) . Pre-vaccination sera did not exhibit significant neutralizing activity to HA-pseudotypes for either influenza virus , though titers against Mex/4108/09 ( GMT 60 ) were higher than those against NJ/76 ( GMT 3 ) , suggesting that influenza viruses with shared epitopes to Mex/4108/09 influenza virus may have circulated previously . We next asked whether subjects with a history of NJ/76 vaccination have significant neutralization titers to 2009 H1N1 today . Accordingly , we analyzed sera from a contemporary cohort of 23 subjects who had a history of NJ/76 vaccination and 19 aged-matched control subjects who did not . As shown in Table 2 and Figure 3A , sera from those who received the NJ/76 vaccine more than 30 years ago showed significant neutralization titers to NJ/76 ( GMT 181 ) , with 52% having neutralization titers >160 . Sera from subjects who did not receive the NJ/76 vaccine had a GMT of only 44 to NJ/76 , although a few individuals showed significant neutralization titers ( >160 ) . We note that the neutralization titers to NJ/76 in sera from subjects who did not receive the NJ/76 vaccine in this contemporary cohort were higher than the pre-vaccination sera in NJ/76 trials , suggesting that natural infection and/or vaccination with seasonal influenza strains during the period 1977–2009 provided a low level of cross-neutralization to NJ/76 . Thus there appears to be residual immunity to NJ/76 in a majority of persons who were previously immunized with NJ/76 vaccine , or immunity may have been boosted by exposures during intervening years . We next assessed cross-neutralization to 2009 H1N1 HA-pseudotypes . Unexpectedly , titers among those immunized with the NJ/76 vaccine were higher against Mex/4108/09 ( GMT 331 ) compared to NJ/76 ( GMT 181 ) , with 83% having neutralizing antibody titers ranging from 161–1456 ( GMT 469 ) . There was a significant correlation between neutralization titers to NJ/76 and Mex/4108/09 ( Figure 3B ) . However , sera from subjects without a history of NJ/76 vaccination had similar cross-neutralization titers to Mex/4108/09 ( GMT 305 ) , with 68% having neutralization titers >160 ( Table 2 and Figure 3A ) . The substantial neutralizing titers to Mex/4108/09 found in a high proportion of subjects in this contemporary cohort , regardless of their vaccination history to NJ/76 , indicated that their cumulative history of influenza infections and vaccinations have involved strains that share neutralizing epitopes with the 2009 H1N1 influenza virus . All 45 subjects in the contemporary cohort received all annual seasonal influenza vaccines for at least the past five years ( 2004/05–2008/09 seasons ) . To investigate potential correlations between neutralizing activity to recent seasonal H1N1 influenza and the 2009 H1N1 viruses , we tested all sera for HA-pseudotype neutralizing activity against the three recent seasonal H1N1 influenza strains , A/New Caledonia/20/1999 ( NCD/20/99 ) , A/Solomon Islands/3/2006 ( SI/03/06 ) , and A/Brisbane/59/2007 ( Bris/59/07 ) . NCD/20/99 was used 7 times in influenza vaccines during the 2000/01 to 2006/07 seasons . SI/03/06 and Bris/59/07 were used in the 2006/07 and 2008/09 seasonal influenza vaccines , respectively ( www . fludb . org/brc/vaccineRecommend . do ? decorator=influenza ) . Neutralization of HA-pseudotypes corresponding to each of these strains is specific , as shown in Table S1 . Using these HA-pseudotypes , 100% of subjects showed significant neutralization titers against NCD/20/99 ( GMT 1237 ) ( Table 3 and Figure 3C ) , consistent with the repeated use of the NCD/20/99 strain in recent seasonal influenza vaccines . Only 49% and 60% had neutralization titers >160 against Bris/59/07 and SI/03/06 , respectively ( Table 3 and Figure 3C ) . By comparison , the GMT of neutralizing titers to 2009 H1N1 is 319 , with 76% having neutralization titers >160 , regardless of vaccination history to NJ/76 ( Figure 3C ) . The neutralization titers to Mex/4108/09 did not correlate with the titers to Bris/59/07 and SI/03/06 ( data not shown ) , but subjects with higher neutralization titers ( >600 ) to NCD/20/99 showed higher cross-neutralization titers to Mex/4108/09 ( p<0 . 05 ) ( Figure 3D ) , suggesting that there may be shared neutralization epitopes between NCD/20/99 and Mex/4108/09 . To look for cross-neutralization between 2009 H1N1 and NCD/20/99 , we analyzed sera collected in 1976 from the NJ/76 vaccine trials for neutralizing activity to NCD/20/99 HA-pseudotypes . Since persons participating in the NJ/76 swine influenza vaccine trial were presumably not previously exposed to NCD/20/99 through natural infection or by vaccination , we considered the presence of neutralizing activity to NCD/20/99 in these sera to be due to cross-neutralizing antibodies . We found that the post NJ/76 vaccination sera had significant cross-neutralization activity to NCD/20/99 ( GMT 320 ) with 45% having neutralization titers >160 and a 4-fold increase over pre-immunization titers , while only 12% of the pre NJ/76 vaccination sera had significant neutralization titers ( Table 1 and Figure 2A ) . The reason that several pre NJ/76 vaccination sera have significant neutralizing activity to NCD/20/99 may be due to prior infections with related viruses . To determine whether NJ/76 vaccination elicits cross-neutralizing activity to other recent seasonal H1N1 viruses , we analyzed the sera for the presence of neutralizing antibodies to Bris/59/07 . Neutralization of Bris/59/07 HA-pseudotypes was seen in only 17% sera with titers >160 and a 4-fold increase over pre-immunization titers . Although the titers to Bris/59/07 were low ( GMT 52 ) after vaccination with NJ/76 , they were significantly higher than the titers in the pre-vaccination group ( GMT 8 ) ( Table 1 and Figure 2A ) . However , NJ/76 vaccination elicited much less cross-neutralization to Bris/59/07 than to NCD/20/99 . The cross-neutralization activity seen in sera after immunization with NJ/76 and seasonal influenza vaccines suggested the presence of shared neutralization epitopes between 2009 H1N1 and NCD/20/99 . Since neutralizing antibodies can target either HA1 or HA2 , we next investigated which subunit of HA accounts for the majority of the cross-neutralization between 2009 H1N1 and NCD/20/99 observed in our sera . First we analyzed the sera from the NJ/76 vaccination trials . The sera with neutralization titers <160 to Bris/59/07 HA-pseudotypes were considered negative for neutralization to either HA1 or HA2 of Bris/59/07 HA . Twenty-one out of 65 post NJ/76 vaccination sera without neutralization activity to Bris/59/07 , but with neutralization titers >160 and a 4-fold increase over pre-immunization titers to NCD/20/99 ( neutralization titers <160 before vaccination ) , were identified ( Table S2 ) and used for mapping . Chimeric HA involving the NCD/20/99 HA1 and Bris/59/07 HA2 subunits ( NCD . HA1-Bris . HA2 ) , as well the Bris/59/07 HA1 and NCD/20/99 HA2 subunits ( Bris . HA1-NCD HA2 ) were constructed and used for making HA-pseudotypes . The infectivity and amount of HA in these chimeric HA-pseudotypes were comparable to the wild-type HA-pseudotypes ( Figure S1 ) . The chimeric HA-pseudotypes showed that: HA1 was responsible for most of the NCD/20/99 cross-neutralization in 2 out of 21 sera ( e . g . 2S5H and 2S5A ) ; HA2 was responsible for most of the NCD/20/99 cross-neutralization in 9 out of 21 sera ( e . g . 2S5G , 2S5F , 2S5B , 2S4H , 2S3D , 2S2E , 2S1A , 1S2B and 1S1B ) ; and both HA1 and HA2 were responsible for much of the NCD/20/99 cross-neutralization in 10 out of 21 sera ( e . g . 2S6E , 2S6B , 2S5C , 2S4G , 2S4F , 2S4B , 2S3E , 2S3C , 2S3B and 1S2A ) ( Figure 4 ) . In many cases , cross-neutralization titers to NCD/20/99 did not simply reflect the sum of the individual neutralization titers to each of the chimeras containing either NCD HA1 or HA2 subunits ( e . g . 2S6B , 2S4G , 2S4F , 2S4B , 2S3E , 2S3D and 2S3C ) , indicating that HA1-HA2 interactions affected neutralization . These data suggested that there may be several targets for cross-neutralization . Nonetheless , the neutralization activity frequently mapped to the HA2 subunit , and in many cases , HA2 appeared to be the major determinant for cross-neutralization . Next we analyzed the sera from the contemporary cohort . Sera with cross-neutralization titers ( >160 ) to Mex/4108/09 , but without neutralization titers ( <160 ) to Bris/59/07 were identified ( Table S3 ) and used for evaluating neutralizing antibodies that may be directed to Mex/4108/09 HA1 and/or HA2 subunits . HA-pseudotypes carrying the chimeric HA consisting of Bris/59/07 HA1 and Mex/4108/09 HA2 ( Bris . HA1-Mex . HA2 ) showed that neutralization titers to Mex/4108/09 HA and Bris . HA1-Mex . HA2 were similar in all comparisons ( samples S1 , S7 , S24 , S31 , S42 , S44 , S45 , S58 and S59 ) ( Figure 5A ) , suggesting that cross-neutralization to Mex/4108/09 involves the Mex/4108/09 HA2 subunit . Curiously , the chimeric Mex . HA1-Bris . HA2 HA-pseudotypes did not have high enough infectivity for neutralization studies , despite good HA incorporation and cleavage of HA0 in the HA-pseudotypes ( Figure S1 ) . Therefore , we could not directly assess the contributions of the Mex/4108/09 HA1 subunit to cross-neutralization . To confirm the reliability of the cross-neutralizing data involving chimeric HA-pseudotypes with the Mex/4108/09 HA2 subunit , we identified sera with neutralization titers ( >160 ) to NCD/20/99 , but without cross-neutralization titers ( <160 ) to Mex/4108/09 ( Table S4 ) . For these sera ( samples S3 , S25 , S39 , S43 , S54 , S56 and S201 ) , neutralization titers for HA-pseudotypes carrying the chimeric NCD/20/99 HA1 and Mex/4108/09 HA2 ( NCD . HA1-Mex . HA2 ) or NCD/20/99 HA were similar ( Figure 5B ) , indicating that neutralization antibodies were directed to the NCD/20/99 HA1 subunit . Therefore , the presence of the Mex/4108/09 HA2 subunit in chimeric HA does not apparently give spurious neutralization results . Again , we were unable to assess neutralization of the complementary chimeric HA-pseudotypes containing Mex/4108/09 HA1 ( Mex . HA1-NCD . HA2 ) due to the poor infectivity of this chimera , despite good incorporation of mature chimeric HA into the HA-pseudotypes ( Figure S1 ) . The difficulties in generating functional chimeric HA involving Mex/4108/09 HA1 further suggests that there are interactions between the Mex/4108/09 HA1 and HA2 subunits that are not present in recent seasonal H1N1 HA . In 1993 [14] and again in a number of recent studies [15]–[20] , neutralizing monoclonal antibodies that are broadly active against many influenza subtypes have been identified and mapped to epitopes in the stalk regions of the HA2 subunit [14] , [16]–[20] . Although some of the cross-neutralization that we observed in our sera appears to map to the HA2 subunit , our data indicated that this cross-neutralization may be strain specific . As shown in Figure 4 , Figure 5A and Figure 6 , we found that sera with cross-neutralization to NCD/20/99 and Mex/4108/09 HA2 did not neutralize Bris/59/07 . Significantly , there are only two amino acid differences in HA2 , at the positions 89 ( 415 in full HA ) and 146 ( 472 in full HA ) between NCD/20/99 and Bris/59/07 HA2 ( Figure 6A ) , suggesting that these two amino acids could influence HA2 antigenicity . When a leucine at residue 89 in HA2 ( 89L ) or an asparagine at position 146 in HA2 ( 146N ) corresponding to NCD/20/99 HA2 were introduced into Bris/59/07 HA2 , the sera without cross-neutralization to Bris/59/07 HA showed neutralization to Bris/59/07 HA2-89L , but not to Bris/59/07 HA2-146N , with titers similar to NCD/20/99 HA and Bris . HA1-Mex . HA2 ( Figure 6B and 6C ) . When both 89L and 146N were presented in Bris/59/07 HA2 , serum titers were the same as those to Bris . HA1-NCD . HA2 in Figure 4 ( data not shown ) . These results demonstrated that the neutralization epitopes in HA2 were influenced by residue 89 in HA2 ( 415 in full HA ) . We then reviewed human H1N1 influenza virus HA sequences ( www . fludb . org/brc/home . do ? decorator=influenza ) and noted that leucine at position 89 in HA2 has been maintained in seasonal H1N1 influenza viruses from at least 1918 to 2005 ( Table 4 ) . During this period , there are only two exceptions: A/Denver/1/1957 from North America has a methionine and A/Canterbury/106/2004 from Oceania has an isoleucine at position 89 of HA2 . The change of leucine to isoleucine at position 89 of HA2 appeared frequently in 2006 with about 37 . 8% strains containing isoleucine , and the change of leucine to isoleucine continued in 2007 with about 34 . 1% strains containing isoleucine . However , by 2008 , isoleucine completely replaced leucine at position 89 in HA2 , raising the possibility that this change may reflect immune escape .
The 2009 H1N1 HA diverges considerably from recent seasonal H1N1 HA and is more closely related to the NJ/76 HA ( Figure 1 ) , raising doubts about the extent of protection that could be afforded by vaccination with recent seasonal influenza vaccines . Our studies show that sera from the NJ/76 swine influenza vaccine trials and contemporary sera from subjects who received recent seasonal influenza vaccines , regardless of whether they had been immunized with the NJ/76 swine influenza vaccine , frequently have cross-neutralizing activity to the 2009 H1N1 . Further , these sera revealed one or more cross-neutralization epitopes that were sensitive to a conservative amino acid change in position 89 in the HA2 subunit , corresponding to a naturally-occurring amino acid variant that emerged in seasonal H1N1 influenza viruses in recent years . Several groups have reported that prior infections or vaccinations can confer some immunity to 2009 H1N1 , though findings vary . There is agreement that individuals >65 years have substantial cross-reactive antibodies to the 2009 H1N1 , consistent with the epidemiology of the 2009 H1N1 pandemic showing that younger age groups were disproportionately affected [4] , but the extent of cross-immunity induced by recent seasonal influenza vaccines is more ambiguous [4]–[9] , [25]–[29] . Differences in methodologies and history of vaccination or infection with NCD/20/99 may have affected the outcomes . Our results involving persons aged 48–64 years ( Table S5 ) extend other reports showing that older persons generally have some pre-existing immunity to the 2009 H1N1 , but more significantly highlight the presence of cross-neutralizing antibodies between 2009 H1N1 and NCD/20/99 . Because all subjects in our contemporary cohort received yearly seasonal influenza vaccines for at least the past five years , and NCD/20/99 was repeatedly used in seasonal vaccines during the 2000/01–2006/07 influenza seasons , we cannot determine the extent to which influenza vaccinations and/or natural infections contributed to the generation of cross-neutralizing antibodies to 2009 H1N1 and NCD/20/99 . To investigate potential cross-neutralizing determinants in NCD/20/99 and 2009 H1N1 , we used chimeric HA-pseudotypes involving HA1 and HA2 subunits of NCD/20/99 and Bris/59/07 and sera that lacked neutralization to Bris/59/07 ( Table S2 ) . Both contemporary and archived sera from the NJ/76 swine influenza vaccine trials contained cross-neutralizing antibodies that depended on the HA2 subunit ( Figure 4 and 5 ) . Most remarkable , we found that the cross-neutralization was influenced by a single conservative amino acid change at position 89 in HA2 , which differed between NCD/20/99 and Bris/59/07 ( Figure 6 ) . Thus , these data reveal a new determinant in the C helix region of the HA2 stalk that modified sensitivity to cross-neutralizing antibodies present in human sera from two different cohorts separated by more than three decades . Growing interest in the generation of broadly neutralizing influenza antibodies has led to the discovery of several new monoclonal antibodies that bind to HA2 [14]–[20] , [30] , [31] . The first reported heterosubtypic neutralizing antibody , C179 , derived from a mouse immunized with the A/Okuda/57 H2N2 strain , was found to be directed to a conformational epitope involving the A helix in the HA2 stalk ( Figure 7A ) and a region in HA1 [14] . More recently , several other HA2 heterosubtypic neutralizing monoclonal antibodies that are potent against strains from H1 and H5 subtype ( Group 1 ) influenza viruses have been isolated using various methods . Some of these antibodies have been also shown to make contacts with the A helix of HA2 [16] , [17] , [30] ( Figure 7A ) . Other HA2 monoclonal antibodies have been shown to bind to a highly conserved pocket in the stalk region containing the fusion peptide [18] or undetermined regions of the HA2 stalk [19] . Another potent broadly neutralizing monoclonal antibody against H3N2 ( Group 2 ) but not H1N1 ( Group 1 ) strains was shown to bind to a peptide corresponding to the C helix region in the HA2 stalk [20] . The HA2 monoclonal antibodies bind to regions in the HA2 stalk and interfere with conformational changes that are needed for virus entry [32] , but they do not block HA attachment to receptors . These HA2 antibodies lack HI activity and were discovered using neutralization assays that sometimes involved HA-pseudotypes [18]–[20] . We [21] , [23] and others [33] , [34] have shown that HA-pseudotypes neutralization titers are highly correlated with microneutralization titers for replicating influenza virus , but the correlate of protection using HA-pseudotype neutralization titers has not been determined . Also , glycoproteins on the surface of HIV-based retroviral particle may be less densely packed and more exposed compared to HA on the surface of influenza viral particles , perhaps making them more susceptible to HA2-directed neutralization compared to influenza virus , as suggested in some studies [18] , [19] . While sensitive screening assays have allowed many groups to fish out broadly neutralizing antibodies , it is generally believed that HA2 heterosubtypic neutralizing antibodies are present at relatively low concentrations , as compared with antibodies directed to HA1 [19] , [35] . The need to change annual seasonal influenza vaccines to match dominant circulating strains indicates that such HA2 cross-neutralizing antibodies may not be present at high enough titers to provide robust protection . It is therefore difficult to discern the degree to which HA2 antibodies in our sera samples could contribute to protection to 2009 H1N1 virus . However , studies in animal models have provided proof of concept that induction [16] , [18]–[20] or passive transfer of HA2 antibodies alone [20] , [36] can provide protection . Appropriately designed vaccines may be able to induce robust immune responses to conserved neutralizing epitopes in HA2 [35] , [37] . Recent examples involving several approaches are showing promise [20] , [38] , [39] . Our finding that the conservative substitution of isoleucine for 89 L reduced sensitivity to cross-neutralizing antibodies present in our sera was surprising . The crystal structure of the A/Cal/04/2009 HA [40] shows that 89 L packs tightly into a poorly exposed crevice underneath the HA1 crown ( Figure 7A ) , making intimate contact with HA1 through a lysine and tyrosine at residues 310 and 308 , respectively ( Figure 7B ) . Substitution of 89 L with isoleucine may cause the interactions between HA1 and HA2 in this region to shift in order to accommodate the alternate side chain ( Figure S2 ) , and in doing so , could directly alter exposure or conformation of the antibody binding site . Alternatively , residue 89 may be evolving in response to immune pressure at distant sites . For example , 89I may reflect an adaptive change in HA2 resulting from direct immune pressure on epitopes in HA1 . The 89I substitution could also impose allosteric changes on nearby or more distant neutralizing epitopes in either HA1 or HA2 . The observation that Bris/59/07 was less sensitive to neutralization by an HA2 antibody compared to NCD/20/99 is consistent with the notion that this residue could influence neutralization by HA2 antibodies [19] . We also note that 89L is not near any of the contact residues for the recently described HA2 monoclonals specific for Group 1 HAs , although it is located on the C helix region of the HA2 stalk that has recently been suggested to contain an epitope for the 12D1 monoclonal antibody that binds H3 strains from Group 2 . Review of the database of human H1N1 HA also offers intriguing clues about the potential significance of the change of leucine to isoleucine at position 89 in HA2 . We note that 89L has been maintained in seasonal H1N1 influenza viruses from at least 1918 until 2006 when it started to change to isoleucine , and 89L disappeared in 2008 ( Table 4 ) . It is tempting to speculate that this change could reflect immune escape . We also note that H3 strains from Group 2 influenza viruses generally have an isoleucine at the corresponding position in HA2 . Interestingly , unlike Group 1 H1N1 HA , a carbohydrate can be seen in the H3N2 HA crystal structure extending in the vicinity of the isoleucine ( coming from N285 ) ( PDB 3HMG ) [41] , [42] , which could conceivably have evolved to shield it from neutralizing antibodies . These observations offer a cautionary note that antigenic drift in this region may arise under strong selection pressure . Nonetheless , the viable substitutions may be limited due to the fact that residue 89 and others in the stalk regions make important contacts in both the native and low pH structures of HA , consistent with the difficulties in generating escape mutants with some of the HA monoclonal antibodies [18] , [20] . Perhaps this explains why H3N2 strains have incorporated a carbohydrate in the vicinity of this region . In summary , our studies showed that cross-neutralizing antibodies to 2009 H1N1 influenza that involve the HA2 subunit could be detected in sera collected in 1976 from NJ/76 swine influenza vaccine trials and sera from persons aged 48–64 who received annual influenza vaccines for at least the past five years . A conservative substitution at position 89 in HA2 , found in drifted seasonal influenza virus variants from the 2006/07 and 2007/08 influenza seasons , abrogated this neutralization . Future studies involving vaccines that elicit strong antibody responses to HA2 will reveal the extent to which mutations can lead to immune escape .
Full-length HA ORF with Q223R mutation from A/Mexico/4108/2009 ( GenBank GQ223112 ) and full-length wild type HA ORFs from A/Solomon Islands/3/2006 ( GenBank EU100724 ) , A/New Caledonia/20/1999 ( GenBank AY289929 ) , and A/Brisbane/59/2007 ( GenBank CY058487 ) were amplified from viruses by reverse transcription-polymerase chain reaction ( RT-PCR ) . Full-length wild type NA ORF from A/California/04/2009 ( GenBank FJ966084 ) was also amplified from virus by RT-PCR . Full-length wild type HA ORF of A/New Jersey/1976 ( GenBank CY021957 ) was chemically synthesized by GenScript ( Piscataway , NJ ) . Chimeric HA carrying HA1 and HA2 from different strains were constructed by ligation of PCR fragments of HA1 and HA2 . The HA and NA ORFs were then placed into the pCMV/R expression plasmid obtained from Dr . Gary J . Nabel ( National Institutes of Health ( NIH ) , Bethesda , MD ) , as described previously [21] . Full-length wild type M2 ORF of A/Puerto Rico/8/1934 ( GenBank EF467824 ) was chemically synthesized by Integrated DNA Technologies ( Coralville , IA ) and placed into pCDNA 3 . 1 ( + ) ( Invitrogen , Carlsbad , CA ) . Codon-optimized human airway trypsin-like protease ( HAT ) gene expression construct ( pCAGGS-HATcop ) was described before [23] . The HIV gag/pol ( pCMVΔR8 . 2 ) and Luc reporter ( pHR'CMV-Luc ) constructs were described previously [43] , [44] and obtained from Dr . Gary J . Nabel ( NIH , Bethesda , MD ) . 293T cells were cultured in Dulbecco's modified eagle medium ( DMEM ) with high glucose , L-Glutamine , MEM non-essential amino acids , penicillin/streptomycin and 10% fetal calf serum . Ethics approval by the Research Involving Human Subjects Committee ( RIHSC ) at the US Food and Drug Administration was obtained for use of the sera involved in this study . Under 45 CFR 46 . 101 ( b ) ( 4 ) , the sera from the 1976 swine influenza trial was included in the category of exempt research because the study used only existing sera , and information was recorded in such a manner that subjects can not be identified , either directly or through identifiers ( RIHSC Protocol #09-043B ) . The sera from the contemporary cohort were obtained with written informed consent from all participants ( RIHSC Protocol #09-110B ) . Two groups of human sera were used in this study . The sera in group one included frozen samples retrieved from storage at FDA/CBER involving 65 pre-vaccination and post-vaccination sera from A/New Jersey/1976 swine influenza vaccine trials conducted in 1976 [22] . The sera in group two were collected in September-December of 2009 from 45 volunteers aged 48–64 years , without a history of vaccinations or influenza symptoms or exposures in 2009 . All subjects in group two received at least five year ( 2004/05 to 2008/09 ) annual seasonal influenza vaccines including A/New Caledonia/20/1999 , A/Solomon Islands/3/2006 and A/Brisbane/59/2007 used for the seasons from 2000/01 to 2008/09 , and 23 subjects among them also received the A/New Jersey/1976 swine influenza vaccine ( Table S5 ) . Sera were heat inactivated by incubation at 56°C for 30 minutes prior to use in neutralization assays . Sera were assessed for neutralizing antibodies to 2009 H1N1 ( A/Mexico/4108/2009 ) and the 2000/09 seasonal H1N1 influenza viruses ( A/New Caledonia/20/1999 , A/Solomon Islands/3/2006 , A/Brisbane/59/2007 ) using an HA-pseudotype neutralization assay , as described below . HA-pseudotypes carrying a luciferase ( Luc ) reporter gene were produced in 293T cells as described previously [21] . 2 . 5 µg of HAT , 2 . 5 µg of A/Puerto Rico/8/1934 M2 , and 4 µg of A/California/04/2009 NA expression plasmids were included in the transfection . HA-pseudotypes were collected 48 hr post-transfection , filtered through a 0 . 45-µm low protein binding filter , and used immediately or stored at −80°C . HA-pseudotype titers were measured by infecting 293T cells with HA-pseudotypes for 48 hr prior to measuring luciferase activity in infected cells ( luciferase assay reagent , Promega ) as described previously [21] . HA-pseudotype titers were expressed as relative luminescence unit per milliliter of HA-pseudotype supernatants ( RLU/ml ) . As previously described [23] , [24] , HA-pseudotypes containing approximately 15 ng/ml of p24 antigen and 12 ng/ml of HA were incubated with heat-inactivated serum samples for 1 hr at 37°C , then 100 µl of HA-pseudotypes and serum mixtures were inoculated onto 96-well plates that were seeded with 2 x 104 293T cells/well one day prior to infection . HA-pseudotype infectivity was evaluated 48 hr later by luciferase assay , as previously described [21] . The serum dilution causing a 95% reduction of RLU compared to control ( IC95-neutralizing antibody titer ) was used as the neutralization endpoint titer [23] . IC95 was calculated using Graphpad Prism software . Data reported were from at least two independent experiments , with each serum sample run in duplicate . To evaluate vaccination responses and potential cross-protection , sera with neutralization titers over 160 that inhibited 95% infectivity were considered highly significant [4] , [23] . The neutralization titers were analyzed with nonlinear regression using GraphPad Prism software . The correlation of neutralization titers was evaluated with Spearman's p , a test for nonparametric correlation . t-test , geometric mean titer ( GMT ) with 95% confidence intervals and corresponding P value were analyzed using GraphPad Prism software . P values <0 . 05 were considered statistically significant .
|
Influenza A viruses mutate to escape neutralization by antibodies . These mutations predominantly occur in the globular head of the hemagglutinin protein , while the stalk is more conserved . Pandemic 2009 H1N1 influenza virus differs from seasonal H1N1 strains that circulated in the past 50 years and resembles a strain that did not circulate but was used in the 1976 swine influenza vaccine . We investigated whether persons immunized with either the 1976 swine influenza or recent seasonal influenza vaccines , or both , have antibodies that cross-neutralize pandemic 2009 H1N1 . Sera from 1976 swine influenza vaccine trials cross-neutralized pandemic 2009 H1N1 and to a lesser extent the A/New Caledonia/20/1999 H1N1 strain that was used in vaccines during the 2000/01–2006/07 influenza seasons . Sera from persons who received several seasonal influenza vaccines containing A/New Caledonia/20/1999 H1N1 cross-neutralized pandemic 2009 H1N1 , regardless of whether they received the 1976 swine influenza vaccine . We found that cross-neutralization between 2009 H1N1 and A/New Caledonia/20/1999 frequently mapped to the hemagglutinin stalk . A mutation in the stalk of strains circulating during the 2007/08–2008/09 seasons abrogates this neutralization . These findings highlight a cross-neutralization determinant influenced by a point mutation in the hemagglutinin stalk and suggest that the stalk may be evolving under direct or indirect immune pressure .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"viral",
"vaccines",
"emerging",
"viral",
"diseases",
"microbiology",
"viral",
"structure",
"adaptive",
"immunity",
"infectious",
"disease",
"control",
"immunizations",
"infectious",
"diseases",
"viral",
"immune",
"evasion",
"biology",
"viral",
"envelope",
"immunity",
"virology"
] |
2011
|
Cross-Neutralizing Antibodies to Pandemic 2009 H1N1 and Recent Seasonal H1N1 Influenza A Strains Influenced by a Mutation in Hemagglutinin Subunit 2
|
The Wnt family of secreted proteins has been proposed to play a conserved role in early specification of the bilaterian anteroposterior ( A/P ) axis . This hypothesis is based predominantly on data from vertebrate embryogenesis as well as planarian regeneration and homeostasis , indicating that canonical Wnt ( cWnt ) signaling endows cells with positional information along the A/P axis . Outside of these phyla , there is strong support for a conserved role of cWnt signaling in the repression of anterior fates , but little comparative support for a conserved role in promotion of posterior fates . We further test the hypothesis by investigating the role of cWnt signaling during early patterning along the A/P axis of the hemichordate Saccoglossus kowalevskii . We have cloned and investigated the expression of the complete Wnt ligand and Frizzled receptor complement of S . kowalevskii during early development along with many secreted Wnt modifiers . Eleven of the 13 Wnt ligands are ectodermally expressed in overlapping domains , predominantly in the posterior , and Wnt antagonists are localized predominantly to the anterior ectoderm in a pattern reminiscent of their distribution in vertebrate embryos . Overexpression and knockdown experiments , in combination with embryological manipulations , establish the importance of cWnt signaling for repression of anterior fates and activation of mid-axial ectodermal fates during the early development of S . kowalevskii . However , surprisingly , terminal posterior fates , defined by posterior Hox genes , are unresponsive to manipulation of cWnt levels during the early establishment of the A/P axis at late blastula and early gastrula . We establish experimental support for a conserved role of Wnt signaling in the early specification of the A/P axis during deuterostome body plan diversification , and further build support for an ancestral role of this pathway in early evolution of the bilaterian A/P axis . We find strong support for a role of cWnt in suppression of anterior fates and promotion of mid-axial fates , but we find no evidence that cWnt signaling plays a role in the early specification of the most posterior axial fates in S . kowalevskii . This posterior autonomy may be a conserved feature of early deuterostome axis specification .
The Wnt family of secreted ligand proteins is involved in a wide range of developmental functions during animal development , from embryonic induction to cell fate specification and the generation of cell polarity [1 , 2] . The presence of Wnt ligands , receptors , and Wnt antagonists in cnidarians , sponges , and ctenophores indicates an early origin and diversification of the Wnt pathway during the radiation of metazoan phyla , before the emergence of bilaterians [3–7] . Wnt ligands can act via noncanonical and canonical pathways . The canonical Wnt ( cWnt ) or Wnt-β-catenin pathway is the best studied and involves ligand binding of both the Frizzled ( Fz ) receptor and Lrp5/6 coreceptor . Activation of this pathway leads to stabilization of β-catenin and trafficking to the nucleus where it activates downstream targets in cooperation with T-cell factor/lymphoid enhancer factor ( Tcf/lef ) transcription factors . Comparative studies have proposed conserved roles of cWnt signaling in basic axial patterning of metazoan embryos , suggesting that cWnt signaling played a fundamental role in the early establishment of metazoan axis formation [3 , 4 , 8–14] . cWnt plays critical axial patterning roles during the development of metazoan embryos: in cnidarians it is first involved in the establishment of the endoderm with a pulse of β-catenin in the animal pole during early animal/vegetal ( AV ) patterning , then later to define the polarity of the oral/aboral axis of the adult . In bilaterians , there is robust evidence from both deuterostomes and lophotrochozoans of a similar role of β-catenin in AV patterning [15–17] , suggesting this is a widely conserved mechanism in eumetazoans . Perhaps the best known axial patterning role of cWnt is during the establishment of the anteroposterior ( A/P ) axis , with compelling comparative data generated across bilaterian lineages [11] . However , these comparative data do not represent a single conserved developmental role of cWnts in A/P patterning and can be divided into two discrete phases that are mechanistically distinct; the first in establishment of the A/P axis , and second , after the A/P axis has been established , the initiation of axis elongation from a posterior growth zone [18] . Within the bilaterians , most of the data on the early role of cWnt signalling in the establishment of A/P pattern is from vertebrates . cWnt signaling has its strongest effect on head patterning , acting as a classical morphogen; it both represses anterior ( forebrain and midbrain ) neural fates and induces posterior ( hindbrain ) neural fates in a dose-dependent manner [19–24] . Wnt antagonists , such as Dickkopf ( Dkk ) and Secreted frizzled-related proteins ( Sfrps ) , expressed in the anterior neurectoderm and in the prechordal plate mesoderm of the organizer , protect the anterior neurectoderm from the posteriorizing effects of Wnt ligands secreted from the mesoderm and posterior neural plate [21 , 25 , 26] . It has thus been proposed that a simple gradient of Wnt activity—high in the posterior and low in the anterior—endows cells with positional information along the A/P axis in the central nervous system [10 , 20 , 23] . Data from several species strongly support this model of brain patterning . However , surprisingly , the role of cWnt signaling in the establishment of the most posterior region of the neural plate has not been extensively investigated , and most focus has been on the effect of Wnts in repression of forebrain and midbrain and promotion of hindbrain fates , with little relevant data on the spinal cord [27] Experimental results from nonvertebrate chordates , the cephalochordates and tunicates , are consistent with data from vertebrates , but suggest that this patterning system is not as fully deployed as it is in vertebrates . In the cephalochordate , Branchiostoma floridae , Wnt ligands are localized posteriorly and antagonists anteriorly [28–30] . However , constitutive activation of the cWnt pathway using glycogen synthase kinase 3 beta ( GSK3β ) inhibitors only represses far anterior markers and expands only blastoporal markers , suggesting that Wnt signaling determines the identity of the two ends of the embryo , but not the intervening regions that are responsive to cWnt in vertebrates [29] . The role of Wnts in early ascidian development remains largely unexplored experimentally . However , the expression of wnt5 posteriorly and the anterior localization of negative regulators such as sfrp1/5 and ror are suggestive of a potential role in A/P patterning [31–33] . Outside of chordates , the role of Wnts in A/P patterning has recently been demonstrated in sea urchins during larval development [34 , 35] and has drawn comparisons with cWnt suppression in the anterior neural plate of vertebrates , suggesting common elements of regulation between the apical pole of sea urchin larvae and the anterior neural plate of chordates [36] . In protostomes , further broad phylogenetic support for an ancient role of Wnt in A/P patterning comes from representatives in ecdysozoans and lophotrochozoans . In lophotrochozoans , this is particularly striking during regeneration in planarians demonstrating a critical role of β-catenin in the decision between regeneration of head or tail following experimental amputations and during homeostasis in maintenance of the posterior [37–39] . In the annelid Platynereis dumerilli , cWnt activation during early embryonic development results in repression of anterior markers [40] . In arthropods , there is no early axial role of cWnt signaling in Drosophila . However , in the basal short germ band insect , Tribolium , analysis of axin , a cWnt signaling repressor , reveals an important role of Wnts in defining the early A/P axis [41] . When considered with the polarized expression of Wnts and their antagonists during Caenorhabditis elegans early development , an ancestral role of cWnt in ecdysozoans A/P development is implied [42] . A recent study of regeneration in a representative acoel , a group that most likely occupies a key phylogenetic position before the protostome/deuterostome split [43] reveals a key role of cWnt in regeneration , very similar to planarians . With the aim of adding an important additional data point to the function of cWnt signaling in A/P axis formation of deuterostome and bilaterians , we have investigated its involvement in specifying embryonic axial properties during the early development of the direct-developing enteropneust S . kowalevskii . Hemichordates are the sister group to echinoderms and are closely related to chordates [44–46] . They occupy a key phylogenetic position for addressing hypotheses of early deuterostome evolution . Previous studies in S . kowalevskii have demonstrated close transcriptional and signaling similarities with vertebrates during early A/P patterning of ectodermal development . The enteropneust body plan is divided into three main domains: a prosome/proboscis that is transcriptionally similar to the vertebrate forebrain , a mesosome/collar , similar to a midbrain , and a metasome/trunk , similar to a hindbrain and spinal cord [47 , 48] . It is this transcriptional network involved in ectodermal regionalization that is regulated by cWnt signaling in vertebrates and raises the obvious question of whether the establishment of the network is similarly regulated by cWnt in enteropneusts . An earlier study demonstrated that β-catenin is a critical component of AV patterning and plays a central role in specifying the endomesoderm [15] in a manner very similar to its early role in echinoderm and ascidian development . We also demonstrated that the early endomesoderm subsequently acts as an early organizer and defines the posterior of the embryo . Preliminary observations from this work revealed that the cWnt pathway was clearly an important component of early A/P patterning . The present manuscript explicitly investigates the role of cWnt signaling in the early specification of the A/P axis . Using a range of experimental approaches , we find strong support for a conserved role of cWnt signaling in the early establishment of the A/P axis in S . kowalevskii . We find three distinct regions in the ectoderm with differing responses to cWnt; first , as in many animals , cWnt signaling inhibits formation of the anterior region by down-regulation of anterior genes of the proboscis and anterior collar . Second , similar to vertebrates , it promotes mid-axial fates by up-regulating genes of the posterior collar , anterior , and mid-trunk . However , surprisingly the terminal posterior ectodermal domain around the blastopore , defined by overlapping expression domains of many Wnt ligands , is initially insensitive to cWnt signaling . We discuss the comparative implications of these findings . Our experiments give robust support for the hypothesis that this pathway was involved in early A/P axis specification deep in bilaterian evolution , predating the diversification of the deuterostome phyla . It also challenges the paradigm of a simple cWnt gradient involved in the specification of the entire A/P axis , and raises the possibility that the most posterior bilaterian ectodermal territory is specified independent of cWnt .
Adult S . kowalevskii were collected intertidally on Cape Cod , MA within the Waquoit Bay Reserve . Animal husbandry and culture techniques were comprehensively described previously [15 , 49] . Embryos were staged by the normal tables of Bateson [50 , 51] and Colwin and Colwin [52 , 53] . Classical embryology experiments and microinjections were carried out as described previously [15] . Targeted blastomere injections were performed under a stereomicroscope using a back-filled needle connected to a glass syringe with plastic tubing . The entire system is filled with mineral oil and the injection is performed manually with the syringe under visual control ( injected solution is colored by 1% fast green FCF ( F-7252; Sigma-Aldrich , St . Louis , MO ) . Injection was performed into identified blastomeres at cleavage stages . Injection success was monitored by co-injection of 1% rhodamine dextran ( D-1817; Molecular Probes , Eugene , OR ) . siRNAs targeting β-catenin and Fz5/8 were described previously [15 , 48] . The open reading frames of S . kowalevskii wnt3 , sfrp1/5 , and dkk1/2/4 were cloned into pCS2+ , then linearized and used for in vitro synthesis of capped RNA using the SP6 Message Machine kit ( Applied Biosystems/Ambion , Foster City , CA ) . The number of experimental embryos examined is indicated in figure panels . The cWnt pathway was activated using the GSK-3β inhibitor 1-azakenpaullone [54] ( 191500; Calbiochem , Sigma-Aldrich ) or the recombinant Wnt3a protein ( R&D Systems , Minneapolis , MN ) . Treatments were carried out as described previously [15 , 55] . The treatments led to robust and homogeneous effects as assessed by in situ hybridization on 2 to 20 embryos per probe . A total of 200 , 000 expressed sequence tags ( ESTs ) were screened from six libraries previously described [47 , 55 , 56] . Partial sequences were further cloned through library screening using PCR . Genbank accession numbers are as follows: wnt1 , EU931645; wnt2 , EU931646; wnt3 , EU931647; wnt4 , GU224244; wnt5 , GU076159; wnt6 , GU076160; wnt7 , GU076161; wnt8 , GU076162; wnt9 , GU076163; wnt10 , GU076158; wnt11 , GU076158; wnt16 , EU931648 . 1; wntA , GU224245; fz5/8 , GU075997; fz1/2/7 , MG711509; fz4 , MG711510; fz9/10 MG711511; dkk1/2/4 , GI:259013424; sclerostin , GU076111 . 1; sfrp1/5 , GU076117 . 1; sfrp3/4 MG682447; r-spondin , GU076102 . 1; notum , GU076072 . 1; wif , GU076157 . 1; dkk3 , MG682446 . Tentative orthology of ESTs was assigned by BLAST . Sequences were aligned with sequences from cnidarians and bilaterians using clustalX [57] . Gene tree analyses were carried out using Bayesean [58] and Neighbor-joining [59] algorithms to assign orthology relationships ( See S1 and S2 Figs ) . Experimental samples of 40 to 50 embryos each were frozen in liquid nitrogen and stored at −80°C . RNA was extracted using the RNAqueous-Micro Total RNA isolation Kit ( Life Technologies , Carlsbad , CA ) using a motorized pestle for initial homogenization of samples . cDNA synthesis was performed with Superscript III ( Life Technologies ) using 50 ng/μL total RNA in a 20 μL volume , using the oligo ( dT ) 20-primer according to manufacturer’s instructions . qPCR reaction-mix was set up using SsoAdvanced Universal SYBR Green Supermix ( Bio-Rad , Hercules , CA ) with 0 . 4 μM final concentration per primer and 0 . 066 ng/μL final cDNA concentration ( cDNA concentration is based on the initial 50 ng/μL total RNA concentration used during cDNA synthesis and dilutions of the cDNA synthesis mix thereafter ) . qPCR was performed in 96-well plates ( Bio-Rad ) with 9 μL total reaction volume per well using the CFX-Connect Real-Time System ( Bio-Rad ) running the following program: 95°C for 3 min for an initial melting , and 95°C for 10 s , 55°C for 40 s for 40 cycles , followed by a melting point analysis . All primer sets were initially optimized for efficiency at 55°C annealing and low probability of primer-dimer product in No-template controls . All cDNAs were tested in a dilution series to determine the area of linear amplification with a larger set of control primers . cDNAs generally behaved linearly with low standard errors to a final dilution of 0 . 00833 ng/μL or lower . Three technical replicates of each sample , a No-Template control , and RT− controls with a subset of primers , were performed for each plate and cDNA . actin , beta-tubulin , odc , and G3PDH were used for sample normalization . Actin was used for inter-plate normalization . Quantitative values of gene up- or down-regulation relative to control were calculated using the 2−ΔΔ CT method [60] using the Bio-Rad CFX Manager software version 3 . 1 . In situ hybridization was carried out as described [47 , 49] . Stained embryos were post-fixed in 10% formaldehyde in 1 X PBS overnight and then sequentially dehydrated into 100% EtOH , followed by several washes in 100% MeOH before clearing in MurrayClear reagent ( 2 parts Benzoyl benzoate and 1 part benzyl alcohol ) and mounted in Permount . Pictures were taken on a Zeiss Axioimager Z1 using a Zeiss Mrc5 camera . Image panels and figures were constructed with Adobe Photoshop and Adobe Illustrator .
To begin our investigation of the function of cWnt signaling during the development of hemichordates , we cloned and described the expression of key components of the signaling pathway ( ligands , receptors , and modifiers ) detected by whole mount in situ hybridization and reverse transcription PCR ( RT-PCR ) , over a range of developmental stages , from oocytes to juveniles . The Wnt ligands are a large family of secreted glycoproteins characterized by an invariant pattern of 22 to 24 highly conserved cysteine residues [61] . We cloned 13 Wnt genes from ESTs sequenced from three developmental stages , from both normalized and non-normalized cDNA libraries [56] , which include wnt1 , wnt2 , wnt3 , wnt4 , wnt5 , wnt6 , wnt7 , wnt8 , wnt9 , wnt10 , wnt11 , wnt16 , and wntA . All members of the proposed ancestral complement of 13 Wnt subfamilies are present in the hemichordate genome , including Wnt11 and Wnt2 that are absent from the sea urchin genome , and WntA , which has been lost from chordates [62] . We cloned one representative from each Wnt subgroup , and have not detected any paralogy duplicates by systematic screening of ESTs and genome assembly ( S1A Fig ) . We first present expression data for Wnt ligands by RT-PCR to determine the temporal expression profiles of each gene from oocyte to day 3 of development , when the main features of the hemichordate body plan are established ( Fig 1 ) . Maternal expression of two ligands , wnt4 and wnt9 , was detected in oocytes and early cleavage stages , and very low levels of wnt1 and wnt8 in oocytes . By late blastula/early gastrula and subsequent developmental stages , all Wnts except wntA were detectable by RT-PCR . We describe the patterns of expression for all 13 Wnt genes ( with the exception of wnt10 , which we failed to detect ) by whole mount in situ hybridization from mid-blastula ( 12 hpf [h postfertilization] at 20°C ) to day three of development ( Figs 2 and 3 ) . The first evidence of zygotic Wnt expression is at mid-blastula: five of the Wnts are detected by in situ hybridization ( wnt4 , wnt6 , wnt8 , wnt2 , and wnt11 ) . Their expression is detected in circumferential bands , strongest at the intersection of the AV hemispheres in the region fated to become the blastopore ( Figs 2Di , 2Ei , 3Ai , 3Bi and 3Ci ) . All five genes are expressed in broad overlapping domains in the animal hemisphere , but none are expressed at the animal pole , the prospective far anterior region . Expression is detected exclusively in the ectodermal , animal hemisphere precursors , not in the vegetal endomesodermal precursors at pregastrula stages . During gastrulation , the expression domains of most of the ligands become more spatially restricted , but generally they retain their relative spacing and order along the newly forming A/P axis of the ectoderm ( Figs 2 and 3 ) . As has been described in other species , many Wnt ligands are expressed around the forming blastopore; wnt1 and wnt4 ( Fig 2Cii and 2Dii ) are both expressed around the inside of the blastopore lip at the boundary of ectoderm and endomesoderm , whereas wnt3 , wnt6 , and wnt16 ( Fig 2Bii , 2Eii and 2Fii ) are expressed further anteriorly in the ectoderm at the exterior edge of the blastopore lip . The remaining expression domains are located further anteriorly in the ectoderm; wnt8 , wnt2 , wnt7 , and wnt5 are all expressed anterior to wnt3 in broadly overlapping domains ( Fig 3Aii , 3Bii , 3Dii and 3Eii ) . Wnt11 has the most anterior limit of ectodermal expression at this stage , but is still restricted from the most apical region ( Fig 3Cii ) . As the embryos elongate after gastrulation , between 36 and 48 h of development , expression of the ligands becomes far more spatially restricted along the A/P axis . Most ligands remain expressed exclusively in the ectoderm with a few exceptions . Two of the ligands are expressed in the two posterior coeloms: wntA ( Fig 3Fiv and 3Fv ) and wnt9 ( Fig 3Giii–3Gv ) , and others are expressed in the endoderm: wnt3 into the posterior endoderm ( Fig 2Biii ) , wnt9 in dorsal , anterior endoderm , around the forming gill slits ( Fig 3Giii–3Gv ) , wnt16 in the ventral posterior endoderm ( Fig 2Fiii ) , and wnt5 at low levels throughout the endoderm at 48 h of development , which later refines to the forming gill slits ( Fig 3Eiv and 3Ev ) . wnt1 , 3 , 4 , and 6 ( Fig 2Biii , 2Ciii , 2Diii and 2Eiii ) all retain their early expression around the blastopore . However , wnt1 , 4 , 6 , and 16 develop additional expression domains around the anterior metasome ( Fig 2Civ , 2Diii , 2Eiii and 2Fiv ) . wnt8 , 2 , 11 , 7 , and 5 all begin to refine their single expression domains to the boundary between the prosome and metasome following gastrulation ( Fig 3Aiii , 3Aiv , 3Biii , 3Biv , 3Ciii , 3Civ , 3Diii , 3Div , 3Eiii and 3Eiv ) , and by day three of development , some of these expression domains are sharply localized in narrow regions , such as in the case of wnt8 , 2 , and 7 ( Fig 3Av , 3Bv and 3Dv ) . wnt7 is not expressed as a single circumferential domain , and is down-regulated along the ventral midline by 48 h of development ( Fig 3Div ) , and up-regulated in a dorsal spot at the base of the proboscis ( Fig 3Dv ) . By day three of development , all ligands with the exception of wntA and wnt10 are detected in one or more of three ectodermal domains of expression; the base of the prosome , the boundary of the metasome and trunk at the developing first gill slit , and in the posterior ectoderm . In summary , the zygotic expression of the Wnt ligands begins at the midblastula stage in the animal hemisphere , close to the boundary with the vegetal hemisphere . Throughout gastrulation , ligands are expressed broadly in the ectoderm from the blastopore to the more anterior regions of the ectoderm , but are always excluded from the most apical/anterior regions . By later developmental stages , this expression largely refines to spatially restricted domains marking a region in the posterior ectoderm , over the developing first gill slit , at the boundary between the trunk and collar , and finally at the boundary between the developing proboscis and the collar . Expression is also detected in the posterior endoderm and mesoderm . An increasing number of modifiers of the cWnt pathway have been described in vertebrates [2 , 63] . Many of these genes encode secreted proteins , which act on the pathway by a variety of means and can act to both potentiate and inhibit cWnt signaling . Comparative studies on the expression and developmental roles of these proteins are far less extensive than studies on the Wnt ligands themselves . We have isolated and determined the expression pattern of several Wnt modifiers in S . kowalevskii and compared their expression and function to those reported in other animals in order to identify possible evolutionarily conserved roles of these pathway modifiers . Three antagonists are described below , and the other modifiers are described in S1 Text and S3 Fig . We cloned four representatives of the Fz receptor family characterized by an extracellular domain , including a signal peptide , and a cysteine-rich Wnt-binding domain , seven transmembrane domains , and a cytoplasmic tail [78] . A genomic survey of the current genome assembly identified only four Fz genes . Previous studies have described four ancestral Fz subgroups: Fz1/2/7 , Fz4 , Fz5/8 , and Fz9/10 [62] . The four Fzs from hemichordates exhibit robust orthology to these subgroups already present in cnidarians ( S1B Fig ) and build further support to the hypothesis that four Fzs are ancestral to the eumetazoan lineage [79] . Similar to the findings from echinoderms , which are a sister group to hemichordates , we were not able to isolate a member of the Fz3/6 group , which is present in vertebrates but absent from cnidarians and sea urchins . This supports the hypothesis that the group arose by duplication during chordate evolution [62] . Partial expression profiles of fz5/8 in S . kowalevskii and early embryonic and larval expression in P . flava have been previously described in the anterior of both embryos [48 , 73] . By RT-PCR , two Fzs , fz5/8 and fz1/2/7 are detected at high levels maternally in newly fertilized oocytes and early cleavage stages ( Fig 1 ) . By late blastula/early gastrula and subsequent developmental stages , all Fzs are expressed . Their expression , like the Wnt ligands , is also highly regionalized along the A/P axis during all developmental stages , suggesting that they play region-specific roles during the patterning of the embryonic A/P axis . The most anteriorly expressed of the receptors is fz5/8 , detected midway through blastula from the animal pole throughout most of the animal hemisphere , but excluded from the most vegetal region of that hemisphere , the region fated as posterior ectoderm ( Fig 5A ) . As gastrulation begins , the expression is increasingly restricted to the most anterior ectoderm . As the vegetal hemisphere invaginates and contacts the anterior ectoderm midway through gastrulation , fz5/8 expression is initiated in the anterior endomesoderm that is fated to become the proboscis/prosome mesoderm , at the same A/P level as the ectodermal expression ( Fig 5C ) . As gastrulation continues , the expression is now clearly restricted in the anterior ectoderm of the prospective prosome , and continues in this domain [48] throughout all stages examined . This expression resembles the early expression of fz5/8 in both P . flava , sea urchin , and the anterior localization of fz8 during mouse development [72 , 73 , 80] . Short interfering RNA ( siRNA ) knockdown of fz5/8 resulted in abnormal patterning of the anterior ectoderm with the expansion of apical markers at the expense of more posterior proboscis markers , demonstrating the importance of Wnt signaling for the posteriorization of the anterior ectodermal territory [48] . fz4 is expressed in a highly restricted domain from the onset of zygotic expression during gastrulation ( Fig 5F ) . Expression is uniquely ectodermal and is localized to a narrow ring in the more rostral region of the anterior ectoderm . This domain refines to a narrow ring at the boundary between the prosome and mesosome , at the posterior boundary of the expression limit of fz5/8 ( Fig 5G and 5H ) . This to some extent resembles the expression of fz4 in chick and mouse , where it is localized in the diencephalon at the boundary with the telencephalon [80] . fz1/2/7 has a much broader domain of expression; it is first detected at blastula stage throughout the animal hemisphere , more broadly than that of fz5/8 ( Fig 5I ) . By the beginning of gastrulation , its expression has been largely down-regulated in the anterior ectoderm and around the blastopore , but remains in the midectoderm ( Fig 5J ) . Expression is also detected in the forming mesendoderm . By 36 h after fertilization , transcripts are detected in overlapping domains with both fz4 and fz5/8 in the anterior domain , with expression mainly in the prospective mesosome , expanding down into the prospective trunk , but not past the ciliated band into the blastoporal area ( Fig 5K ) . By day three of development , most transcripts are detected in the mesosome and at the base of the proboscis , with only weak expression in the anterior trunk . At this late stage , expression is now visible in the anterior mesoderm of the proboscis ( Fig 5L ) . fz9/10 expression begins at blastula in the most vegetal portion of the animal hemisphere ( Fig 5M ) . At gastrulation , expression remains in a similar domain in the posterior ectoderm , just anterior to the forming ciliated band ( Fig 5N ) , and by 36 h of development , the expression domain has expanded anteriorly to the entire trunk and part of the collar , overlapping extensively with the expression domain of fz1/2/7 ( Fig 5O ) . By day three of development , expression is no longer detectable . The interactions of Fzs and Wnts is poorly understood , and it is proposed that the overlapping expression of many of the vertebrate Fzs leads to functional redundancy [81 , 82] . In S . kowalevskii , Fz expression is highly regionalized along the A/P axis and suggests that the different receptors are responsible for patterning different regions of the axis . Notably , none of the receptors are expressed around the blastopore , where many of the ligands are localized . The functional significance of this observation is explored in later experiments . The expression of Wnt ligands and their antagonists is strongly localized along the developing A/P axis of S . kowalevskii , and their relative domains are similar in location to their orthologues during vertebrate development; ligands are expressed posteriorly and antagonists are expressed anteriorly , with Fz receptors expressed in staggered domains ( Fig 6 ) . We carried out a series of experiments to test whether the cWnt pathway is involved in the early specification of the hemichordate A/P axis . To test for the timing of Wnt activity during early A/P patterning , we further developed the 1-azakenpaullone experiments . Concurrent treatments were carried out by adding the inhibitor every 2 h , with the first beginning at early blastula ( 11 hpf ) , and the treatment ending at 64 hpf ( S6 Fig ) . Embryos were cultured at 20°C and fixed at 96 hpf when all the major body divisions had formed , and the A/P and dorsoventral ( D/V ) axes were clearly morphologically differentiated . Graded phenotypes were observed and correlated with the treatment initiation time; the more severe phenotypes resulted from earlier and longer treatments ( S6I Fig ) . The first treatment initiated at 11 h led to the most dramatic phenotype that we have observed . The embryos exhibited major morphological defects; all anterior markers down to anterior trunk failed to express , even the anterior trunk marker en ( S6ii Fig ) , but msx—normally a broad trunk marker—was expressed up to the most anterior region of the embryo ( S6Gii Fig ) . hox9/10 , a posterior trunk marker , was expressed almost up to the anterior of the treated embryos ( S6Hii Fig ) . Later treatments produced weaker phenotypes with a gradual restoration of molecular marker expression from posterior to anterior . hox9/10 expression was restricted posteriorly for treatments starting at 13 hpf or later , and msx expression did not expand anteriorly for treatments starting at 17 hpf or later ( S6G and S6H Fig ) . These experiments suggest that the ectoderm is most sensitive to cWnt activation during the blastula and gastrula stages . It is the period when Wnt genes are expressed broadly in the ectoderm ( Fig 6 ) . We have further analyzed the above experiment by examining gene expression at 21 hpf ( onset of gastrulation ) , when little morphogenesis has occurred ( S6B–S6D Fig ) . A similar gradation in phenotypes was observed for the anterior markers sfrp1/5 and six3 . Interestingly , expression of the posterior marker hox9/10 was unchanged even for the earliest treatment starting at 11 hpf , while we observed an ectopic expression anteriorly when its expression was analyzed at 96 hpf ( compare S6Dii and S6Hii Fig ) . To better understand this discrepancy , we treated embryos with a range of 1-azakenpaullone concentrations starting at early blastula ( 12 hpf ) and analyzed A/P ectodermal markers expression using quantitative PCR ( qPCR ) at early gastrula stages ( 24 hpf; Fig 8A ) . This restricted the treatment to early establishment of the A/P axis . The second treatment again began at 12 h but was extended to 48 h into early embryo elongation from 24 to 48 h , after the early specification of the A/P axis ( 48 hpf; Fig 8B ) . For both treatments , proboscis/prosome markers were strongly repressed ( Fig 8A and 8B ) . Collar markers were initially strongly down-regulated , whereas their later expression was either moderately activated or repressed depending on the inhibitor concentration ( Fig 8A and 8B ) . This observation suggests that anterior suppression regulates fates down to the collar and its exact position varies according to the concentration of 1-azakenpaullone . All anterior/central trunk markers are activated in a dose-dependent manner at both stages . Early expression of posterior Hox genes , during the initial establishment of the posterior territory , is insensitive to cWnt activation , with the exception of hox11/13b . By contrast , the same genes are strongly activated when their expression is analyzed at 48 hpf . We confirmed this global gene expression analysis by in situ hybridization of selected markers ( Fig 8C ) . The anterior gene six3 is not detected following either treatment . The expression of the trunk marker msx was expanded anteriorly for both treatments while the posterior marker hox9/10 was only ectopically activated when analyzed at 48 hpf during embryo elongation following gastrulation . Overall , the results from this section indicate that A/P markers can be organized into three groups according to their sensitivity to cWnt activation: ( 1 ) anterior genes are repressed ( proboscis ) or activated or repressed depending on concentration ( collar ) , ( 2 ) intermediate genes ( anterior/central trunk ) are activated , and ( 3 ) posterior genes ( posterior trunk ) are initially insensitive before being activated following gastrulation during embryo elongation . In hemichordates , we have shown that endomesoderm is the source of early signals posteriorizing the ectoderm that otherwise adopt an anterior character ( Fig 9 ) [15] . We tested whether activation of the Wnt pathway was sufficient to posteriorize naive ectoderm explants that lack inputs from the endomesoderm . Embryos were cut at the 32-cell stage and the animal hemispheres cultured in isolation . They developed anterior fates demonstrated by the expression of the apical marker foxQ2-1 throughout the explant ( Fig 9Biii ) , and failed to express posterior transcriptional markers such as en , msx , and hox9/10 ( Fig 9Bvii , 9Bxi and 9Bxv ) . However , treating these explants with 1-azakenpaullone from midblastula stages ( 12 h of development ) resulted in repression of the apical marker foxQ2-1 ( Fig 9Biv ) and the ectopic activation of the more posterior markers , msx , and en ( Fig 9Bvii and 9Bxii ) throughout the ectoderm . However , we were unable to activate the most posterior marker hox9/10 ( Fig 9Bxvi ) . The expression of this marker was actually unchanged upon cWnt activation in whole embryos ( Fig 9Bxiv ) , whereas expression of both en and msx was expanded ( Fig 9Bvi and 9Bx ) . These results are entirely consistent from the 1-azakenpaulone treatments in intact embryos , and further suggest that there is a differential response to Wnt signaling along the A/P axis , with three distinct domains . We carried out reciprocal experiments to test for the effects of reducing Wnt activity . We first injected capped mRNA of the Wnt antagonists dkk1/2/4 and sfrp1/5 . Both overexpressions led to very similar phenotypes; expansion of the anterior proboscis territory and reduction in the size of the trunk ( Fig 10i , S1 and S2 Movies ) , which is the reciprocal phenotype from Wnt3 overexpression ( Fig 7Hi–7Iiii ) . The phenotypes differed in that the proboscis took on a more bulbous appearance with a narrower trunk following dkk1/2/4 injection ( Fig 10iF and 10iH ) when compared to sfrp1/5-injected embryos ( Fig 10iB and 10iD ) . This may be due to the difference in specificity of the secreted antagonists as Dkk1/2/4 specifically antagonizes the canonical pathway by binding to Kremen/lrp , whereas Sfrp1/5 binds the Fz receptor and inhibits all Fz-mediated Wnt signaling . The molecular markers support the phenotypic transformations with both foxQ2-1 and six3 expanding in injected embryos ( Fig 10iB , 10iF and 10iH ) , and the trunk marker en expressed further posteriorly with weaker expression when compared to controls following injection of sfrp1/5 ( Fig 10iD ) , consistent with down-regulation of this trunk marker . Previous data have revealed that knockdown by siRNA of fz5/8 results in an anteriorized phenotype , as assayed by the expansion of the apical marker SkFGF-1; the effect is restricted to the proboscis matching the expression domain of fz5/8 [48] . We further tested this phenotype by assaying the expression of another apical marker foxQ2-1 and show a similar expansion ( Fig 10iI and 10iJ ) . Conversely , the expression domain in experimental embryos of rx , a proboscis ectodermal marker normally excluded from the most apical domain , is restricted to a more posterior domain of expression ( Fig 10iK and 10iL ) [48] , suggesting that a gradient of Wnt activity is involved in patterning the proboscis ectoderm . The overexpression of Wnt antagonists results in a phenotype complementary to cWnt activation , but trunk fates cannot be repressed entirely . Because this could result from an incomplete blockade of cWnt activity , we performed additional experiments utilizing targeted injections of siRNAs designed to β-catenin . However , β-catenin is necessary and sufficient to specify endomesoderm , and injection of siRNA against β-catenin before the first cell cycle disrupts the establishment of the AV axis , resulting in fully animalized embryos that fail to gastrulate [15] . To avoid disrupting germ layer formation , β-catenin siRNA was injected into single blastomeres at the 4-cell stage . The first two cleavage planes occur along the AV axis and define the left/right and dorso-ventral axes , respectively [53] . Gastrulation and endomesoderm formation ( revealed by foxA expression ) were relatively normal if injection was delayed until the 4-cell stage ( Figs 10ii and S7 ) . Expression of foxQ2-1 , which is normally restricted to the anterior-most ectoderm ( Fig 10iiA ) , expanded posteriorly in descendants of blastomeres injected at the 2 , 4 , and 8-cell stage ( Fig 10iiB and 10iiC and S7 Fig ) . Notably , expression did not expand down to the presumptive posterior ectoderm , despite the presence of the siRNA in all ectodermal descendants in the injected quadrant ( Fig 10iiC ) . Expression of the trunk marker msx was lost in all descendants of injected cells ( Fig 10iiE and 10iiF ) , while the posterior marker hox9/10 was activated even in the absence of cWnt signalling ( Fig 10iiH and 10iiI ) . These targeted β-catenin siRNA experiments further support that there is a threshold of Wnt sensitivity in the trunk ectoderm . Wnt inhibition is sufficient to ectopically expand the most anterior genes , but only the presumptive anterior ectoderm can expand following cWnt suppression . Anterior/midtrunk gene expression requires active cWnt signaling , whereas posterior genes such as hox9/10 are unaffected by suppression of cWnt during the early establishment of A/P pattern ( Fig 11 ) . The inability to anteriorize posterior ectoderm by elimination of Wnt signaling , and the early activation of posterior genes even in the absence of cWnt signaling is likely to be due to non-cWnt posteriorizing signals emitted by endomesoderm .
A variety of analyses from genome and EST projects have established that the ancestral bilaterian complement of Wnts was likely 13 distinct families . Comparisons between bilaterian clades have revealed both Wnt gene diversification and loss during bilaterian body plan evolution . Within the deuterostomes , in chordates , the ancestral complement of Wnt gene families is 12; amphioxus has 13 ligands , and humans have 19 , which can be classified into 12 distinct subfamilies [83] . The only subgroup missing from the vertebrates and amphioxus is WntA . In the tunicates , data from ascidians suggest that several Wnts were lost during the extensive genomic rearrangements in the tunicate lineage . Outside of chordates , 11 Wnts have been isolated from sea urchins with wnt2 and 11 absent from the genome of S . purpuratus [62] , and RNAseq data suggests this is conserved in another species of urchin [72] . From comprehensive arthropod sampling and recent work on onychophorans , along with limited lophotrochozoan sampling , the likely ancestral repertoire of protostome Wnt genes was 12 , excluding wnt3 [84 , 85] , with several groups showing secondary losses of individual ligands . The most parsimonious reconstruction of the Wnt gene family complement in bilaterians is 13 . Sequence data from S . kowalevskii conclusively supports this hypothesis , and provides the first data for a bilaterian possessing all of the 13 Wnt ligand families . Comparative studies from placozoans , sponges , and ctenophores suggest that two Fzs were ancestral to metazoans [6 , 13 , 79 , 86] . Data from sea urchins [62] , sea stars [87] , chordates , hemichordates ( our study ) , and annelids [85] support the conclusions from cnidarians that the Fz complement fz5/8 , fz1/2/7 , fz4 , and fz9/10 is the ancestral state for bilaterians and eumetazoans , and that the Fz3/6 group present in chordates represents a lineage-specific duplication from the Fz1/2/7 group [62 , 79] . We cloned and characterized a range of extracellular Wnt modifiers . Of the characterized Wnt antagonists , we cloned dkk1/2/4 , dkk3 , sfrp1/5 , sfrp3/4 ( frzb ) , WISE/sclerostin , wnt inhibitory factor ( wif ) , and notum/wingful , and of the Wnt agonists , we cloned r-spondin . Comparative genome surveys that explicitly discuss extracellular Wnt modifiers are scarce , and outside of chordates there are few reports of the presence or absence of these genes in other groups . Notably , a comprehensive survey in sea urchins [62] has identified all the Wnt modifiers that we identified except sclerostin , and thus S . kowalevskii has many of the antagonists that have been described in a variety of vertebrate species . The current data from hemichordates , when combined with data from other metazoans , provides strong support for the hypothesis that cWnt signaling was intimately involved in the early evolution of the A/P axis . In many bilaterian groups , Wnts have been proposed to play a conserved role in anterior suppression and posteriorization , and strong supporting experimental evidence has been generated from chordates , sea urchin , annelids , planarians , and acoels [20 , 23 , 34 , 37–40 , 43] ( reviewed in [10 , 11 , 14] ) . Functional data from cnidarians also demonstrates that the evolution of the role of Wnts in axis polarization was likely a very deep innovation of metazoan embryos [77 , 88–90] , even prior to bilaterians . We present comprehensive expression data for the entire complement of Wnt ligands throughout all early developmental stages ( Figs 2 and 3 ) , as well as of Wnt modifiers ( antagonists and agonists ) ( Fig 4 , S3 Fig ) and Fzs ( Fig 5 ) . The relative expression domains of Wnt ligands and their antagonists resemble those of other bilaterians at opposite ends of the forming axis ( Fig 6 ) . In S . kowalevskii , the predominant expression of Wnt ligands posteriorly , excluded from the most anterior domains , the broad expression of the four fz genes in localized ectodermal domains down the A/P axis , and the predominant localization of Wnt antagonists in the anterior of the embryo , resembles expression patterns of many different metazoan embryos [11] . However , our data contrast with the comprehensive expression data from sea urchins , which shows most of the ligand expression domains in early development in the endomesoderm , with a subset of the ligands with localized domains in the ectoderm [72] . In this study , we focus on the role of Wnts in ectodermal patterning , but it is important to note that Wnt ligands , Fzs , and Wnt modifiers are also expressed in the endoderm and mesoderm ( Figs 2 , 3 and 5 ) , and likely also play important developmental roles in those germ layers ( S4 Fig ) . Although the relative localization of the ligands and their antagonists during A/P patterning is perhaps not surprising based on other comparative studies , the functional data we present , both embryological and molecular , show some unexpected results . While we have not addressed the function of individual Wnt ligands or identified which ligands activate the cWnt pathway , our data reveal close similarities to the regulatory role of cWnt during vertebrate neuraxis A/P patterning that are not shared by invertebrate chordates , and some surprising results for the role of cWnt in posterior specification . In both vertebrates and S . kowalevskii , the ectoderm is initially fated to become anterior unless exposed to posteriorizing signals from the endomesoderm . Early embryonic manipulation of cWnt signaling , either positively or negatively , results in dramatic and complementary changes to the relative proportions of the major body regions , the proboscis , collar , and trunk ( Figs 7 and 10 ) . Over-activation of the pathway or overexpression of Wnt3 ligand results in reduction or loss of the proboscis and anterior collar ( Fig 7 ) , and overexpression of secreted antagonists of the cWnt pathway results in an over expansion of the anterior proboscis fates and reduction in the size of the trunk ( Fig 10 ) . The severity of anterior truncation from Wnt over-activation was concentration-dependent; a high level of 1-azakenpaullone resulted in loss of both anterior collar and proboscis fates , whereas lower concentrations resulted in loss of only the proboscis markers ( Fig 7 ) . This supports a model of a Wnt gradient , specifying the A/P pattern of the anterior embryo as has been proposed in vertebrates [20] . Despite the morphological disparity between vertebrates and hemichordates , cWnt is regulating the same region of the transcriptional network that defines the ectodermal A/P axis in both species . Proboscis/anterior collar markers , like forebrain/midbrain markers , are down-regulated , whereas posterior collar/pharynx/anterior trunk markers , like hindbrain markers , are up-regulated by cWnt signaling . The timing of Wnt activity during this early phase of A/P specification is also very similar between vertebrates and hemichordates . In S . kowalevskii , the onset of zygotic Wnt expression at midblastula ( 11 h ) corresponds to the onset of posteriorizing activity . This was determined by timed treatments with the GSK3β inhibitor 1-azakenpaullone ( S6 Fig ) . Treatments beginning at early blastula , just before the onset of zygotic Wnt expression , have the most severe anterior truncation , and this time period likely represents the onset of A/P patterning . Earlier treatments impact AV patterning and the amount of endomesoderm specified by β-catenin . Incremental delays of onset of exposure result in less severe anterior truncations through gastrulation , suggesting that the anterior ectoderm becomes increasingly refractive to cWnt with time . A key finding of our work is the subdivision of the ectoderm into three domains regarding their sensitivity to cWnt ( Fig 11 ) . Below , we compare these observations with results from other metazoans . The S . kowalevskii anterior-most ectoderm is defined by a set of transcription factors ( six3 , rx and , foxQ2-1 ) , whose anterior expression is conserved over a broad phylogenetic range [40] . Early cWnt activation or ligand overexpression leads to anterior marker inhibition and a “proboscis-less” embryo ( Figs 7 and 8 , S5 and S6 Figs ) , and overexpression of anteriorly localized Wnt antagonists results in expansion of this anterior territory ( Fig 10 , S7 Fig ) . This demonstrates that cWnt needs to be actively inhibited to allow correct patterning of anterior territories . However , the repressive effect of cWnt is likely tightly regulated in normal development in the anterior ectoderm to produce graded Wnt levels and correctly pattern proboscis and collar fates . qPCR data indicates that the most anterior fates are the most strongly inhibited by cWnt activity , with more caudal proboscis and collar markers less strongly down-regulated supportive of a local repressive gradient in the anterior ectoderm ( Fig 8B ) . In addition , loss-of-function data suggest some anterior markers require some cWnt signaling for transcription; knock down of fz5/8 demonstrated that rx expression contracts posteriorly when Wnt signaling is locally repressed in the anterior ectoderm , and anterior clones of β-catenin-deficit cells fail to express rx , but activate ectopically the apical marker FGFsk-1 [48] . Further , more sophisticated experiments would be required to further dissect this issue . Moreover , the initial broad expression of anterior markers in the ectoderm together with the observation that ectodermal cells separated from the endomesoderm at blastula stages develop into anterior-most ectoderm [15] suggests that anterior identity could correspond to a default or intrinsic fate . In comparison with vertebrates , it appears that the hemichordate embryo does not depend on an endomesodermal source of Wnt antagonists for anterior patterning; the anterior ectoderm suffices . The implication of cWnt in anterior suppression has been documented in a wide range of phylogenetically diverse groups of animals from chordates , echinoderms , arthropods , and annelids [20 , 36 , 40 , 41 , 97] . A similar situation may operate during regeneration in acoels and planarians and for the specification of the aboral pole in cnidarians [37–39 , 43 , 88] . We have shown that cWnt is necessary and sufficient for the activation of markers of the midaxial identities in the posterior collar/anterior trunk along with anterior and central class Hox genes ( Figs 8 and 9 ) . This territory corresponds transcriptionally to the chordate hindbrain and anterior spinal cord [47] . Although anterior suppression has been broadly demonstrated in bilaterians , a role for cWnt in promotion of more posterior territories by up-regulation of midaxial markers , rather than by repression of anterior ones , comes mainly from vertebrates in which Wnts promote hindbrain and anterior spinal cord fates . We propose that this region of both embryos represents a homologous embryonic midaxial territory and is regulated by a cWnt-dependent regulatory program . Importantly , this conclusion could not have been strongly supported with current knowledge from nonvertebrate deuterosomes . In indirect-developing larval species of echinoids , asteroids , and hemichordates , most of the larval ectoderm is fated to be anterior [34 , 91 , 92] , without an equivalent midaxial trunk territory . Consequently , the experimental focus in echinoderms has been mainly on the establishment and patterning of the anterior neural territory [34 , 35] . Nevertheless , in urchin larvae , Wnt5 activates a posterior ectodermal marker in the larval ectodermal [93] , suggesting some role of Wnt , even in a body plan without a trunk . Strong similarities have not been found for tunicates , and it is possible that this patterning device has been greatly modified or changed its function . In cephalochordates , however , cWnt activation by Gsk3β inhibitor treatments leads to an expected anterior truncation , but the reported effect is rather moderate , affecting only the most anterior region of the neural tube , and it is thought that most of the axis is patterned independent of cWnt signaling [29] . Our data is thus very similar to vertebrate neural data except that patterning influences the entire ectoderm rather than a localized central nervous system . We thus propose that cWnt has been instrumental in the specification of the anterior axial and midaxial identities at least at the base of the deuterostomes , and that secondary modifications have occurred in echinoderms and invertebrate chordate lineages . Outside of deuterostomes , there is support for a role of cWnt in trunk identity during planarian regeneration [94 , 95] , but most other experimental studies in arthropods deal with posterior growth rather than initial establishment of the A/P axis [18] . We have previously reported that in S . kowalevskii , endomesoderm sends posteriorizing signals to the ectoderm in mid/late blastula stages that would otherwise adopt an anterior identity [15] . Gain- and loss-of-function experiments presented here demonstrate that initiation and early expression of posterior trunk markers such as hox9/10 , hox11/13a , c , are insensitive to cWnt signaling . Moreover , targeted knock down of β-catenin by siRNA starting at the 2- , 4- , and 8-cell stages only expands the expression domains of anterior markers down to a region corresponding to the midtrunk , suggesting that the posterior trunk is resistant to anteriorization by cWnt inhibition ( Fig 10ii , S7 Fig ) . These results suggest that posterior ectoderm identity is specified independent of cWnt and that other factors are involved in specification of posterior trunk . In vertebrates , Nodal , bone morphogenetic protein , fibroblast growth factor , and retinoic acid signaling are all involved in this process [96 , 97] and we are currently investigating the roles of these pathways during the early development of S . kowalevskii . By combining the results of both gain- and loss-of-function , we tentatively mapped the boundary between cWnt-dependent and cWnt-independent ectodermal domains at the anterior limit of hox9/10 expression ( Fig 11 ) . Based on similarities in gene expression , this limit roughly corresponds to the posterior spinal cord in vertebrates . Our results are very similar to what is observed in comparable experiments performed in vertebrates ( reviewed in [27 , 98] ) When X . laevis animal caps are cut following RNA injection of xwnt3a and noggin , the forebrain markers XAG-1 , xanf-2 , and otxA were inhibited , and the hindbrain markers markers en-2 and krox-20 were induced . However , this approach does not activate the spinal cord marker hoxB9 [22 , 99] . Moreover , when the Wnt pathway is activated in whole X . laevis or chick embryos , it results in posteriorization of the anterior neural plate and expansion of hindbrain markers , whereas repression of cWnt has the reciprocal phenotype; repression of hindbrain fates and expansion of forebrain fates . However , posterior spinal cord markers were not examined in these studies , making it difficult to make firm conclusions about the role of Wnt signaling in posterior neural territories [20 , 22 , 23 , 25 , 96 , 99 , 100] . Our results suggest that the early specification of deuterostome posterior embryonic fates may be more dependent on the activity of other secreted ligands and that cWnt plays a less important role than previously proposed . We argue that the consideration of the role of cWnt signaling in posteriorization largely conflates several developmentally distinct roles of Wnts during AV and A/P patterning . Three distinct roles of the cWnt pathway in early embryonic axial patterning of metazoans have been proposed to have deep evolutionary origins in animal evolution ( reviewed in [10 , 11 , 14] ) , and so far no single species displays all three of these . A clear developmental distinction may help for making coherent hypotheses about the role of this complex pathway in the establishment of metazoan axes . The role with the earliest onset in development involves the action of β-catenin in the establishment of the AV axis and the specification of the endomesoderm . This role has been described in sea urchins , hemichordates , cnidarians , and nemerteans [15–17 , 101] . We described elsewhere this conserved function in S . kowalevskii [15] . Importantly , we have shown that endomesoderm specification by β-catenin is essential for subsequent ectoderm A/P patterning . We propose that endomesoderm sends posteriorizing signals to the overlying ectoderm that on its own harbors a “default” anterior fate that is independent of Wnt signaling . These unidentified signals are likely controlling posterior identity and posterior trunk formation , and are responsible for the initial cWnt insensitivity of this part of the embryo ( Fig 11 ) . The second role is dealt with explicitly in this study and involves the role of cWnt in the early establishment of the A/P axis . The role of cWnt in anterior suppression has broad phylogenetic support . In deuterostomes , posteriorization of ectodermal fates by cWnt activity has been firmly established in vertebrate CNS in multiple species [20 , 23 , 97] , and demonstrated to a limited extent in echinoids [93] , and now in hemichordates . Whether this function of cWnt is broadly conserved in bilaterians is still unclear because it has not been formally investigated during embryogenesis in animals in which anterior suppression is evidenced; however , work in planarian and acoel regeneration are supportive of a broader role in posterior specification [37 , 38 , 43 , 94] . Finally , the third role of the cWnt pathway is later in development , following the early crude establishment of the A/P axis . Both arthropods and vertebrates deploy a conserved regulatory network of genes localized in a terminal growth zone mediated by the action of Wnts , which likely represents an ancestral developmental strategy for posterior growth ( reviewed in [11 , 18] ) . We are currently investigating a later role of Wnts in posterior extension of the trunk in hemichordates . Following the early establishment of A/P polarity , hemichordates undergo an extended period of posterior growth to elongate the trunk . Wnt genes continue to be expressed in the posterior and potentially mediate this morphological extension , which mechanistically may be homologous to posterior growth in arthropods and chordates . If this is supported by further experiments , then early development in S . kowalevskii would be regulated by all three of the proposed conserved roles of cWnt signaling in axial patterning , endomesoderm specification , early establishment of A/P axis , and posterior growth .
|
The anteroposterior ( A/P ) axis is a conserved feature of bilateral animals and is defined in the anterior by a head and in the posterior by a trunk . The secreted family of Wnt proteins and their antagonists play a conserved role in setting up this axis during early development in many bilaterians . The widely accepted model for the role of Wnt signalling in A/P axis specification is by the establishment of a simple activity gradient whereby high levels of Wnt lead to posterior fates and low levels to anterior fates . In this study we further test this model , examining the role of Wnt signaling in the acorn worm , a representative of hemichordates that belongs to the superphylum Deuterostomia along with chordates . We find strong evidence supporting the hypotheses that Wnt signaling represses anterior fates and promotes more caudal fates in the ectoderm , in a similar way to central nervous system development in vertebrates However , we find that the most posterior territory is established independently of Wnt signalling , which is inconsistent with the prevailing model of Wnt function in A/P patterning . We conclude that these data are not inconsistent with vertebrate patterning data and may represent a conserved feature of deuterostome axis patterning .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"taxonomy",
"vertebrates",
"animals",
"blastulas",
"animal",
"taxonomy",
"developmental",
"biology",
"data",
"management",
"embryos",
"zoology",
"ectoderm",
"embryology",
"computer",
"and",
"information",
"sciences",
"gastrulas",
"signal",
"transduction",
"eukaryota",
"acorn",
"worms",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"wnt",
"signaling",
"cascade",
"hemichordata",
"cell",
"signaling",
"organisms",
"signaling",
"cascades"
] |
2018
|
Anteroposterior axis patterning by early canonical Wnt signaling during hemichordate development
|
Type 1 Serine/Threonine Kinase Receptors ( STKR1 ) transduce a wide spectrum of biological signals mediated by TGF-β superfamily members . The STKR1 activity is tightly controlled by their regulatory glycine-serine rich ( GS ) domain adjacent to the kinase domain . Despite decades of studies , it remains unknown how physiological or pathological GS domain modifications are coupled to STKR1 kinase activity . Here , by performing molecular dynamics simulations and free energy calculation of Activin-Like Kinase 2 ( ALK2 ) , we found that GS domain phosphorylation , FKBP12 dissociation , and disease mutations all destabilize a D354-R375 salt-bridge , which normally acts as an electrostatic lock to prevent coordination of adenosine triphosphate ( ATP ) to the catalytic site . We developed a WAFEX-guided principal analysis and unraveled how phosphorylation destabilizes this highly conserved salt-bridge in temporal and physical space . Using current-flow betweenness scores , we identified an allosteric network of residue-residue contacts between the GS domain and the catalytic site that controls the formation and disruption of this salt bridge . Importantly , our novel network analysis approach revealed how certain disease-causing mutations bypass FKBP12-mediated kinase inhibition to produce leaky signaling . We further provide experimental evidence that this salt-bridge lock exists in other STKR1s , and acts as a general safety mechanism in STKR1 to prevent pathological leaky signaling . In summary , our study provides a compelling and unifying allosteric activation mechanism in STKR1 kinases that reconciles a large number of experimental studies and sheds light on a novel therapeutic avenue to target disease-related STKR1 mutants .
Serine/Threonine Kinase Receptors ( STKRs ) , also known as Transforming Growth Factor beta ( TGF-β ) receptors , are ubiquitous transmembrane proteins that play central roles in many biological processes ranging from cell differentiation , migration , proliferation and adhesion to development [1 , 2] . These receptors phosphorylate transcription factor Smad proteins in response to extracellular growth factors including TGF-β , activins , inhibins , nodal , and Bone Morphogenic Proteins ( BMPs ) . In mammals , twelve known STKRs are organized into either type 1 receptors ( STKR1 ) or type 2 receptors ( STKR2 ) . STKR1s encompasses seven members named Activin-Like Receptors 1–7 ( ALK1–7 ) , while STKR2s consist of five members: Activin Receptor Type 2A , Activin Receptor Type 2B , BMP Receptor Type 2 , TGF-β receptor Type 2 and Anti-Mullerian Hormone Receptor Type 2 . Extracellular ligands usually promote the formation of a heteromeric complex consisting of two STKR1s and two STKR2s , although in some cases the existence of a pre-formed complex has been reported [3] . This complex-ligand aggregate enables constitutively active STKR2 [4–6] to phosphorylate a glycine-serine ( GS ) regulatory region in the intracellular domain of STKR1 . In turn , STKR2-mediated phosphorylation of the STKR1 GS region enables STKR1 kinase domain to bind adenosine triphosphate ( ATP ) and transfer its γ-phosphate to downstream Smad substrates . Activated ALK2/3/6 phosphorylate downstream Smad1/5/8 to transduce BMP signaling whereas activated ALK4/5/7 phosphorylate downstream Smad2/3 to transduce TGF-β and activin signaling ( Fig 1A and 1B ) [7] . In the absence of ligand , STKR1 kinase activity is physiologically inhibited by binding of a negative regulator protein FKBP12 to the GS domain . Gain-of-function mutations in the GS and kinase domains are somehow able to bypass FKBP12-mediated inhibition and produce aberrant STKR1 signaling that is associated with various diseases such as heterotopic ossifications and cancer [8–11] . How STKR1 GS domain modifications by FKBP12 dissociation , STKR2 phosphorylation , or disease-causing mutations lead to activation of the kinase catalytic domain located about 30Å away is unclear . To this aim , structural studies have shown that in the inactive forms of ALK2 and ALK5 , ATP binding is prevented by an inhibitory salt-bridge between an aspartate residue located in the conserved DLG kinase motif and an arginine located in the activation loop ( A-loop ) [9 , 12] . However , it remains unknown whether this salt bridge represents a common inhibitory mechanism amongst all STKR1s , and if so , how the salt bridge is allosterically regulated by physiological and pathological modifications of the distant GS domain . In this study , using bioinformatics , molecular dynamics ( MD ) simulations and experimental assays , we show that this salt bridge acts as a common electrostatic lock among all inactive STKR1s . GS domain modifications allosterically destabilize this electrostatic interaction to induce STKR1 kinase activation . Experimental disruption of this electrostatic lock by mutations of the conserved A-loop arginine in three tested STKR1s produce disease-mimicking leaky signaling in the absence but not presence of the FKBP12 repressor , indicating this electrostatic lock functions as a secondary endogenous repressor downstream of FKBP12 .
To gain insight into the STKR1 activation mechanism , we aligned nine high-resolution X-ray structures of constitutively active STKR2s ( ActRIIB and BMPRII ) , FKBP12-bound inactive STKR1s and FKBP12-free STKR1s with substrate analogs bound to their catalytic sites . The overall structures of the kinase domains overlap extremely well except for the A-loop ( Fig 1C ) . In the inactive structures , an Arg ( R ) residue on the A-loop forms an inhibitory salt-bridge with the Asp ( D ) residue in the conserved DLG motif as previously reported [9 , 12] . In contrast , in constitutively active STKR2s and the FKBP12-free ALK5 structures , this A-loop R residue is systematically flipped away from the ATP binding site . Interestingly , a positively-charged residue Arg or Lys at this position is strictly conserved in all STKR1 but not present in any constitutively active STKR2s ( Fig 1D ) . These observations strongly suggest that the R-D salt bridge is ubiquitous in all known STKR1 members and constitutes an electrostatic lock to physiologically prevent STKR1 kinase catalytic function in the absence of ligand . If our hypothesis that the R-D salt bridge acts as an electrostatic lock in STKR1 kinase inhibition is correct , disrupting this electrostatic lock would be necessary during STKR1 kinase activation . To examine this concept , we first carried out multiple MD simulations in explicit solvent to study whether FKBP12 dissociation and GS loop phosphorylation ( Phosp ) disrupt the electrostatic lock in ALK2 . To this aim , four structure systems were prepared from the same crystal structure of inactive FKBP12-ALK2WT , namely FKBP12-ALK2WT; ALK2WT ( no FKBP12 ) ; FKBP12-ALK2WT-Phosp; and ALK2WT-Phosp ( no FKBP12 ) . Phosphorylation sites of ALK2WT-Phosp were selected based on a previous ALK5 study as they are conserved among STKR1 [7] ( see Methods ) . The length of each simulation varies between 280 to 300 nanoseconds ( ns ) . In either the absence of FKBP12 or presence of phosphorylation , the R-D lock between D354 and R375 rapidly breaks during the simulation , but remains only stable in FKBP12-ALK2WT simulations ( Fig 2A–2D ) . This unique stability of the R-D lock in FKBP12-ALK2WT was double-checked using a duplicated simulation of 280 ns starting from a different snapshot and initial velocity ( S1A Fig ) . Those simulation results suggest that activation of ALK2 kinase by either FKBP12 dissociation or GS phosphorylation indeed destabilizes the electrostatic lock in ALK2 . Interestingly in the simulation of ALK2WT ( no FKBP12 ) , the R ( Cζ ) -D ( Cγ ) distance varies between ~4 Å and ~8 Å , indicating the salt-bridge reversibly forms and breaks in the absence of FKBP12 ( Fig 2B ) . In contrast , in phosphorylated systems with or without FKBP12 , the R-D salt-bridge breaks at around 200 ns and oscillates above 8 Å , consistent with the requirement of physiological phosphorylation of GS domain for full activation of ALK2 ( Fig 2C and 2D ) . We also observed that GS domain phosphorylation results in partial FKBP12 dissociation ( S1B Fig ) , consistent with previous reports that FKBP12 only binds to unphosphorylated STKR1s [7 , 13 , 14] . The common ALK2R206H and ALK2Q207E gain-of-function mutations in the GS domain produce leaky BMP signaling and are associated with Fibrodysplasia Ossificans Progressiva ( FOP ) and childhood brainstem tumor diffuse intrinsic pontine glioma ( DIPG ) [15 , 16] . To investigate whether the ALK2 R-D lock is allosterically altered by these mutations , we mutated in silico the FKBP12-ALK2WT crystal structure to ALK2R206H and ALKQ207E and performed MD simulations . Because R206H is partially buried inside the protein structure at the FKBP12 binding interface , its pKa ( consequently its protonation state ) may fluctuate depending on local side chain rearrangements . We , therefore , calculated pKa of R206H using constant-pH molecular dynamics simulation [17] and PROPKA method ( using 10 , 000 conformations ) [18 , 19] . Both constant-pH simulation and PROPKA predict an average pKa value of around 6 . 3 , similar to the His pKa value in aqueous solvent ( S2 and S3 Figs ) . The dominant protonation state of R206H predicted by constant-pH simulation ( uncharged state with hydrogen on the Nδ atom ) is used for FKBP12- ALK2R206H MD simulation . It is very interesting that despite the fact that R206H and Q207E are located about 30 Å away from the R-D salt bridge , both mutations rapidly destabilize this salt bridge during the 300ns MD simulation ( Fig 2E and 2F ) . Unlike phosphorylation , R206H and Q207E did not seem to affect FKBP12 binding during the simulations ( S1 Fig ) , and the distribution of the R-D distance in the R206H mutant is similar to the ALK2WT simulation , suggesting that R206H alleviates the inhibitory effect of FKBP12 rather than disrupting its binding to ALK2 . This mechanistic hypothesis agrees well with experimental evidence that R206H does not dissociate ALK2/FKBP12 interaction [9] . In contrast , the R-D distance distribution of the Q207E mutant is similar to the phosphorylation systems , suggesting that the negative charge introduced by Q207E could partly mimic GS domain phosphorylation , potentially lead to partial FKBP12 dissociation in a longer time scale . This mechanism fits well with experimental data that FKBP12 removal does not significantly increase leaky BMP signaling through ALK2Q207E [9] . To overcome sampling limitations of the MD simulations , we calculated the free energy profiles or potential of mean force ( PMF ) along the R-D distance in the ALK2WT system using Hamiltonian Replica-Exchange Umbrella Sampling ( H-REUS ) molecular dynamics simulations [20–23] ( Fig 3 ) . The reaction coordinate of the umbrella sampling is the distance between center of mass of hydrogen atoms of one NH2 moiety of R375 guanidine sidechain and one oxygen atom of D354 carboxylic group . A correlation plot for R375-D354 salt bridge distance metrics vs . umbrella sampling restraint center reveals clearly that there is a sidechain flipping at 4 Å window and the R-D bridge breaks between 4 Å and 5 . 5 Å windows ( S7 Fig ) . Three main free energy wells were found at ~4 Å , ~8 Å , and ~13 Å , which are in excellent agreement with the R-D distance distribution from our brute-force MD simulations ( Fig 2A–2F ) . The first free energy barrier in PMF profile is between the energy minima at ~4 Å and ~8 Å , corresponding to the reversible disruption of the R-D salt-bridge , which is seen in the 300 ns equilibrium simulation of ALK2WT without FKBP12 ( Fig 2B ) and R206H mutant ( Fig 2F ) . A higher free energy barrier in PMF profile located at ~10 Å corresponds to R375 passing under the αC helix during flipping out . Only the brute-force simulations of phosphorylated systems ( Fig 2C and 2D ) and Q207E mutant ( Fig 2E ) sampled an R-D distance above 10 Å ( Fig 2 ) , indicating that phosphorylation and Q207E mutation significantly lower a 2nd free energy barrier corresponding to Arg flipping . To understand how phosphorylation at GS loop allosterically controls the R-D distance at catalytic site , we performed wavelet analysis feature extraction ( WAFEX ) analysis [24] by using the MD trajectories from four systems: FKBP12-ALK2WT , ALK2WT , FKBP12-ALK2WT-Phosp , and ALK2WT-Phosp ( Fig 4 ) . By examining the motion of kinase domain in both physical space ( indicated by residue ID on y-axis ) and temporal space ( indicated by simulation time on x-axis ) , we found that ALK2WT-Phosp system undergoes much larger scale motion than FKBP12 dissociation alone ( ALK2WT ) or GS loop phosphorylation with FKBP12 present ( FKBP12-ALK2WT-Phosp ) . In addition , WAFEX identified that the most intensive clustered motion ( shown in red color ) in ALK2WT-Phosp is located at αC helix and β3 sheet . The trajectory frames can be further split into three frame sets based on the discontinuity of the motions shown in the WAFEX diagram ( see Methods ) ( Fig 4 ) . The jump in the motion of αC helix around 200 ns correlates well with the jump in salt bridge distance in ALK2WT-Phosp ( Fig 2D ) . Based on this temporal key information , we performed principal component analysis ( PCA ) using frame set 2 to monitor the change in the motion . PCA revealed that ALK2WT-Phosp system exhibits significantly larger scale motion in GS domain , L45 loop , and αC helix , in comparison with ALK2WT-FKBP12 ( S1 Movie ) . Most interestingly , the direction of motion captured by PCA revealed a separation between αC helix and residue R375 in the R-D lock , as shown by the projections of the R-component vectors ( i . e . eigenvectors normalized by their eigenvalues ) onto the α carbons ( Fig 5 Top ) . This separation motion was quantified by calculating natural log of R375 to αC helix separation for first 15 principal components ( Fig 5 Bottom ) . In frame set 2 , the separation is significantly larger in phosphorylated system in the top three principal components , which constitutes 60–75% of the global motion ( see cumulative covariance in S4 Fig ) . Such separation represents an opening motion of the ATP binding site induced by phosphorylation of GS domain , which explains the larger R-D distances observed in the MD simulation of phosphorylated systems . To experimentally verify our MD simulation results , we sought to test the functional consequences of substituting the arginine in the R375-D354 lock of ALK2 to glutamine ( ALK2R375Q ) and glutamate ( ALK2R375E ) , two common residues found at the homolog position in STKR2 , which are constitutively active ( Fig 1D ) . In addition , since the FOP-related mutation ALK2R375P was previously characterized [9] , we have also included this mutation in the study as a control . The ALK2R375Q , ALK2R375E and ALK2R375P mutants as well as ALK2WT were expressed into BMP-responsive C2C12/BRE-luc cells respectively , followed by treatments with FKBP12 inhibitor FK506 . Compared to cells expressing ALK2WT , FK506 treatment alone significantly increased BMP-dependent luciferase activity ( RLU = 22 . 67±8 . 19 , n = 3 , p< 0 . 01 ) in cells expressing ALK2R375P , which is consistent with a previous report [9] ( Fig 6A ) . Interestingly , the two other mutations ALK2R375Q and ALK2R375E produced comparable luciferase activities to the ALK2R375P ( RLU = 23 . 67±3 . 18 , n = 3 , p< 0 . 001 for ALK2R375Q and RLU = 18±8 . 50 , n = 3 , p< 0 . 001 for ALK2R375E ) , in good agreement with R-D salt-bridge in ALK2 acting as an additional repressor beyond FKBP12 to prevent ligand-independent leaky signaling . In the ALK2WT transfected cells , a small , albeit insignificant , increase of luciferase activity with FK506 treatment is consistent with our simulation results that FKBP12 dissociation in ALKWT produces a less stable R-D salt-bridge but still maintains a relative high free energy barrier for Arg flipping , thus allowing the salt bridge to quickly reform following disruption . To examine whether an electrostatic lock exists in other BMP type I receptors , we substituted the corresponding arginine in the A-loop of ALK3 and ALK6 to proline as this mutation produces the strong phenotype amongst our tested ALK2 mutations . The ALK3R401P and ALK6R317P mutants were expressed into C2C12/BRE-luc cells respectively , followed by treatments with FK506 . Similarly to ALK2 mutants , FK506 treatments dramatically increased luciferase activities in cells transfected with ALK3R401P and ALK6R317P compared to ALK3WT and ALK6WT , respectively ( Fig 6B ) . These experiments further demonstrate that the presence of the Arg in the A-loop of ALKs promotes a common inhibitory mechanism independent of the FKBP12 repression . It is noteworthy that the luciferase activities induced by FKBP12 dissociation and Arg mutations are nearly not as high as the ones induced upon addition of BMP6 ligand . This is consistent with previous reports that the phosphorylation on GS domain may facilitate the binding of downstream Smads [7 , 25] . We noticed that in comparison to the ALK2WT basal luciferase activity ( RLU = 38 . 33±4 . 98 , n = 3 ) , basal activity ( i . e . in the absence of ligand and in the presence of FKBP12 ) of ALK2R375P ( RLU = 22 . 67±8 . 19 , n = 3 ) shows no significant difference , while the basal activity of ALK2R206H is significantly higher ( RLU = 192±14 . 50 , n = 3 ) ( Fig 6 ) . This is consistent with clinical observations that FOP patients bearing the ALK2R375P mutation generally have milder symptoms compared with patients with the ALK2R206H mutation . Interestingly , R206H is located on the GS domain while the R-D lock is located at the catalytic site 30 Å away . To investigate how ALK2R206H allosterically unlocks the R-D salt bridge , we performed network analysis to determine long-distance correlated residue motions between R206 and R375 in our simulations . Since the geodesic metric network analysis only searches sub-optimal paths within a specified cutoff and excludes contributions of contact edges in the entire network , we performed an alternative current-flow betweenness analysis , which provides scoring metric including contributions of all paths ( see Methods ) [26–29] . Comparison of ALK2WT versus ALK2R206H mutant networks is shown in Fig 7 . Fig 7A shows two-dimensional representations of the correlation and current-flow betweenness scores . In the upper plot , the full graph shows that the network of residue contacts involved in transmitting motion between 206 and 375 residues is more focused ( i . e . fewer pathways but with bright white or red color ) in ALK2WT than ALK2R206H . The high transmission sub-networks extracted from the upper panels is shown in the lower panels of Fig 7A . Interestingly , the contacts between FKBP12 and the kinase domain that are present in ALK2WT ( indicated by arrows ) disappeared in ALK2R206H . To elucidate the relevance of this missing path , the high transmission sub-paths were projected onto the three-dimensional representation of the protein structure . Fig 7B shows that the missing contacts in the mutant correspond to correlations between T86 of FKBP12 and F246 on ALK2 αC helix . This indicates that the R206H mutation allosterically disrupts FKBP12-mediated stabilization to the R-D lock . Also of note is the edge between P-loop Y219 and R375 show up in ALK2R206H , as hydrogen bonding between the Y219 and R375 can be directly observed after R-D lock breaks and R375 enters a metastable intermediate state . In conclusion , the network analysis indicates that the pathological effect of the R206H mutation is to bypass the inhibitory effect of FKBP12 rather than to prevent its binding to the GS domain . This mechanism is consistent with previous co-immunoprecipitation experiments showing that the R206H mutation did not disrupt physical binding of FKBP12 to the GS domain [9] . Consistent with the previous studies and current luciferase assay , FKBP12 dissociation in ALK2R206H produces a significant increase of luciferase activity as compared to the basal conditions , indicating that R206H does not entirely abolish FKBP12-mediated inhibition .
STKR1 are physiologically activated upon ligand-induced GS domain phosphorylation which promotes dissociation of the FKBP12 repressor , ATP binding to the catalytic site and Smad substrates recruitment . Here we show how the kinase catalytic site becomes active or inhibited in physiological or pathological situations . Modifications in the regulatory GS domain are propagated toward the catalytic domain where they stabilize or destabilize an inhibitory arginine-aspartate salt bridge positioned next to the ATP binding site . Interestingly , this salt bridge is strictly conserved in all STKR1s but is absent in STKR2s which are constitutively active , indicating the requirement of this electrostatic interaction for kinase inhibition . The salt bridge distance free energy profile reveals three possible states which correspond to the three functional states of the kinase receptors as illustrated in Fig 8 . In absence of ligand and in presence of the FKBP12 repressor , the Asp-Arg salt-bridge is formed and stable , leading to the inhibited form of the STKR1 . FKBP12 dissociation and pathogenic mutations in the GS domain lead to transient formation and disruption of the salt-bridge . This state is associated with a leaky or partially active state also called “de-inhibited” state . Finally , the Arg side chain flipping induces a kinase conformation where the salt bridge become permanently disrupted , allowing Asp to participate in ATP coordination . This constitutes the physiological active state of the kinase upon phosphorylation . It is important to notice that experimental removal of the salt bridge alone by mutating Arg did not lead to full kinase activation in a luciferase assay because downstream Smad substrates binding to STKR1 depends upon GS domain phosphorylation [7 , 25] . In addition , conformational differences between WT STKR1 and their pathogenic mutants ( such as FOP mutant ALK2R206H ) revealed in our study may offer a novel strategy to design small molecules specific to leaky mutant receptors without interrupting the WT STKR1 activity , thus minimizing potential drug side effects . In summary , our study provides a compelling and unifying general mechanism for STKR1 regulation , and sheds light on a novel therapeutic avenue for targeting disease-related STKR1 mutants . Five conformational changes are necessary to enable the fully active form of majority of kinases: A-loop opening , DFG-in motion , K-E pair formation , αC helix inward motion , and R-spine formation [30 , 31] . Interestingly , all those features already exist in the inactive forms of STKR1 ( i . e . FKBP12 bound and GS domain unphosphorylated ) , suggesting that the inactive STKR1 adopts a conformation similar to that of the fully active kinases ( Table 1 and Fig 9A and 9B ) . Unlike STKR1s , serine/threonine kinases PKA and Cdk or tyrosine kinases Src or Abl require A-loop phosphorylation for their full activation . Their DFG motifs located at the beginning of the A-loop are tightly coupled to the phosphorylation state of the A-loop . The phosphorylation-induced DFG-in conformation is required for the binding of divalent cations involved in nucleotide recognition , and in forming the regulatory hydrophobic spine ( R-spine ) in those kinases . In inactive STKR1s , a DLG motif , homolog to the DFG motif , adopts a DFG-in like conformation , but with one main difference that Asp residue is locked by the R-D salt bridge in inactive STKR1 conformation ( Fig 9C ) . Unlocking this salt bridge enables Asp of DLG motif to align perfectly with the DFG-in conformation of active tyrosine kinases in order to coordinate divalent cations . In summary , we observed that STKR1s adopt an active-like conformation , but are still physiologically inhibited by FKBP12 and the R-D salt bridge . The fact that STKR1 activation requires a much smaller conformational change than other kinases further emphasizes the physiological importance of the R-D lock in STKR1s .
All simulations were done using NAMD2 . 9b [32] or AMBER16 [33] with CHARMM C36 force field [34] . WAFEX analysis is done using CPPTRAJ [35] and R [36] . PCA is done using pytraj and visualized via the Normal Mode Wizard plugin of the VMD program [37 , 38] . Constant pH simulations were conducted using the constant pH functionality with Generalized Born implicit solvation under the AMBER16 . In network analysis , the flow betweenness scores and corresponding networks were computed using an R script and the NetworkView plugin in VMD [28] . The initial structure of wtALK2 was taken from the crystal structure of the ALK2-dorsomorphine in complex with FKBP12 ( PDB ID: 3H9R ) of Homo sapiens species . Using VMD , small molecular inhibitor was removed from the crystal structure to eliminate its inhibitory effect on ALK2 , and test the inhibitory effect of FKBP12 alone . In this crystal structure , ALK2 partial A-loop ( residues 362 to 374 ) , and the β-turn between β4 and β5 ( residues 273 to 275 ) were missing and therefore transplanted from PDB ID 3Q4U . The pKa values were calculated using PROPKA . CHARMM-GUI [39] was used to read in the PDB file and generate solvated systems . The phosphorylation site in ALK2 GS domain was selected based on experimental data of ALK5 , which defines the exact amino acids of the GS domain ( Thr185 , Ser187 , Ser189 and Ser191 ) that are phosphorylated during activation [7] . Each system was solvated in a rectangular water box with 12 Å edge distance from protein surface . Each system was neutralized with K+ and Cl- ions at a physiological salt concentration of 150 mM . All simulations employed the all-atom CHARMM C36 force field for proteins and ions , and the CHARMM TIP3P force field [40] for water . All brute-force simulations were performed with NAMD2 . 9b using periodic boundary conditions at constant temperature and pressure ( NPT ensemble ) of 300 K and 1 atm using Langevin thermostat and Andersen-Hoover barostat . Long-range electrostatic interactions were treated using the particle-mesh Ewald ( PME ) method . A smoothing function is applied to van der Waals forces between 10 Å and 12 Å . The dynamics were propagated using Langevin dynamics with langevin damping coefficient of 1 ps-1 and a time step of 2 fs . The non-bonded interaction list was updated on every integration step using a cutoff of 13 . 5 Å . The SHAKE algorithm was applied to all hydrogen atoms . The molecular dynamics equilibrium was set to relax the atomic system by releasing the harmonic constraints ( force constant 50 kcal/mol/Å2 ) stepwise ( every 200 ps ) on water and ion molecules , protein side chains , and eventually the protein backbone . To assess the protonation state of H206 in the R206H mutant and any consequential perturbation of the surrounding electrostatic environment incurred by this mutation , constant pH simulations were conducted using the constant pH ( CPH ) functionality of the GPU accelerated PMEMD program with Generalized Born implicit solvation [41 , 42] as available under the AMBER16 modeling and simulation suite . Prior to constant pH simulation , the PDB of the FKBP12-ALK2R206H mutant was preprocessed for use in CPH simulations under AMBER . The hydrogen atoms of the appropriate HIS residue were stripped and the residue renamed to HIP in order to denote it as a titratable HIS residue . The system was then processed for use under constant pH simulation as described in the AMBER16 constant pH tutorial and the corresponding paper by Mongan et . al . [17] . The system was minimized , then heated and equilibrated for a total of 6 ns , followed by 100 ns of production simulation . Protonation state data for H206 was then extraction from the CPH production simulation using the cphstats tool and imported into R for further analysis . The energetics of R-D lock dissociation was investigated using Hamiltonian replica exchange umbrella sampling ( H-REUS ) simulations . The last snapshot of ALK2WT system trajectories was used as a starting point for the HREUS-1D simulations . A set of 1 . 5 Å width windows was constructed using a harmonic restraint to enforce R-D separation distances . Prior to running H-REUS simulation , each window was allowed to equilibrate for a total of 10 ns at its specified separation distance . This was accomplished by iteratively increasing the restraint distance so that each window was equilibrated starting from the endpoint of the previous window’s equilibration trajectory . This was done to minimize distortions generated during rapid separation of the R-D lock . An exchange rate of 0 . 2 picoseconds ( ps ) was employed . System energies were output after each exchange event , and separation distances were recorded every 0 . 02 ps . Trajectory frames were written every 100 ps . The total simulation time is 30ns per window . All simulations were performed using the PMEMD module of the AMBER modeling and simulation package with support for MPI multi-process control and GPU acceleration code . All simulation parameters were implemented to match the standard MD simulations performed in NAMD described above , except the barostat method and VDW cutoff parameter . The standard AMBER cutoff 10 Å radius was used rather than the NAMD implementation which utilizes a smooth function between 8 and 12 Å , in order to avoid significant slowdown in GPU code algorithms . The AMBER simulations used a Monte-Carlo barostat rather than the standard langevin barostat employed for NAMD simulations . This again was necessary to ensure optimal performance of the GPU accelerated simulation code . The PMF was generated utilizing Weighted Histogram Analysis Method ( WHAM ) [43] . The PMF data output by WHAM for each sub-trajectory was collected and analyzed using R . Cumulative PMFs over every 2 ns were then plotted to ensure convergence of the PMF at the region of interest . The last 4 cumulative PMF are shown . Only data for the first 9 windows are displayed since the PMF is apparently still converging at much higher separations . The recent paper by Heidari et . al [24] describes a novel WAFEX analysis method for elucidation of the temporal and spacial location of relevant motions within a trajectory . Prior to performing WAFEX analysis , the trajectories from the NAMD based simulations were stripped of all solvent molecules and only the heavy atoms of the remaining protein were retained . Further , after a preliminary run , it was noted that torsional motion between the FKBP and kinase domain dominated the wavelet output and obscured other relevant motions of the kinase domain when the binding protein was present and included in the analysis . Thus , only the kinase domain was considered in this analysis , with the binding protein being stripped from the trajectory prior to conducting WAFEX . A total of 60 timescales ranging from 10 ps ( trajectory time step ) , up to 300 ns ( approximate trajectory duration ) were used . The Morlet wavelet was employed , necessitating a correction value of 1 . 01 as described in the paper by Heidri et . al [24] . The default chi squared cutoff of 1 . 6094 was used for the noise reduction threshold . A minimum cluster size of 350 points with an epsilon value of 20 . 0 was specified for the clustering algorithm . The WAFEX analysis along with preliminary preprocessing of trajectories was conducted using the modified CPPTRAJ program . The wavelet and clustering data produced by CPPTRAJ was then processed in R in order to merge the wavelet intensity data with the clusters predicted by the clustering analysis and produce corresponding clustered intensity plots . Silhouette scoring plot for temporal clustering of WAFEX data was generated by computing silhouette scores for consecutive k-cluster cuts taken from the hierarchical clustering of the frame-wise WAFEX data ( S5 Fig ) . Hierarchical clustering was computed using Euclidean distance norms over the log of the wavelet intensities at each frame along with Ward similarity scoring as implemented in the hclust function of the fastcluster package in R . The consistency of temporal clustering of WAFEX data by eye and by automated hierarchical clustering cut to 3 clusters is shown in S6 Fig . Since previous analyses indicate that the trajectories produced are likely exhibiting dynamic transition events , it was important to perform principal component analysis ( PCA ) over relevant subsets of each trajectory in order to ensure that the motions predicted by PCA correspond to the relevant conformation changes occurring in each individual event . For FKBP12-ALK2WT system , these sets were determined to be the times ranging from 0 to 120 ns , 120 to 220 ns and 220 to 250 ns . For ALK2WT-Phosp systems the time ranges were chosen to be 0 to 100 ns , 100 to 180 ns , and 180 to 250 ns . These sub-trajectories were loaded using pytraj ( a python implementation of CPPTRAJ and PTRAJ ) and the coordinates of backbone atoms were extracted by stripping all other atoms from the trajectories . As with wavelet analysis , the FKBP binding domain was stripped as well . Additionally , the first and last 16 residues of the kinase domain were stripped since these regions corresponded to unstructured tails , which exhibited whip-like motions that would dominate the principal components and obscure more relevant motions . The pca function of the pytraj library was then used to perform principal component analysis on each trajectory subset . The resulting eigenvalue and eigenvector data was then exported to data tables for further analysis in R . Additionally , the PCA data for each trajectory subset was output as NMD format files for visualization via the Normal Mode Wizard plugin of the VMD program . The extent to which each principal component contributed to separation between the R375 residue of the R-D lock and the portion of the αC helix acting as a steric barrier to Arg flipping was then quantified from the PCA data output by pytraj using R . To do so , the unit vector pointing from the center of mass of the steric barrier region of αC toward the center of mass of R375 was computed for the averaged structure generated over each trajectory subset . The average of the dot products of this vector with the R-component vectors [44] of each atom of αC was computed as well as the average of the dot products of the separation unit vector with the R375 atom R-component vectors . These two averages were then subtracted to yield the net separation motion described by each principal component . Finally , the Normal Mode Wizard plugin of the VMD visualization program was used to load the NMD format output files generated from the first 15 PCA modes computed using pytraj to produce renderings of representative principal component modes in the R375 / αC region to facilitate visual comparison of the relative motions of R375 and αC helix . Current-flow betweenness provides an alternative scoring metric [27 , 29] . This metric starts by constructing a ‘correlation resistance’ network by taking the reciprocal of the correlation for each pair of edge in the contact correlation network . The graph laplacian of this network is then constructed . This corresponds to the matrix constructed such that for any distinct pair of residues in the correlation resistance network , [i , j] , the corresponding off diagonal entry of the graph laplacian matrix ai , j is the inverse of the value of the edge in the resistance network . The inverse of this graph laplacian can then be used to compute a ranking for the relative importance of a given contact edge [i , j] in the correlation network , with respect to the transmission of motion between a given set of source residues S and target residues T . The following equation describes the procedure for computing the effective flow betweenness score for a given contact pair ( i , j ) given the appropriate pseudo inverse matrix: Ebtw ( i , j ) =∑s∈S∑t∈T|Li , s−1+Lj , t−1−Li , t−1−Lj , s−1| where L-1 is the pseudo inverse of the graph laplacian of the correlation resistance network and S and T are the sets of source and target nodes between which motion transfer is being considered . It should be noted that this equation assumes that sets of source and target nodes , S and T , are distinct . If not , the summation must be updated to include only distinct combinations of pairs . Since Laplacian matrices constructed from such graphs are , in general , not directly invertible , the Moore-Penrose pseudo-inverse is used as the effective inverse of the correlation resistance network’s graph laplacian . While the Bozzo and Franceschet paper [26] also discusses methods for approximating the pseudo inverse for large methods , direct computation was used here , since our network is relatively small ( under 500 nodes ) . In brief , this method of computing betweenness scores is equivalent to modeling the correlation network as if it were a network of electrical resistors , wherein each edge is represented as a resistor with resistances equal to the reciprocal of the absolute value of that edge’s correlation . The pseudo inverse of the graph laplacian for this network will then represent the relative difference in electrical potential between each pair of nodes . By selecting a set of nodes as the ‘sources’ and another set as the targets , or ‘grounds’ , it is then possible to predict the corresponding ‘relative flow’ across each edge as if 1 unit of ‘current’ ( e . g . atomic motion ) were sent from the source ( s ) to the target ( s ) . This ‘relative electrical flow’ can serve as a scoring function ( betweenness ) to rank each edge or node in order of its relative importance with respect to transmission of motion between the source and target nodes . These betweenness scores can then be used to isolate the most relevant sub-networks that transmit motion between either the FKBP binding interface or residue 206 , to residue R375 of the R-D saltbridge lock . The betweenness scores and corresponding optimal sub-networks were computed from the correlation and contact map matrices generated by the NetworkView plugin using R and the corresponding igraph [45] R library , and then visualized in two-dimensional format in R and three dimensional format in VMD . pCDNA3 plasmids harboring ALK2WT , ALK3WT and ALK6WT were purchased from Addgene ( #80870 , #80873 and #80882 ) . Point mutations in ALK2 , ALK3 and ALK6 were made in-house using the Q5-Site Directed Mutagenesis Kit from New England Biolabs . Mutagenic oligonucleotides were designed using the NEBaseChanger online tool ( New England Biolabs ) and purchased from Integrated DNA Technologies . Mutations were confirmed by automated Sanger sequencing ( Genewiz ) . Transient transfection was performed using Fugene HD transfection reagent ( Promega ) according to manufacturer’s instructions . Briefly , 7 . 5 × 103 C2C12/BRE-luc cells ( stable mouse C2C12 cells expressing the Id1 promoter-firefly luciferase reporter ) were seeded in 96-well plates to culture in the growth medium DMEM/10% FBS without antibiotics . After overnight culture , the cells were transfected with pcDNA3 . 1 ( + ) , human wild type ALK2 , ALK2R206H , ALK2R375P , ALK2R375Q , ALK2R375E , wild type ALK6 , ALK6R317P , wild type ALK3 and ALK3R401P respectively , along with the Renilla luciferase reporter pRL-TK plasmid . Five hours after transfection , cells were starved in DMEM containing 0 . 5% FBS for 4 h . Then the cells were untreated or treated with 1μM FK506 ( Cayman ) or 50 ng/ml BMP6 ( R&D ) overnight before lysis . Luciferase activities were determined according to the Dual-Luciferase® reporter assay system ( Promega ) using Renilla for normalization of transfection efficiency . Data are presented as the mean ± S . E . M ( standard error ) , each performed in triplicate . Student’s unpaired t-test was used for two groups of statistical analysis . A p-value <0 . 05 was considered statistically significant . The R scripts used to generate the current flow betweenness scores and sub-optimal networks can be found on GitHub at: https://github . com/LynaLuo-Lab/network_analysis_scripts .
|
Kinases play central role in essential physiological process and are attractive therapeutic drug targets . One of the important kinase families is Type 1 Serine/Threonine Kinase Receptors ( STKR1 ) , which control gene expression in response to extracellular growth factors . The activities of STKR1 are tightly controlled by their regulatory domain , which is distant from the kinase catalytic site . The underlying molecular mechanism is elucidated here . We identified that formation or disruption of a highly conserved charge-charge interaction located near the ATP binding site , mediates the physiological inhibition or activation of STKR1 . We find that the stability of this charge-charge interaction is remotely controlled by interactions propagated from the distant regulatory domain . Several disease-causing mutations are located at the regulatory domain . We demonstrate how those mutations bypass these endogenous STKR1 inhibition mechanisms to produce pathological phenotypes . This study provides a general activation mechanism in STKR1 kinases , thus may benefit understanding the molecular mechanism of diseases and drug development .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"phosphorylation",
"luciferase",
"crystal",
"structure",
"molecular",
"dynamics",
"enzymes",
"electricity",
"condensed",
"matter",
"physics",
"enzymology",
"multivariate",
"analysis",
"electrostatics",
"mathematics",
"statistics",
"(mathematics)",
"crystallography",
"thermodynamics",
"research",
"and",
"analysis",
"methods",
"solid",
"state",
"physics",
"proteins",
"mathematical",
"and",
"statistical",
"techniques",
"principal",
"component",
"analysis",
"oxidoreductases",
"salt",
"bridges",
"chemistry",
"free",
"energy",
"physics",
"biochemistry",
"electrochemistry",
"post-translational",
"modification",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"chemistry",
"statistical",
"methods"
] |
2017
|
Polymodal allosteric regulation of Type 1 Serine/Threonine Kinase Receptors via a conserved electrostatic lock
|
The organization of the Escherichia coli chromosome into a ring composed of four macrodomains and two less-structured regions influences the segregation of sister chromatids and the mobility of chromosomal DNA . The structuring of the terminus region ( Ter ) into a macrodomain relies on the interaction of the protein MatP with a 13-bp target called matS repeated 23 times in the 800-kb-long domain . Here , by using a new method that allows the transposition of any chromosomal segment at a defined position on the genetic map , we reveal a site-specific system that restricts to the Ter region a constraining process that reduces DNA mobility and delays loci segregation . Remarkably , the constraining process is regulated during the cell cycle and occurs only when the Ter MD is associated with the division machinery at mid-cell . The change of DNA properties does not rely on the presence of a trans-acting mechanism but rather involves a cis-effect acting at a long distance from the Ter region . Two specific 12-bp sequences located in the flanking Left and Right macrodomains and a newly identified protein designated YfbV conserved with MatP through evolution are required to impede the spreading of the constraining process to the rest of the chromosome . Our results unravel a site-specific system required to restrict to the Ter region the consequences of anchoring the Ter MD to the division machinery .
The large size of genomes compared to cell dimensions imposes an extensive compaction of chromosomes compatible with various processes of DNA metabolism such as gene expression , replication and segregation of the genetic information . The organized bacterial chromosome , named the nucleoid , is a compact structure that presents a particular orientation inside the cell preserving the linear order of genes on DNA [1] . Although our understanding of the global organization of the bacterial chromosome is still limited , different levels of organization have been identified . At the molecular level , binding of a number of proteins including nucleoid-associated proteins ( NAPs ) , Structural Maintenance of Chromosomes ( SMC ) proteins , transcriptional regulators , chromosome organizers or RNA polymerase to the DNA molecule results in the formation of a bacterial chromatin with different properties according to the identity of proteins bound [2]–[4] . The excess of negative supercoils generates the formation of plectonemes that impose a secondary structure to the DNA molecule [5] . The size of such structures has been estimated to range in size from 10 kb to 100 kb depending on the studies [6]–[8] . At a higher level , there are several levels of organization required to fully condense the E . coli chromosome . Recent studies have implicated the H-NS protein in the formation of two clusters per chromosome in which H-NS regulated genes are sequestered [9] . Using FISH and genetic approaches , a long range organization based on the existence of four insulated macrodomains ( MD ) and two less constrained regions called non-structured ( NS ) regions has been uncovered [10] , [11] . MDs have been defined as large regions in which DNA interactions occurred preferentially and DNA interactions between the different MDs are highly restricted . In NS regions , DNA sites can interact with both flanking MDs [11] . The Ori MD contains oriC while the opposite Ter MD contains the replication terminus and the chromosome dimer resolution dif site . The Ter MD is flanked by the Left and Right MDs whereas the Ori MD is flanked by the two NS ( NSRight and NSLeft ) regions [11] . This organization influences the segregation of sister chromatids and the mobility of chromosomal DNA [12] . Structuring of the Ter MD relies on the binding of the MatP protein to a 13 bp motif called matS repeated 23 times in the 800 kb long domain . This protein accumulates in the cell as a discrete focus that colocalizes with the Ter MD . In the absence of MatP , DNA is less compacted , the mobility of markers is increased and segregation of the Ter MD occurs early in the cell cycle [13] . Moreover , recent studies showed that these two last aspects are also affected in a zapB mutant . ZapB , which is associated to the division machinery , stabilizes the Ter MD at mid-cell through a direct interaction with MatP ( Espéli et al . , submitted ) . Interestingly , bioinformatic analyses showed that matS sequences are present in different enterobacteria and Vibrio species [13] while MatP has been shown to belong to a group of proteins ( including SeqA and MukBEF ) exclusively identified in bacteria carrying Dam methyltransferase activity [14] . The molecular basis for structuring the other MDs remains to be characterized . However , repeated sequences analogous to matS have not been found in the other MDs and other models should be considered . In order to investigate the effect of the relative position of the MD on its properties , we designed a new method reminiscent of a cut-and-paste transposition event that allows the relocation of any chromosomal fragment at a defined position of the genetic map . Using this “transposition” technique , we generated a number of new chromosomal configurations . Very interestingly , we noticed that transposing the NS regions close to the Ter MD leads to a dramatic decrease in the mobility of fluorescent markers . We further identified two palindromic sequences called tidR and tidL ( found in the Right and Left MDs , respectively ) that are required in order for the NS regions to keep their specific properties . In the absence of these sequences , a constraining effect dependent upon the association of the Ter MD to the division machinery and acting in cis is detected at a very large distance , up to 750 kb of the Ter MD . Strikingly , we identified a protein , called YfbV , also exclusively found in bacteria possessing the Dam methylase which is required to prevent this constraining process of the NS regions and flanking Right and Left MDs . Overall , this study shows that specific determinants ( two palindromic sequences and a protein ) are required in order to prevent a MatP dependent constraining effect to affect chromosome arms flanking the Ter MD , therefore “insulating” the Ter MD from other parts of the chromosome .
In several studies , genomic rearrangements have been used to unveil the principles of chromosome organization [15]–[18] . In these cases , recombination between inverted recombining sites was used to promote genomic rearrangement . Unfortunately , this method perturbs several parameters at the same time: for example , inversion of the segment encompassing the Ori and Right MDs changes the relative position of the involved regions in Ori and Right MDs but also affects the orientation of more than one thousand genes including those in the intervening NS region . To limit the number of parameters affected by genetic rearrangements , we devised a new strategy reminiscent of a cut-and-paste transposition event that allows relocating any chromosomal fragment at a defined position of the genetic map without changing gene orientation . It relies upon the site-specific recombination Int system from bacteriophage λ . The Int integrase mediates phage integration in the chromosome by recombination between phage attP and bacterial attB attachment sites ( integrative recombination ) generating two hybrid sites , attL and attR . Recombination between attL and attR ( excisive recombination ) requires the additional presence of the Xis excisionase . We used directly repeated attL and attR sites flanked by the 5′and 3′ parts of lacZ , respectively [11] , to delimitate the genomic fragments to relocate . Recombination between these two sites leads to the formation of two circles , one carrying attB in frame within a functional lacZ gene , and one carrying attP . An additional attB site ( called attB′ ) was integrated at the target site . The int and xis genes were cloned under the control of a thermosensitive promoter [11] . The transposition takes place in two steps ( Figure 1 ) . First , excisive recombination between attL and attR sites promoted by Int and Xis generates two circles: one carrying attP ( the excised segment ) and one carrying attB ( the chromosome with a deletion of the excised segment ) . As the excised segment carries essential genes , its loss is fatal to the cell . Therefore , the second step of the transposition process consists in its reinsertion into the chromosome , either in the reconstituted attB fused to lacZ or in attB′ ( giving rise to attL′ and attR′ sites ) ( Figure 1 ) . The two events can be discriminated by the absence or presence of β-galactosidase activity . Conditions that provide a large amount of Xis and Int recombinases were applied to ensure that the excision of the segment between attR and attL occurs at a high frequency , greater than 95% . The level of lethality was greater than 90% , indicating that reinsertion of the excised segment occurred at low frequency . This is probably due to the presence of Xis that inhibits integrative recombination . However the amount of recombination was sufficient to obtain every transposition we tried . Insertion in the attB′ site and formation of the attL′ and attR′ sites can be detected by the appearance of blue clones on plates containing X-Gal and can be demonstrated at the DNA level by PCR analysis ( Figure S1 ) . Using fluorescent microscopy , we showed in a previous study that the dynamic behaviour of loci was different according to their belonging to a MD or a NS region . The system used derived from the bacteriophage P1 partition module and involved the ParB-GFP fusion protein interacting with parS sites inserted in the chromosome [19] . By time-lapse imaging during a 5 min period with a 10 sec interval , tracking of parS tags inserted in different regions of the chromosome and decorated by a fluorescent ParB-GFP fusion protein allowed to measure the travelled distance by each tag and its diffusion coefficient; it revealed that constraints on mobility are higher in MDs than in NS regions [12] . In order to investigate the effect of position on these different behaviours , travelled distances ( Figure 2 ) and diffusion coefficients ( Table S1 ) were measured in different strains ( Table 1 ) for a number of loci ( NSR-1 , NSR-2 , NSR-5 , Right-2 , Right-5 , Ter-3 , Left-2 , Left-1 , NSL-3 , NSL-4 , Ori-3; for their exact location , see Table S2 ) in different chromosomal configurations . Results are shown in Figure 2 . We first transposed most of the NSRight region between the Ter and Right MDs ( strain LC13-R127-BO-NSR ) . Such rearrangement had no dramatic effect on cell growth , cell and nucleoid aspects; a mild effect was detected in competition experiments , resulting in a 1 to 10 ratio in competition assays after 80 generations ( Figure S2 ) . Remarkably , mobility was specifically affected in the transposed NSRight region ( markers NSR-2 and NSR-5 ) whereas it was unchanged in other regions of the chromosome , including the NSLeft region ( Figure 2B ) . We next transposed most of the NSLeft region between the Ter and Left MDs ( strain LR146-R124 BL-T2 ) . Strikingly , the marker NSL-3 in the NSLeft region specifically showed a decreased mobility , whereas markers in other regions ( including NSR-2 in the NSRight region ) were not affected ( Figure 2C ) . We also constructed a strain in which the two NS regions are positioned on the same replication arm of the chromosome ( strain LC13-R18inv/LR146-R31 in Table 1 ) . Strikingly , markers located in the two NS regions showed a decreased mobility ( Figure 2D ) . This indicates that the constraining of the NS regions is not linked to their position relative to the Right or the Left MDs . Rather it may depend on proximity of the Ter MD . Structuring of the Ter MD relies on the binding of the MatP protein to its target sequence matS . To determine whether the constraining process which originates from the Ter region requires its structuring into a MD , we analysed the changes in mobility of markers in cells devoid of MatP . We deleted the matP gene in the LC13-R127-BO-NSR strain , and assayed the mobility of markers located in different regions of the chromosome ( Figure 2E ) . Remarkably , markers located in the NSRight region next to the Ter were mobile , in contrast to what is observed when MatP was present ( see Figure 2B for comparison ) . The mobility of a Ter marker was also increased in a matP mutant , as expected , whereas the mobility of markers located in the Right and Left MD was not affected ( Figure 2E ) . These results indicated that the constraining process associated to the Ter region required the MatP protein . Experiments described above suggest that the Right and Left MDs may insulate the NS regions from a constraint linked to the Ter MD proximity . In order to test this hypothesis and investigate what regions of the Right MD was required to protect the NSRight region , we moved various parts of the Right MD by transposition and tested the effect of remaining parts of the Right MD on the mobility of markers in the NSRight region . Results are presented in Figure 3A . They showed that this mobility was not affected when the 221 kb region closest to the Ter MD ( coordinates 914 kb to 1135 kb ) remained between the Ter MD and the NS region ( strain R127-LR14-BO-NSR ) whereas the 206 kb region nearest from the NSRight region ( coordinates 600 kb to 806 kb ) was not sufficient to prevent a decrease in mobility of the NS region ( strain LC13-R17-BO-NSR ) . In this strain , the mobility of a marker located in this 206 kb segment of the Right MD was the same as in a wild type chromosomal configuration ( data not shown ) . This indicated that the structuring of the Right MD by itself was not affected , and that it was not sufficient to protect the NS region . This also revealed that specific insulation determinants were present in the 221 kb segment adjacent to the Ter MD . We further dissected what part of this 221 kb segment carried these determinants , using the same strategy . We showed that they were present in the 142 kb segment flanking the Ter MD ( coordinates 993 kb from 1135 kb; strain R127-LR132-BO-NSR ) but not in the 133 kb segment flanking the Ter MD ( coordinates 1002 kb to 1135 kb; strain R127-LR128-BO-NSR in Figure 3A ) . This demonstrated that insulation determinants were present in the 9 kb region located between coordinate 993 kb to 1002 kb , which encompasses 10 genes ( ssuCDAE-ycbQRSTU ) predicted as not essential for the viability of E . coli . Deletion of this 9 kb region was engineered ( strain Δrins1 ) and we measured the mobility of different markers . A specific reduction of mobility was observed for the marker NSR-2 in the NSRight region , confirming that insulation determinants are present in that 9 kb region . Interestingly , these results showed that the constraining effect promoted by the Ter MD was also detectable in the native chromosome configuration provided that the insulating determinants were removed indicating that immediate proximity of the Ter MD was not required . Deletions spanning the ssuCDAE-ycbQRSTU region were generated to map more precisely the insulation determinants and the effect of these deletions on the mobility of markers located in the NSRight region was assayed . The first series of deletions indicated that determinants were present in the ssuCD region ( Figure 3B ) . More precise deletions were subsequently generated and the results showed that only a 12 bp palindromic sequence ( GCTGACGTCAGC ) located in the 3′ end region of ssuC was required for the insulation phenomenon ( Figure 3C ) . Indeed , deletion of these 12 bp induced a specific decrease in mobility of markers NSR-2 and NSR-5 , whereas other markers ( such as NSL-4 in the NSLeft region ) were not affected ( strain ΔtidR in Figure 3D ) . This sequence was called tidR for ter MD insulation determinant Right . Its deletion had no important consequences for cell morphology and nucleoid distribution ( Figure S3 ) . As described above , the constraining process seemed to be at work on both sides of the Ter MD . In order to map insulation determinants responsible for preventing a change of properties in the NSLeft region , a number of chromosomal rearrangements involving the Left MD and the NSLeft region were generated ( strains LR146-R124-BL-T and LC13-R524 ) . Results showed that insulation determinants were present in the 282 kb segment flanking the Ter MD ( strains LR146-R124-BL-T in Figure 3E ) and presumably present in the 135 kb segment flanking the Ter MD ( strain LC13-R524 in Figure 3F and data not shown ) . A bioinformatic analysis of this 135 kb fragment allowed the identification of a 12 bp sequence that differed from tidR by a single nucleotide ( GTTGACGTCAGC ) that was called tidL for ter MD insulation determinant Left . This sequence was located in the tar gene , encoding a methyl-accepting chemotaxis protein . The effect of tidL deletion on marker mobility could not be assessed as disruption of the tar-tsp-cheRBYZ operon perturbed the number and disposition of fluorescent tags; tidL insulation properties were demonstrated by inserting it at an ectopic position ( see below ) . To determine whether tidR/L could act at another position , the 500 bp fragment carrying tidR deleted in strain ΔRins2 . 1 was inserted in the middle of the NSRight region , between markers NSR-2 and NSR-5 , in a Δrins2 strain ( Figure 4A ) . The NSR-5 marker which is closest to the Ter MD showed a decrease in mobility compared to the wild type situation , similar to what was observed in a ΔtidR strain . Remarkably , the NSR-2 marker , farthest from the Ter MD , displayed a mobility similar to the wild type situation , in contrast to what was observed in a ΔtidR strain . These results indicated that the fragment containing tidR protected from the constraining effect originating from the Ter MD at various positions , in the Right MD or in the NSRight region . They suggest that tidR works in cis by impeding the progression along the chromosome of a process that changes the dynamic properties of the DNA molecule . To test whether the 12 bp tidR/L sequence was sufficient to protect from the constraining process , tidL was inserted at the same location between markers NSR-2 and NSR-5 ( Figure 4B ) . Insertion of the tidL sequence protected the marker NSR-2 from the constraining effect indicating that the 12 bp sequence was sufficient to impede the constraining process ( Figure 4B ) ; similar results were observed with the 12 bp tidR sequence ( data not shown ) . The effect was specific for tidR and tidL as a 12 bp control shuffled palindromic sequence GACGCTAGCGTC had no effect on impeding the change of mobility of NSR-2 ( Figure 4C ) . These results allowed the definition of the 12-mer GYTGACGTCAGC consensus sequence that insulated the NS regions from a long range cis-acting constraining effect promoted by the Ter MD . tidR and tidL were the only occurrences of this sequence found in the E . coli genome . By bioinformatic analyses , the presence of tidRL flanking the Ter MD was found conserved only in Shigella species; other enterobacterial genomes lack tidRL sequences ( e . g . Yersinia pestis ) or carry multiple copies ( e . g . in Salmonella typhimurium ) . MatP belongs to a group of 18 proteins exclusively identified in bacteria with Dam methyltransferase activity [14] . We wondered whether one of these proteins might be involved in the insulation mechanism preventing the spreading in cis of the constraining process . In order to test this hypothesis , we analyzed the mobility of markers in mutants of each of the genes encoding proteins of unknown function . Strikingly , the mobility of markers located in the NSRight region and in the NSleft region was reduced in a yfbV mutant , whereas mobility of markers located in the different MDs was not affected ( Figure 5A , see Figure 2A for comparison ) . This protein might thus be implicated in the insulation process promoted by both tidR and tidL . Interestingly , the yfbV gene encodes a predicted protein of 151 amino acids with two internal trans-membrane segments imposing the following topology: a 45 aa N-terminus cytoplasmic domain , a trans-membrane domain composed of two trans-membrane segments separated by a small periplasmic loop of 3 aa , and a 60 aa C-terminal cytoplasmic domain . The reduction in mobility of markers located in the NSRight region and in the NSleft region observed in a yfbV mutant was not seen in a yfbV matP double mutant ( Figure 5B ) , confirming the involvement of yfbV in impeding the spreading of the constraining process originating from the Ter MD . It was recently shown that the Ter MD is associated with components of the divisome during part of the cell cycle . This association relies on the interaction of MatP with ZapB ( Espeli et al . , submitted ) . In order to probe the role of this association in the constraining process , we analysed the mobility of markers in cells devoid of ZapB ( Figure 6A ) . In a zapB mutant , the mobility of a Ter marker was increased , whereas mobility of markers located in the other MDs or in the NS regions remained the same ( See Figure 2A for comparison ) . Remarkably , deleting the tidR palindrome in this zapB mutant context did not lead to a decreased in mobility of makers located in the NSRight region , in contrast to what was observed in a wild type background ( Figure 6B , see Figure 3D for comparison ) . This suggests that the constraining process affecting the NS regions in the absence of tidR and tidL may depend upon the association of the Ter MD with the divisome . This interaction lasts for about half of the cell cycle , and mobility of markers was routinely measured during that period . In order to confirm that the constraining process really required this association , we measured the mobility of markers during the part of the cell cycle when the Ter MD did not interact with the divisome , i . e . in small cells in which the Ter MD has not yet been segregated ( Figure 6C ) . In these cells , the mobility of a Ter marker was slightly increased , which was consistent with the fact that association with the divisome added a level of constraint onto the Ter MD ( Espéli et al . , submitted ) . Strikingly , a marker in the NSRight region was as mobile in small ΔtidR cells as in small wild type cells ( Figure 6D ) . Altogether , these results demonstrated that the constraining process observed in the NSRight region when tidR is absent depends on the association of the Ter MD with the division machinery . It was previously shown that MatP associates with the entire 800 kb Ter region ( Mercier et al . , 2008 ) . We wondered whether the constraining process originating from the Ter MD required the recruitment of the constrained NS regions within the Ter MD territory . To test this hypothesis , we analysed the co-localization of different markers with a MatP-mCherry fusion protein ( Figure 6E ) . In the wild type chromosomal configuration , the Ter-2 marker was co-localized with the MatP focus , whereas the Right-2 marker was not ( Figure 6E ) . In a strain where the NSRight region has been transposed next to the Ter region ( strain LC13-R127-BO-NSR ) , the NSR-2 marker did not co-localize with the MatP focus as observed for marker Right-2 in the wild type configuration ( Figure 6E ) . These results indicated that constraining of the NS region did not result from a recruitment of the constrained region in the Ter MD territory . Long distance DNA collisions revealed by the frequency of λ site-specific recombination between λ attR and attL sites inserted at different chromosomal locations can be used to reveal chromosome conformation [11] . Inactivation of MatP leads to increased collisions between markers in the Right MD and the Ter MD [13] . To assess whether the constraining of the NSRight region upon deletion of tidR could affect long range DNA collisions inside the NSRight region or between the NSRight region and the flanking Right MD , interactions between attL inserted in the NSRight region and attR sites located at different positions in wt or Δrins2 cells . Remarkably , interactions were as frequent in both genetic contexts , either within the NSRight region or between the NSRight region and the Right MD ( Table S3 ) . The organization of the chromosome into macrodomains influences the segregation of sister chromatids and the mobility of chromosomal DNA [12] . Above is described the effect of the absence of the insulation determinants upon DNA mobility . We then examined its impact on the extent of co-localization following replication , by measuring the co-localization index ( see Materials and Methods ) . In the wild type strain , this co-localisation index was high for markers located in MDs ( >0 . 1 ) in contrast to markers in NS regions ( <0 . 02 ) . Remarkably , only the index of co-localization of the NSRight region marker increased in a ΔtidR mutant , whereas it increased for markers present in both the NSRight and NSLeft regions in the yfbV mutant ( Table 2 ) . The increase of co-localization index for markers in NS regions was correlated with a higher number of cells carrying only one focus , localized close to mid-cell .
In various studies , inversion of chromosomal segments has been used to change gene positioning or chromosome configurations . Although these approaches have been very useful to reveal various features of gene and genome organization , genetic inversions affect several parameters: gene orientation relative to replication , gene dosage through the change of position relative to oriC , disruption of genetic organization at the two sites of recombination . We have devised a new method to rearrange the bacterial chromosome . The experimental design involved two steps , excision of a segment followed by the reinsertion at another location , and the desired chromosomal configuration can be directly detected on plates with a coloured indicator . This method allows the transposition of fragments ranging in size from a few kilobases to hundreds of kilobases . Because the orientation of genes and sequences are conserved , it allows the transposition of any segment on the chromosome; we succeeded in making all the rearrangements we wanted to perform . The only limitation of the method is the requirement for an essential gene in the transposed segment . In the absence of such a gene , all clones that sustained the deletion will grow making difficult the detection of clones that recombined the excised molecule . Markers in the MD and NS regions can be distinguished by their mobility and by the extent of colocalization after replication . The juxtaposition of NS regions to the Ter MD , the deletion of the insulators or the inactivation of YfbV provoked a change of these two properties in the NS regions . Remarkably , the constraining process appeared to spread in cis as insertions of the insulators in the middle of the NSRight region insulated the distal part but not the part proximal to the Ter MD . This effect was regulated during the cell cycle as it was promoted by the interaction of the Ter MD with the division machinery at mid-cell . The spreading of the constraining process can not be visualized in Right and Left MDs of wt cells because of the reduced mobility of markers in these regions . However , in yihI mutant cells in which the mobility of markers in the Right and Left MDs is released to a level similar to that of NS regions ( Valens et al . , in prep ) , mobility of markers of the Right MD located between the Ter MD and tidR were constrained by the MatP-associated constraining process ( data not shown ) . These results indicate that , in the absence of insulation barrier tidR , the constraining process affects not only the NSRight region but also the Right MD in agreement with a process spreading in cis . It is not yet known how MatP organizes the Ter MD at the molecular level . The association of the Ter MD with the division machinery modifies its properties and probably affects the overall structure of the Ter region; targeting of the Ter MD to the division machinery could promote the formation of loops between matS sites via the interaction of MatP with ZapB assembled in the FtsZ ring ( Espéli et al . , submitted ) . In the absence of the insulator , the macromolecular complexes assembled at mid-cell provoked a change of DNA properties detectable at several hundreds of kilobases ( Figure 7 ) . The change of properties at a long distance might either result from a change at the chromatin level by the binding/tracking of protein ( s ) or from modifications of physical properties of the DNA molecule . We do not favour the first hypothesis as it is hard to conceive how the effect would be related to the targeting of the Ter region to the division machinery and how binding or tracking of a protein could act in cis over a region of several hundreds of kilobases . A plausible hypothesis implies modifications of physical properties of DNA . Using two different topological reporter systems that reveal changes in local DNA supercoiling ( expression of a lacZ gene under the control of the supercoiling sensitive promoter PgyrA and site-specific resolution reaction between two γδ res sites catalyzed by the γδ resolvase [6] , [20] ) , we failed to detect modifications in the topological properties of the DNA molecule in the absence of insulation ( data not shown ) . We must therefore consider a hypothesis involving an affect on other mechanical properties of the DNA molecule . Although the molecular mechanism is not yet identified , we can imagine that it requires two “anchorages” between which DNA is constrained . While the association of the Ter MD to the division apparatus could constitute the first anchorage , the second one would implicate the tethering of other chromosomal regions in the cell . Because the process is effective up to the NS regions , the second anchorage might involve the Ori region . By acting as a mechanical relief , tidRL would restrain the constraining process and would restrict it to the Ter region . A number of processes are known to occur in the terminus of the chromosome; they include replication termination [21] , decatenation of entangled chromosomes [22] , chromosome dimer resolution [23] and coupling of replication termination with cell cycle progression [15] . It is tempting to speculate that the changes in DNA properties resulting from the association with the division machinery of the Ter region facilitate certain DNA metabolic processes required for the proper segregation of the chromosome . A challenge now is to characterize the molecular architecture of the Ter MD associated to the division machinery , to identify the process ( es ) occurring at this stage and reveal the contribution of the change of DNA properties for the efficiency of the process ( es ) involved . By genetic [11] and cytological [12] approaches , we have revealed a chromosome organization into structured MDs and less structured NS regions . Our subsequent analyses have revealed the molecular nature of the Ter MD structuring [13] . Here we have identified a new system that is related to the process involved in the structuring of the Ter MD . This system allowed restricting the effects associated to the interaction of the Ter region with the division machinery . Two 12 bp motifs that flank the Ter MD and one protein co-occurent with MatP have been shown to be required for the insulation process . The two sequences tidR and tidL were found in the MD flanking the Ter MD , at 130 kb on the right side and at 50 kb on the left side , respectively . This positioning probably reflects the potential of MatP to act at a distance greater than 50 kb [13] that likely precludes the positioning of the insulators closer to the matS sites located at the borders of the Ter MD . Remarkably , tidR and tidL conserved their activity when moved several hundred kilobases away from their original position; their presence close to the Ter MD might thus not rely on specific positional requirements for the insulation activity but rather serve to prevent the spreading of the constraining process to the flanking regions . tidR and tidL are the only occurrences of the 12-mer GYTGACGTCAGC consensus sequence in the E . coli genome . Allowing one nucleotide difference in the tidRL consensus sequence revealed the presence of 34 sites in the entire genome , dispersed in the diverse MD and NS regions ( Figure S4 ) . Based on their position on the genetic map and the properties of markers in various strains , 28 were predicted to be non functional for the insulation property . For 6 variants ( 2 in the Ter MD and 4 in the Ori MD ) , it was not possible to predict the functionality . At least two types of insulation mechanisms could be envisioned at this stage . In both cases , the insulator would block the unidirectional progression of the constraining process since it acts in cis . In the first type , a nucleoprotein complex formed at the insulator would act as a barrier and block the progression of a factor tracking along the DNA molecule . This process would be reminiscent of that selected in boundary elements that prevent the spreading of heterochromatin via the binding of proteins to specific sequences in yeast or vertebrate cells [24] , [25] or of that at work with the Tus/ter system that block replication fork progression in bacteria [26] . The second type may result in barrier function by interfering with the spreading of changes affecting physical properties resulting from the MatP-matS complexes . In the last case , anchoring of the DNA molecule to a cellular structure may be required to act as a point of physical relief . We showed that the gene yfbV is required for the insulation properties of tidR and tidL . It is not yet known how YfbV could promote insulation at tidRL sequences; however it is tempting to speculate that the anchorage of YfbV in the membrane might be used to tether the insulators to a fixed structure , e . g . the membrane , within the cell . Further analyses will be required to characterize the role played by YfbV in the insulation process .
The bacterial strains and plasmids used in this study are listed in Table 1 . E . coli strains were grown at 30°C in Lennox broth , or in minimal medium A supplemented with 0 . 12% of casaminoacids and 0 . 5% of glucose . Antibiotics were added when necessary . The different deletions targeted in the chromosome were performed by the one-step insertion-deletion technique [27] . These deletions were done in strains carrying the plasmid pKD46 or in the strain DY330 [28] . Two cassettes used to perform deletion constructions carried either a chloramphenicol or a rifampicin resistance gene . Both resistance genes are flanked by frt sites allowing their subsequent deletion [27] . Deletion coordinates are indicated in Table 1 . The deletions were verified by PCR . The strains used for transposition carried three attL , attR and attB′ sites derived from the λ site-specific integration module . The three att sites were inserted in the same orientation . Each site was flanked by a cassette carrying an antibiotic resistance gene ( chloramphenicol , kanamycin and apramycin , respectively ) . In addition , the attL site was flanked by the 5′ part of lacZ and the attR site was flanked by the 3′ part of lacZ . The strains are transformed by the plasmid pTSA-CXI that expressed the λ recombinase [11] to promote the transposition . The strains were grown at 30°C and shifted to 39°C to induce the production of the recombinases . After the recombination step , the strains were plated on LB medium supplemented with Xgal and blue colonies were selected . The transposition reactions were verified by PCR verification with appropriate primers ( Figure S1 ) . Cultures were grown in minimal A medium in the presence of glucose and casaminoacids without IPTG to maintain at a minimal level expression of gfp-parB present on plasmid pALA2705 [19] . Movies were recorded automatically on a Leica microscope . Autofocus was performed at every time point on the phase contrast image and GFP fluorescence was recorded on the plane with the best phase contrast focus . Image analysis was performed with ImageJ software using the manual tracking plugin ( http://rsb . info . nih . gov/ij/index . html ) . The XY co-ordinates of the two poles and of the foci were recorded manually and processed automatically with Excel ( Microsoft ) software . The travelled distance was estimated over a period of 5 minutes by adding up the absolute values of the distances for all 10 sec interval as described before [12] . The ( x , y ) coordinates of the foci at every time point were recorded and the distance travelled in the 10 s interval calculated . For the mobility measurements at the home position , 30 foci were analyzed ( 30 cells with one focus for markers in the Ter MD and 15 cells with two foci for markers in other regions of the chromosome ) . To measure the mobility of markers , the distance travelled by various foci was recorded over a period of 5 minutes with 10 sec intervals when markers were at home position . In the growth conditions used , at home position , markers of Ori , Right and Left MDs as well as markers from NS regions were segregated in the two halves of the cells whereas markers of the Ter MD are found at mid-cell [12] . This index is calculated by comparing the theoretical number of genes ( nbth ) [29] to the experimental number of genes , i . e . the number of foci detected by microscopy ( nbobs ) [12] . The normalization of the index is obtained by the formula: [ ( nbth−nbobs ) /nbth] .
|
The large size of genomes compared to cell dimensions imposes an extensive compaction of chromosomes compatible with various processes of DNA metabolism , such as gene expression or segregation of the genetic information . Most bacterial genomes are circular molecules , and DNA replication proceeds bidirectionally from a single origin to an opposite region where replication forks meet . In the bacteria Escherichia coli , the long-range organization of the chromosome relies on the presence of mechanisms that structure large regions called macrodomains . The macrodomain containing the terminus of replication is structured by a specific organization system involving the binding of the protein MatP to 23 matS sites scattered over the 800-kb-long Ter region . In this report , we describe a site-specific insulation system that restricts to the Ter region the consequences of the mechanism structuring the Ter macrodomain . We identified two 12-bp sequences flanking the Ter macrodomain and one protein that are required to isolate the Ter region from the other parts of the chromosome .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"biology",
"microbiology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] |
2012
|
Long-Range Chromosome Organization in E. coli: A Site-Specific System Isolates the Ter Macrodomain
|
The morphogenesis of retroviral particles is driven by Gag and GagPol proteins that provide the major structural component and enzymatic activities required for particle assembly and maturation . In addition , a number of cellular proteins are found in retrovirus particles; some of these are important for viral replication , but many lack a known functional role . One such protein is clathrin , which is assumed to be passively incorporated into virions due to its abundance at the plasma membrane . We found that clathrin is not only exceptionally abundant in highly purified HIV-1 particles but is recruited with high specificity . In particular , the HIV-1 Pol protein was absolutely required for clathrin incorporation and point mutations in reverse transcriptase or integrase domains of Pol could abolish incorporation . Clathrin was also specifically incorporated into other retrovirus particles , including members of the lentivirus ( simian immunodeficiency virus , SIVmac ) , gammaretrovirus ( murine leukemia virus , MLV ) and betaretrovirus ( Mason-Pfizer monkey virus , M-PMV ) genera . However , unlike HIV-1 , these other retroviruses recruited clathrin primarily using peptide motifs in their respective Gag proteins that mimicked motifs found in cellular clathrin adaptors . Perturbation of clathrin incorporation into these retroviruses , via mutagenesis of viral proteins , siRNA based clathrin depletion or adaptor protein ( AP180 ) induced clathrin sequestration , had a range of effects on the accuracy of particle morphogenesis . These effects varied according to which retrovirus was examined , and included Gag and/or Pol protein destabilization , inhibition of particle assembly and reduction in virion infectivity . For each retrovirus examined , clathrin incorporation appeared to be important for optimal replication . These data indicate that a number of retroviruses employ clathrin to facilitate the accurate morphogenesis of infectious particles . We propose a model in which clathrin contributes to the spatial organization of Gag and Pol proteins , and thereby regulates proteolytic processing of virion components during particle assembly .
To establish a productive infection in host cells , retroviruses have evolved strategies that employ numerous host factors to facilitate their replication . Recently , several groups have applied genome-wide RNAi screens to identify hundreds of candidate host factors that may facilitate human immunodeficiency virus-1 ( HIV-1 ) and murine leukemia virus ( MLV ) infection [1] , [2] , [3] . Other strategies to identify host factors that facilitate virus replication include the identification of proteins that bind to viral proteins [4] and analysis of the proteomes that are incorporated into virions [5] , [6] , [7] . Indeed , host proteins involved in HIV-1 budding , such as Tsg101 [8] and ALIX [9] can be found in virions . Additionally , proteins that modulate virion infectivity such as cyclophilin A ( CypA ) [10] , and Hsc70 [11] can also be demonstrated to be virion components . However , while proteomic analyses of purified HIV-1 or MLV particles have revealed dozens of virion-associated host proteins , no biological significance has been attached to the virion association of many of them . One such protein is clathrin , which previous reports suggest is only passively incorporated into particles [6] . Clathrin has been intensively studied in the context of cell biology ( reviewed in [12] ) . It is a cytosolic protein that functions in vesicle genesis and transport and , specifically , mediates endocytosis from the plasma membrane and cargo trafficking from the trans-Golgi network ( TGN ) . Clathrin is comprised of a trimer of 180 kDa heavy chains ( HC ) that are arranged with their N-terminal adaptor binding domains at the extremities of each leg of a triskelion , while clathrin light chains ( LC ) bind to heavy chains close to their C-termini . Clathrin adaptors ( such as the AP family ) govern the sorting of specific cargoes into clathrin-coated vesicles and recruitment of clathrin to membranes ( reviewed in [13] ) . Many of these adaptors contain motifs such as LΦXΦ[DE] ( Φ indicates a bulky hydrophobic residue ) , or in the case of AP180 , repeated motifs with the sequence DLL , which bind to the clathrin N-terminal β–propeller domain , and facilitate the recruitment of clathrin to the plasma membrane [14] , [15] . Here , we demonstrate that clathrin is abundantly and specifically incorporated into a range of diverse retrovirus particles , including HIV-1 , SIVmac , MLV and M-PMV through interactions with Gag or Pol proteins . Indeed , several retroviral Gag proteins were found to encode peptide motifs that drive clathrin incorporation and mimic those found in cellular clathrin adaptor proteins . We also show that mutations in the motifs that mediate clathrin recruitment have a range of effects on the accuracy of particle morphogenesis , or on virion infectiousness , depending on the particular retrovirus that was examined . In several cases , we demonstrate that these effects can be recapitulated by reducing the available levels of clathrin in cells .
Initially , to discover potential host factors involved in HIV-1 replication , we set out to identify cellular proteins that are incorporated into HIV-1 particles , using a slightly different strategy compared to previous studies . To minimize the “noise” in such experiments , such as contaminating cell debris , or passively incorporated proteins , and prevent degradation of incorporated cellular proteins by the viral protease , plasmids expressing codon-optimized HIV-1 Gag alone , or protease-inactive ( D25A , PR- ) GagPol proteins were transfected into 293T cells . This generated a very high yield of virus-like particles ( VLPs ) without apparent cytotoxicity . Silver and Coomassie blue staining of SDS-PAGE gels ( Figure 1A ) , loaded with iodixinal gradient-purified GagPol VLPs revealed that five cellular proteins were abundantly incorporated into GagPol VLPs and were resistant to digestion by externally applied subtilisin . Mass spectroscopic analysis of the excised bands identified these proteins as CypA , Hsp70 , Hsp90 , AIP1/ALIX and clathrin HC ( Figure 1A ) , all of which have previously been identified as components of HIV-1 particles [5] , [6] , [9] , [10] , [11] . Strikingly , and in contrast to the other proteins that were found in HIV-1 GagPol VLPs , both silver staining and Western blot analysis revealed that clathrin HC was undetectable in particles simultaneously generated using the HIV-1 Gag protein alone ( Figure 1A ) . Moreover , visual inspection of silver or Coomassie stained gels suggested that clathrin HC was packaged into GagPol VLPs at exceptionally high level , i . e . the level of clathrin HC in virions approached that of GagPol ( Figure 1A , right panel ) . To verify this finding with authentic virion particles , the T-cell lines CEMX174 and MT2 were infected with VSV-G pseudotyped , Env-defective HIV-1 at a multiplicity of infection of ∼1 , washed extensively and progeny virions were harvested 40 h later ( Figure 1B ) . Western blot analysis revealed that clathrin HC was abundantly incorporated into these virions . Indeed , in experiments where virions were simultaneously generated via infection of T-cell lines or transfection of 293T cells and the amount of clathrin incorporated into virions directly compared by quantitative fluorescence-based Western blotting ( LI-COR ) , we found that the T-cell derived virions incorporate as much or more clathrin than the 293T-derived particles ( Figure S1 ) . Moreover , we determined that 6 to 8% of the total clathrin HC in the HIV-1 infected T-cell cultures appeared to be present in extracellular virions rather than cells ( Figure S1 ) . Consistent with previous findings that clathrin LC binds to the C-terminal portion of clathrin HC and forms a stable complex with it [12] , clathrin LC was also incorporated into HIV-1 virions . Indeed , when green fluorescent HIV-1 VLPs were generated by coexpression of HIV-1 GagGFP and GagPol in cells stably expressing fluorescently tagged clathrin LC ( DsRed-clathrin LC ) , about 40% of the VLPs were labeled with sufficient DsRed-clathrin LC to be visualized by deconvolution microscopy ( Figure 1C , Table S1 ) . In contrast , only a few VLPs generated by coexpression of GagGFP and Gag were red-fluorescent . In similar experiments , authentic virions were generated by transfection with a proviral plasmid encoding YFP embedded in the stalk region of the MA domain of Gag into cells stably expressing DsRed-clathrin LC . In this case , 70–80% of YFP+ virions contained sufficient DsRed-clathrin LC to be visualized as colocalizing red fluorescent puncta ( Figure 1D , Table S1 ) . Given that clathrin incorporation into HIV-1 particles appeared Pol-dependent , we next attempted to map the sequences responsible for its incorporation , using particles generated by expressing protease-defective GagPol proteins and detection of virion associated clathrin using Western blot or microscopy assays . Precise deletion of the reverse transcriptase ( RT ) or integrase ( IN ) domain in this context did not markedly affect VLP release but , surprisingly , both manipulations abolished clathrin incorporation ( Figure 1C and Table S1 ) . Similarly , several Pol truncations or point mutations , constructed in the context of protease-defective proviral plasmid , also inhibited or abolished clathrin incorporation ( Figure 1E and Table S1 ) . Specifically , mutations in reverse transcriptase ( RT ) , including L234A , that inhibit RT dimerization [16] prevented clathrin incorporation as did deletion of the IN C-terminal domain ( CTD ) as well as two so-called ‘class II’ IN mutations , N184L and F185K [17] . Western blot analysis of virions indicated that , while the levels of GagPol protein in cells and virions were not apparently affected by these mutations , clathrin incorporation was inhibited ( Figure 1E ) . We even found that the presence of the non-nucleoside RT inhibitor efavirenz , which stimulates RT dimerization [16] , during the production of VLPs , inhibited clathrin incorporation into virions ( Table S1 ) . In contrast , no effect on clathrin incorporation was observed for the active site IN point mutant E152K ( Figure S2 ) . Taken together , these findings demonstrate that multiple Pol domains are strictly required for clathrin incorporation into HIV-1 VLPs . This suggests that the overall conformation of Pol is critical for clathrin packaging , and effectively made it impractical to map small motifs in Pol that might be responsible for clathrin recruitment . Clathrin adaptors GGA and AP-1 , 2 and 3 are known to bind to the N-terminal 7-bladed β–propeller domain of clathrin HC [13] . This clathrin domain , encoded by residues 1–494aa is known to fold autonomously when expressed in the absence of other clathrin domains [18] . Co-expression of N-terminal 1–494aa or 1–363aa fragments of clathrin along with HIV-1 GagPol resulted in specific incorporation of these clathin fragments into wild-type VLPs , but not into IN δCTD VLPs ( Figure 1F ) , mimicking the property of endogenous full-length clathrin . Thus , the N-terminal adaptor binding domain of clathrin HC was sufficient to drive its incorporation into HIV-1 GagPol VLPs . To determine whether clathrin incorporation into virions is a general feature of primate lentiviruses , we next determined whether it also occurred in SIVmac . In contrast to results obtained with HIV-1 Gag , analysis of VLPs generated using only SIVmac Gag showed that clathrin was abundantly incorporated into VLPs in the absence of Pol ( Figure 2A ) . Mapping experiments in which chimeric SIVmac/HIV-1 Gag proteins ( Figure 2B ) were used to generate VLPs revealed that Gag proteins encoding the SIVmac p6 domain , specifically SIV ( HIV MA ) and HIV ( SIV p6 ) , yielded VLPs containing clathrin ( Figure 2B and 2C lane 4 and lane 6 ) . Conversely , the reciprocal chimeric Gag proteins , HIV ( SIV MA ) and SIV ( HIV p6 ) , generated VLPs that did not incorporate clathrin ( Figure 2B and 2C lane 3 and lane 5 ) . Thus , the differential abilities of HIV-1 and SIVmac Gag VLPs to incorporate clathrin were clearly governed by the p6 domain . Inspection of the SIVmac p6 protein sequence revealed two ‘DLL’ motifs ( positioned at p6 residues 21–23 and 51–53 ) that are absent in HIV-1 p6 ( Figure 2D ) . Because the clathrin adaptor AP180 employs multiple copies of a DLL motif in its C-terminal domain to directly bind to clathrin HC [15] , we mutated either or both DLL motifs in SIVmac p6 and tested the ability of the mutant Gag proteins to drive clathrin incorporation into VLPs . Western blot analysis indicated that mutations in both motifs ( D21A , D51A ) or the second motif only ( D51A or L52S ) reduced clathrin incorporation to almost undetectable levels , while mutation in the first DLL motif alone had little effect ( Figure 2E ) . To determine the clathrin domain that mediates packaging into SIVmac Gag VLPs , expression plasmids encoding the 1–494aa or 1–363aa clathrin HC N-terminal domains were coexpressed with SIVmac Gag . As shown in Figure 2F , SIVmac VLPs efficiently incorporated the clathrin HC N-terminal 1–494aa fragment , but unlike HIV-1 , the 1–363aa fragment was poorly incorporated ( Figure 2F , bottom panel ) . The significance of the difference in clathrin sequence requirements for incorporation into HIV-1 versus SIVmac virions is unclear at present , but in both cases the N-terminal adaptor binding domain appeared to be responsible for driving incorporation . Next , we tested whether SIVmac Pol , like HIV-1 Pol , could also drive clathrin incorporation . To facilitate clathrin detection , an HA tagged clathrin N-terminal domain ( 1–494aa ) was co-transfected with plasmids expressing either wild-type SIVmac GagPol , or mutants in which either ( i ) the DLL motifs in p6 were mutated ( DLL- GagPol ) , ( ii ) IN was mutated ( N184L ) in a way analogous to that which blocks clathrin incorporation into HIV-1 GagPol VLPs , or ( iii ) both p6 and IN were mutated . SIVmac DLL- GagPol exhibited greatly diminished clathrin packaging into VLPs ( Figure 2G ) . However , some clathrin incorporation was observed in SIVmac DLL- GagPol VLPs , and this incorporation was completely abolished by additional mutation at IN residue N184 ( Figure 2G ) . Thus , in SIVmac , both Gag and Pol contribute to clathrin incorporation into virions , but Gag appears to play the dominant role and drives significantly more clathrin incorporation than Pol . The finding that clathrin was specifically packaged into HIV-1 and SIVmac VLPs prompted us to ask whether clathrin incorporation is a general property of retroviruses . Inspection of a variety of retroviral Gag protein sequences revealed that some , but not all , encoded putative clathrin binding peptides , including DLL and LLTLD motifs in their Gag proteins . In particular , a prototype gammaretrovirus , MLV and a prototype betaretrovirus , M-PMV were selected for further investigation . Putative clathrin binding motifs , DLL and DLISLD respectively , were found in their Gag proteins proximal to their late domains ( a DLL motif at 156–158aa in MLV Gag and a DLISLD motif at 129–133aa of M-PMV Gag respectively , Figure 3A ) . Each of these viruses was found to package either endogenously expressed clathrin HC , or coexpressed N-terminal adaptor-binding domain fragments of clathrin HC ( 1–494aa or 1–363aa , Figure 3B , C , D , E ) into virions . This incorporation was specific because mutations in respective putative clathrin binding sites in Gag ( DLL to ALL in MLV or DLISLD to DAASLD in M-PMV ) dramatically reduced this incorporation . Notably , the yield of virion particles was unaffected by mutations in these clathrin recruiting motifs ( Figure 3B , C , D , E ) . Mutations in other candidate clathrin HC binding sites in Gag ( 533LLTLD537 at the C-terminus of MLV Gag and 30DLL32 in the matrix domain of M-PMV , respectively ) caused no reduction in clathrin incorporation , indicating that these other candidate motifs do not play a critical role in clathrin packaging ( unpublished observations ) . Overall , these findings demonstrated that clathrin can be specifically incorporated into virions from widely divergent retroviruses , and that this incorporation is driven by the adaptor-binding domain at the N-terminus of clathrin HC . Moreover , clathrin incorporation sometimes occurs via the action of peptide motifs in viral structural proteins that mimic those found in cellular clathrin adaptors . To probe the role of clathrin in retrovirus replication , we adopted a variety of approaches , including the analysis of viral mutants that were defective for clathrin incorporation , as well as depletion of clathrin using siRNA based approaches . Attempts to deplete clathrin using siRNA were complicated by the fact that it is highly abundant , has a relatively long half-life ( 20 h–50 h ) [19] , [20] and is essential for various cellular functions and for cell viability . Therefore , while we were able to reproducibly deplete about 70–80% of endogenous clathrin HC ( Figure S3A ) , it proved nearly impossible to completely deplete the intracellular clathrin to a sufficient extent such that a clathrin-deficient viral phenotype could be analyzed . Moreover , since clathrin plays a critical role in a number of cellular pathways , including trafficking of proteins through the secretory pathway , distinguishing the direct effects of depletion from indirect effects is challenging . Therefore , in order to investigate the role that clathrin plays in retrovirus life cycles , we employed multiple approaches including characterizing viruses with mutations that prevent clathrin incorporation , reducing clathrin expression by siRNA depletion ( Figure S3A ) , and overexpressing the C-terminal domain of the clathrin adaptor AP180 to induce clathrin sequestration ( Figure S3B ) [21] , [22] . Several mutations in HIV-1 Pol were found to block clathrin incorporation , including mutations in RT at the dimer interface ( L234A ) , as well as class II and CTD-truncating mutations in IN ( N184L and δCTD ) ( Figure 1E ) . Unfortunately , these mutations are pleiotropic , and may therefore have multiple effects on viral replication by perturbing the tertiary structure and function of the Pol protein . Although these mutations did not have discernable effects on Gag and GagPol precursor protein levels in cells or VLPs when the viral protease was inactivated by mutation ( Figure 1E ) , these mutations induced aberrant proteolytic cleavage of Pol proteins and reduced levels of GagPol precursor in cells and virions when protease was active [23] , [24] ( Figure 4A , Figure S4 ) . Specifically , when HIV-1 proviruses were expressed in 293T cells , the WT and Pol-mutant viruses generated similar levels of Gag protein and its processed derivatives , but the levels of GagPol precursor , partly processed intermediates and mature IN proteins , detected using an anti-IN antibody , were diminished in the mutants that failed to package clathrin into virions ( Figure 4A , Figure S4 ) . Similarly , in a series of constructs that were made to express GagPol protein with an HA-epitope fused at C-terminus of Pol , reduced GagPol levels were detected using an anti-HA antibody when mutations that blocked clathrin were introduced , and aberrant processed derivatives were detected ( Figure 4B ) . While the RT and IN mutations both blocked clathrin incorporation and induced PR-dependent Pol depletion , it was not clear whether or how these two effects were causally related to each other , given the pleiotropic nature of these mutations . However , these results did suggest the possibility that clathrin recruitment by Pol might inhibit its viral protease-dependent depletion . To investigate this possibility , we examined Pol processing in cells where the clathrin HC N-terminal domain ( 1–494aa ) or the full-length clathrin HC was overexpressed . Notably , for two Pol mutants , namely RT ( L234A ) and IN ( F185K ) , clathrin HC ( 1–494aa ) or full length clathrin HC overexpresssion led to an increase in the levels of GagPol derived proteins , including mature IN ( Figure 4B and 4C ) . Conversely clathrin overexpression had no , or only a slight effect , on IN ( N184L or δCTD ) mutants ( Figure 4B and 4C ) . Importantly , the effects of clathrin HC overexpression were specific to Pol; there was no significant effect on Gag precursor or processed derivative levels . To further explore the idea that clathrin recruitment might stabilize Pol in the presence of an active viral protease , we fused the clathrin recruiting domain from SIVmac Gag , namely p6 ( mutated in such a way so as not bind to Tsg101 and ALIX ) , to the C-terminus of HIV-1 GagPol ( IN δCTD ) . This chimera expressed higher levels of Pol protein expression than an equivalent construct containing the SIVmac p6 domain in which the DLL motifs were mutated ( Figure 4D ) . Together , these results are consistent with the notion that clathrin recruitment results in the stabilization of Pol proteins in the presence of an active viral protease . In support of this idea , Western blot analyses revealed that clathrin HC depletion using siRNA reduced the levels of GagPol precursor , IN , and partly processed intermediate Pol proteins in cells transfected with a WT HIV-1 proviral construct ( Figure 4E ) . Importantly , the levels of Gag protein ( which are translated from the same viral mRNA species ) were not reduced by this manipulation . Furthermore , AP180C overexpression also reduced the levels of intracellular GagPol , IN and intermediate proteins without affecting Gag levels ( Figure 4F ) , again suggesting that clathrin specifically stabilizes Pol proteins in HIV-1 infected cells . When introduced into GagPol expression vectors that were then used to transfect 293T cells , the class II HIV-1 IN mutations that blocked clathrin incorporation ( N184L , F185K or δCTD ) had only minor effects on overall particle yield ( Figure 5A ) . However , as expected , viruses encoding these mutations exhibited extremely low infectivity ( Figure 5B ) . More importantly , depletion of clathrin HC using siRNA did not affect overall particle yield from 293T cells ( Figure 5C ) , but caused significant decrease in infectiousness ( 5 to 20-fold ) of HIV-1 particles generated by cotransfection with an HIV-1 proviral plasmid ( Figure 5D ) . Alternatively , overexpression of AP180 C-terminal domain ( AP180C ) , which binds clathrin and induces its mislocalization [21]–[22] , reduced the incorporation of clathrin into HIV-1 particles ( Figure S5 ) and also reduced the infectiousness of HIV-1 particles generated from a proviral plasmid by >100 fold ( Figure 5E ) without affecting physical particle yield ( Figure 5C ) . However , clathrin HC depletion using siRNAs , or perturbation by AP180C overexpression , had comparatively modest , but nevertheless significant , effects ( ∼3-fold ) on the infectivity of VSV G-pseudotyped HIV-1 particles ( Figure 5F and 5G ) . Thus , while clathrin depletion and sequestration clearly impacted Pol protein levels and , consequently , virion infectivity , there was a significant discrepancy in the magnitude of the infectivity impairment induced by clathrin perturbation when VSV-G pseudotyped HIV-1 particles were examined as compared to those generated using the natural HIV-1 envelope . Given clathrin's role in the secretory pathway , it was possible that clathrin perturbation might have effects on the HIV-1 envelope protein that impact virion infectivity , independent of its effects on Pol . In fact , generation of HIV-1 particles in the presence of AP180C resulted in virions that contained primarily unprocessed gp160 envelope protein , and very little gp120 ( Figure S6 ) . Therefore , it appears likely that clathrin affects the trafficking of the HIV-1 Env protein , or cellular factors required for Env maturation . While these results indicate caution in the interpretation of the effects of clathrin perturbation on virion infectivity , results with M-PMV ( see below ) suggest the clathrin perturbation using siRNA or AP180C expression does not cause a non-specific effect on the infectivity of retroviruses , and in particular the function of the VSV-G envelope protein . Thus , the effect of clathrin siRNA and AP180C on VSV-G pseudotyped HIV-1 infectivity ( Figure 5F and 5G ) should be independent of effects on the envelope protein . In SIVmac , MLV and M-PMV Gag proteins , we noticed that motifs responsible for clathrin incorporation into virions were situated proximal to motifs responsible for recruitment of factors ( ESCRT pathway-associated proteins or ubiquitin ligases ) involved in viral budding ( Figure 2D and 3A ) . In the case of SIVmac , one of the two DLL motifs is positioned overlapping the putative ALIX binding site at the C-terminus of p6 ( Figure 6A ) . Therefore , to facilitate an examination of the role for clathrin in SIVmac replication , we first identified residues that were critical for ALIX recruitment , as well as residues that were critical for clathrin recruitment that could be mutated without disrupting ALIX recruitment . An SIVmac Gag expression plasmid was subjected to scanning mutagenesis throughout residues 41–60 of p6 ( Figure 6A ) , with mutations selected so as not to alter the underlying p6-Pol ( p6* ) protein sequence . Mutant SIVmac Gag proteins were coexpressed with HA-tagged ALIX , and ALIX incorporation into VLPs was assessed by Western blot analyses ( Figure 6B ) . Mutations P44L Y45S , L52S and D51A/L52S dramatically affected ALIX incorporation into VLPs , while the mutations D21A , D51A , L58P and F59S , did not ( Figure 6B ) These results are consistent with a recent study that also mapped the ALIX binding site in SIVmac p6 [25] . Thus , the SIVmac p6 domain was capable of recruiting clathrin and ALIX into VLPs through overlapping peptide sequences ( D21LL D51LL for clathrin and P44Y45 L52 for ALIX ) , but mutants were readily identified that separately inhibited these activities . Therefore , in the ensuing studies , an SIVmac Gag protein encoding the D21A and D51A mutations in p6 was termed ( DLL- ) and used as a clathrin–recruitment defective mutant while the P44L/Y45S mutant was termed ( PY- ) ( Figure 6A ) and used as a ALIX-recruitment defective mutant . To investigate the potential role of clathrin recruitment by Gag in SIVmac replication , we first compared the infectivity of particles generated by wild type and ( DLL- ) SIVmac proviral plasmids in a single round infectivity assay . A modest reduction ( 2–3 fold ) in infectiousness was observed as a consequence of DLL motif mutations , with no obvious effect on physical particle yield ( Figure 6C and 6D ) . Similar results were obtained when VSV-G pseudotyped SIVmac particles were used ( unpublished observations ) . However , in spreading replication assays conducted in C8166 cells , this modest difference was apparently amplified , and the SIVmac ( DLL- ) virus was highly attenuated compared to the wild-type counterpart ( Figure 6E ) . Given the proximity of the DLL motifs to the L-domains in SIVmac Gag ( Figure 6A ) , we next tested the effects of mutations predicted to abolish the recruitment of Tsg101 ( PTAP- ) , clathrin ( DLL- ) and ALIX ( PY- ) , either alone or in combination , on particle release . Initially this was done in the context of SIVmac Gag , in the absence of the viral protease ( Figure 6F ) . Consistent with a previous report [26] , mutation of the PTAP motif alone had a surprisingly modest effect on SIVmac particle yield . Moreover , mutation of the ALIX binding sites had no discernable impact on particle yield , when introduced either alone or in combination with the PTAP mutation ( Figure 6F and 6G ) . Strikingly , while the DLL- mutation had no effect on VLP yield when introduced alone , the combination of this mutation with the PTAP- mutation dramatically diminished VLP yield ( Figure 6F ) . This finding held true when the SIVmac Gag protein was expressed using a human codon optimized construct ( Figure 6G ) . Overexpression of ALIX has previously been reported to rescue PTAP-mutant HIV-1 particle release [27] , [28] and we found that overexpression of ALIX restored the defect in VLP release associated with the PTAP- DLL- double mutant ( Figure 6G ) . This activity required an intact ALIX binding site , because ALIX overexpression failed to rescue the budding defect in the PTAP-DLL-PY- triple mutant ( unpublished observations ) . This finding indicates that the PTAP-DLL- double mutant is fully competent to assemble into virions , but requires ALIX overexpression to complete budding . Importantly we also found that overexpression of AP180C recapitulated the effect of the DLL- mutation in SIVmac Gag , and inhibited the release of VLPs assembled using PTAP- Gag ( Figure 6H ) . AP180C overexpression did not inhibit particle release when wild-type Gag was used and this finding suggested that clathrin facilitates the completion of assembly and/or budding of SIVmac particles , particularly when budding is impaired or retarded by inhibition of Tsg101 recruitment . To further characterize the defects in VLP release imposed by the PTAP and DLL- mutations , we analyzed 293T cells expressing wild-type and mutant SIVmac Gag proteins via scanning electronic microscopy . To eliminate variations from transfection levels , Gag expression constructs were utilized which contained an internal ribosomal entry site ( IRES ) linked to a GFP coding sequence on the same mRNA as Gag , and cells expressing equivalent amounts of GFP were chosen for analysis . Virions assembled using Gag proteins bearing individual PTAP- or DLL- mutations exhibited no gross morphological abnormalities , and numerous apparently spherical particles were observed on the plasma membrane of cells ( Figure 6I and Figure S7 ) . However , PTAP- DLL- double mutant Gag proteins exhibited a clear morphogenesis defect ( Figure 6I and Figure S7 ) . Specifically , cells exhibited hemispherical protrusions from their surfaces , but complete spherical particles were almost never observed . Crucially , this morphological defect could be induced using the PTAP- single mutant SIVmac Gag protein ( but not the wild type Gag protein ) upon overexpression of AP180C ( Figure 6J and Figure S8 ) . Together , these data strongly suggest that clathrin plays a facilitating role in the morphogenesis of SIVmac virions that is modest when measured in the context of a single cycle of SIVmac assembly , but is sufficient to strongly enhance replication , and becomes particularly evident when budding is inhibited or slowed by mutation of the PTAP L-domain . When SIVmac Gag proteins were expressed in the context of an active viral protease , either using GagProtease ( Figure 7A ) or full-length GagPol ( Figure 7B ) expression plasmids , the DLL- single mutation had little or no effect on levels of cell associated Gag protein ( Figure 7A and 7B ) . However , when the DLL- mutation was present in combination with the PTAP- mutation , the steady state cell-associated levels of the Gag precursor and processed derivatives were significantly diminished ( Figure 7A and 7B ) . Remarkably , overexpression of ALIX rescued this defect and restored the level of PTAP- DLL- mutant SIVmac Gag proteins to those expressed by the wild type Gag-protease and GagPol expression plasmid ( Figure 7A and 7B ) . Notably , the ability of ALIX to restore PTAP- DLL- mutant SIVmac Gag protein levels required an ALIX binding site , as this effect was not observed when a PTAP- DLL- PY- mutant SIVmac GagPol expression plasmid was used ( Figure 7B ) . The effect of the DLL- mutation on SIVmac Gag expression levels was partially recapitulated by clathrin sequestration using AP180C . Specifically , AP180C overexpression caused a reduction in the levels of Gag protein expressed by a PTAP- PY- mutant , but not wild-type , GagPol expression plasmid ( Figure 7C ) . Taken together , these findings lead to the conclusion that clathrin interaction with SIVmac Gag facilitates virion morphogenesis , with consequent effects on Gag protein stability in the presence of the viral protease , especially when budding is impaired . The effects of the DLL- mutation on SIVmac Gag protein levels were specific to Gag , and did not affect Pol protein levels ( Figure 7B ) . Similarly , a mutation in SIVmac Pol ( analogous to the HIV-1 IN N184L ) that blocked Pol dependent clathrin incorporation ( Figure 2G ) caused a reduction in the level of SIVmac Pol protein to an almost undetectable level , while Gag levels were unaffected ( Figure 7B ) . In contrast to the effect of ALIX on SIVmac Gag expression , this Pol expression defect could not be rescued by overexpression of ALIX . Thus , the effects of mutations that reduce protease-dependent Gag stability and clathrin incorporation were independent of those that affected Pol stability and clathrin incorporation . Like SIVmac , MLV has a DLL motif proximal to its L-domain that is responsible for clathrin incorporation ( Figure 3A ) . Transfection of an MLV GagPol expression plasmid in which the DLL motif was mutated to either ALL or DAA , along with plasmids encoding VSV-G and a GFP expressing MLV vector resulted in no defect in the yield of physical particles ( Figure 8A ) , but mutant particles were >100-fold less infectious than WT particles ( Figure 8B ) . Similar results were obtained using full-length MLV proviral plasmids ( unpublished observations ) . Similarly , siRNA mediated clathrin depletion did not affect particle yield ( Figure 8C ) . However , clathrin depletion resulted in only a modest reduction of MLV infectivity ( Figure 8D ) . This may be attributable to incomplete clathrin knockdown ( approximately 20% of normal clathrin levels remained in siRNA transfected cells , Figure S3A ) . Consistent with this notion , coexpression of WT and DLL- mutant MLV GagPol at various ratios revealed that particles were nearly fully infectious , even when a small fraction of the Gag proteins contained therein harbored an intact DLL motif ( Figure 8E ) . Therefore , we hypothesize that the residual clathrin recruited by WT virus assembled in siRNA treated cells may be nearly sufficient to fulfill its functional role . Western blot analysis showed that substitutions in the MLV DLL motif had no effect on cell- or extracellular virion- associated levels of viral precursor Gag ( Pr65 ) or processed p30 CA proteins when expressed in the context of MLV GagPol or proviral plasmids ( Figure 8A and unpublished observations ) . This held true when the DLL mutation was introduced alone , or in combination with mutations in PPPY and/or PSAP motifs ( unpublished observations ) . However , Western blot analysis with an anti-p12 monoclonal antibody revealed a dramatic decrease in the levels of the mature p12 protein associated with virions ( Figure 8F ) . Conversely the partly processed MA-p12 intermediate was present at equivalent abundance in wild type and mutant virions . This finding excluded the possibility that the p12 antibody failed to recognize the DAA mutant p12 sequence and instead suggested the possibilities that either the MA-p12 junction was not efficiently processed in the DAA mutant , or that the DAA mutation created a p12 that was aberrantly cleaved and as a consequence could not be recognized by the p12 antibody . To investigate these possibilities , a small amount of wild-type or DLL-mutant MLV Gag-Pol expression plasmid was co-transfected with increasing amounts of wild-type or mutant Gag expression plasmids ( in a Gag-Pol ∶ Gag ratio of 1∶0 . 5 to 1∶16 , ( Figure S9 ) ) . This analysis suggested that the DAA mutant and WT Gag proteins could be processed in trans by coexpressed GagPol bearing the WT p12 sequence , yielding mature p12 protein in extracellular virions with comparable efficiency . Conversely , the DAA mutant GagPol protein failed to generate the fully processed p12 in virions , even when a WT Gag protein was provided in trans ( Figure S9 ) . In a similar experiment , we used a different MLV-related GagPol expression plasmid ( from XMRV ) whose p12 sequence is poorly recognized by the anti MLV p12 antibody ( Figure S10 ) . When XMRV GagPol was coexpressed with WT or DLL mutant MLV Gag proteins , the WT and DLL mutant MLV p12 proteins were detected at approximately equivalent levels in the resulting virions ( Figure S10 ) . Thus , both the WT and mutant MLV MA-p12 junction can be efficiently processed by MLV or XMRV protease in trans resulting in the appearance of the mature p12 protein in virions . In other words , it appeared that the DLL motif regulates the activity of the MLV protease in cis , but not in trans , and the absence of the DLL motif in cis to the protease caused aberrant Gag processing and absence of p12 in virions . While the absence of the p12 protein in DLL-mutant MLV virions appeared symptomatic of the defect induced by the clathrin binding site mutation , it was not clear whether this lesion was the direct cause of the infectivity defect . To test this , we generated MLV particles using a construct , similar to one previously described [29] , in which the MA-p12 cleavage site was mutated ( S1KK ) so that it could not be cleaved , resulting in the absence of the mature p12 protein in virions ( Figure 8G ) . In this context , the DLL- mutation retained its deleterious effect on MLV infectivity ( Figure 8H ) . Thus , these findings suggested the existence of a generalized morphological defect that included , but was not limited to , p12 deficiency in MLV particles consequent to mutation of the clathrin binding site . Electron microscopic analysis of WT and DLL mutant MLV expressing cells did not reveal any gross morphological alterations in assembling or assembled particles ( Figure S11A ) . Therefore , to test for more subtle morphogenesis defects , we examined whether DLL- mutant particles retained the ability to saturate the TRIM5α restriction factor . This assay should test for the presence of a stable , morphologically accurate capsid lattice , which is expected to be required for efficient binding to , and saturation of TRIM5α . Thus , increasing amounts of N-tropic MLV particles generated using WT or DLL- mutant GagPol expression plasmids ( Figure S11B ) were applied to human TE671 cells , which express a TRIM5α protein that can recognize and restrict N-tropic MLV capsids [30] , [31] . Simultaneously , a fixed and equal amount of GFP-expressing WT B-tropic or N-tropic indicator virus was applied to monitor TRIM5α saturation ( Figure 8I ) . Wild-type N-tropic MLV particles efficiently saturated TRIM5α and thereby facilitated N-tropic MLV infection . However , particles containing a mutation in the DLL motif in p12 were approximately 10-fold less active in this TRIM5α saturation assay ( Figure 8I and S11B ) . The finding that the mutant virions were less ‘visible’ to human TRIM5α suggests that their cores were not properly formed , or unstable and , therefore , that the clathrin-binding motif in p12 is important for accurate MLV particle morphogenesis . Like SIVmac and MLV , the betaretrovirus M-PMV harbours a motif ( DLISLD ) proximal to its L-domains that is responsible for recruiting clathrin into virions . Notably , however , the morphogenesis pathway for M-PMV is quite different to that of SIVmac and MLV , in that complete immature capsids are assembled in the cytoplasm and move thereafter to the plasma membrane for envelopment . Mutation of the clathrin-recruiting motif in M-PMV Gag reduced the infectiousness of M-PMV virions by 5-fold , without affecting the yield of physical particles ( Figure 9A and 9B ) or causing gross morphological abnormalities that could be visualized by electron microscopic examination ( unpublished observations ) . Notably , clathrin depletion using siRNA or perturbation by overexpression of AP180C had a negative effect on M-PMV infectivity of similar magnitude to that of the clathrin binding site mutation ( Figure 9B and 9C ) . Importantly , however , the effect of depletion or sequestration of clathrin on M-PMV infectivity was specific to the WT virus; these manipulations had no effect on the infectivity of the mutant M-MPV that could not recruit clathrin ( Figure 9B and 9C ) , suggesting that the clathrin binding site mutation and the clathrin perturbation affected the same process . It was surprising that the clathrin knockdown had a greater effect on M-MPV infectivity compared to MLV infectivity ( compare Figure 8D and 9B ) , while M-PMV was less affected by clathrin binding site mutations than was MLV ( compare Figure 8B and 9B ) . However , experiments in which M-MPV virions were generated using mixtures of WT and DLISLD mutant proviral plasmids showed that reduction in infectivity was approximately proportional to the fraction of the Gag protein that was mutant versus WT ( Figure 9D ) . Thus , M-MPV infectivity appeared more sensitive to partial depletion of clathrin , or partial removal of clathrin binding sites from virions , than did MLV . Nonetheless , optimal infectiousness and replication of both viruses was clearly dependent on clathrin recruitment by their respective Gag proteins .
In this study , we identified clathrin as a component in a variety of retrovirus particles , including members of the lentivirus ( HIV-1 and SIVmac ) , gammaretrovirus ( MLV ) and betaretrovirus ( M-PMV ) genera . Specifically , the Gag proteins of SIVmac , MLV and M-PMV recruit clathrin HC using DLL or DLISLD motifs , which mimic those commonly found in classical clathrin adaptors . Additionally , HIV-1 Pol and , to some extent , SIVmac Pol were capable of recruiting clathrin into virions , although the motifs responsible could not be mapped because clathrin incorporation appeared to be dependent on the conformational integrity of multiple domains of the HIV-1 Pol protein . Phenotypic characterization of mutant viruses and a combination of other techniques to perturb clathrin in virus-producing cells ( siRNA-based clathrin HC knockdown or AP180C overexpression ) , revealed that clathrin can have a range of apparently distinct effects on virion morphogenesis , depending on the particular retrovirus examined . These effects are summarized in Table S2 . In the case of HIV-1 , mutations or drugs ( efavirenz ) that affect RT dimerization as well as class II IN mutations are known or thought to affect protease activation [17] , [23] , [24] , [32] , [33] , [34] , [35] . We found that at least some of these mutations blocked clathrin incorporation , and also found that clathrin overexpression could ameliorate deficits in the levels of Pol protein harboring some of the aforementioned mutations . Moreover , depletion or sequestration of clathrin could reduce the levels of Pol proteins in cells . This suggests that clathrin may be directly involved in regulating protease activity , or perhaps in stabilizing or retaining the products of Pol proteolytic cleavage during HIV-1 morphogenesis . Similarly , mutations in the clathrin-recruiting motif in the p6 domain of SIVmac Gag , as well as sequestration of endogenous clathrin , decreased the level of the viral Gag protein in the presence of the viral protease . However , these effects were only truly evident when the viral PTAP motif was also mutated . Again this finding is consistent with the notion that clathrin regulates proteolysis of the viral protein , or stabilizes the products of proteolysis , and the magnitude of the effect is exaggerated to easily detectable levels if budding is blocked or retarded . Notably , the apparent protease-dependent instability of the PTAP- DLL- mutant Gag protein could be largely reversed by overexpression of ALIX , strongly suggesting that the PTAP- DLL- mutant Gag protein is not generically unstable or acutely sensitive to proteolysis simply as a direct consequence of the introduced mutations . Rather , this observation reinforces the notion that protease-dependent SIVmac Gag depletion , consequent to a failure to recruit clathrin , is exaggerated by retarding the rate of particle budding . In the case of MLV , there was no apparent effect of the DLL- mutation on Gag precursor stability , even when the proximal PPXY L-domain motif was mutated . However , there was a very clear effect on the products of Gag proteolysis , in that p12 was absent from virions . Moreover , the DLL mutation had major effects on the infectivity of MLV particles , as well as the accuracy of their morphogenesis , or their stability , as evidenced by the relative inefficiency with which MLV cores were apparently recognized by the TRIM5α restriction factor . However the precise nature of the biochemical lesion responsible for this infectivity defect remains to be completely defined . The absence of p12 from virions could contribute to the infectivity defect , as p12 has been shown to be a component of MLV preintegration complexes [36] . However , mutations that prevent the cleavage of p12 from MA did not abolish the effect of the DLL-mutation on MLV infectivity , suggesting additional effects of clathrin on particle morphogenesis . In addition to effects on viral protein proteolysis , clathrin appears to affect the morphogenesis and release of SIVmac particles in the absence of the viral protease . Again , observing this effect required the PTAP-motif to be mutated , and the effect could be suppressed by overexpression of ALIX . The ability of ALIX overexpression to suppress the two effects of clathrin binding site mutation on SIVmac , namely; ( i ) an assembly defect in the absence of the viral protease and ( ii ) Gag instability in the presence of the viral protease , suggests that they are different manifestations of the same defect in clathrin recruitment . This apparent ability to affect both protease-dependent and protease-independent processes influences the generalized models that can be invoked to explain the role of clathrin in retrovirus morphogenesis ( see below ) . It is interesting that the D51L52L53 motif in SIVmac Gag overlaps with the ALIX binding site , raising the possibility that ALIX and clathrin might compete with each other . Moreover , inspection of multiple viral strains of the SIVsm/SIVmac/HIV-2 lineage reveals that all p6 sequences contain 1 , 2 or 3 copies of a DLL motif . In those that contain a single DLL motif , it appears that one copy has been displaced by a PTAP motif . These observations are suggestive of interplay between clathrin and the ESCRT machinery that merits future investigation . A caveat to the aforementioned conclusions is that mutations , particularly in HIV-1 or SIVmac Pol might have pleiotropic effects , and it therefore was difficult to assign cause and effect in situations where , for example , Pol mutations both blocked clathrin incorporation and caused decreases in the levels of Pol proteins . Nonetheless , depletion or sequestration of clathrin could , in several cases , at least partly recapitulate the effects of mutations that blocked clathrin recruitment . The pleiotropic effect of HIV-1 Pol mutations likely arises from the fact that they cause premature protease activation , and because protease has multiple substrates in Gag and Pol , a variety of effects on the accuracy of particle assembly and Pol protein incorporation , and therefore particle infectiousness , are predictable consequences . It is possible that the pleiotropic effect of these mutations is a consequence of their effects on clathrin recruitment . Indeed , overexpression of clathrin , or fusion of a clathrin binding site in cis could increase the levels of HIV-1 Pol proteins that were apparently destabilized by Pol mutations . However , it was not possible to restore the infectiousness of Pol-mutant particles using this approach . One factor that must be considered in arriving at a proposal for general model for a role of clathrin in retroviral replication was that the magnitude of the effect on infectious virion yield clearly varied according to which retrovirus was examined . Additionally , clathrin was not found to be efficiently incorporated into all retroviruses that we tested . For instance , equine infectious anemia virus ( EIAV ) and human endogenous retrovirus K ( HERV-K ) did not efficiently incorporate clathrin ( unpublished observations ) . Moreover , the apparent effect of clathrin on retrovirus replication usually varied according to whether the experimental manipulation was clathrin binding site mutation , clathrin depletion , or clathrin sequestration . In some cases ( e . g . SIVmac Gag ) , the effect of mutations in defined motifs responsible for clathrin recruitment was relatively modest in a single cycle of replication , while in others ( e . g . MLV ) the effect was dramatic . In the case of HIV-1 , SIVmac and MLV the effects of mutations that abolish clathrin incorporation were much greater than the effects of clathrin depletion or sequestration . While this may be due to pleiotropic effects of the mutations that were introduced , it is also true that clathrin depletion or sequestration using the techniques employed herein was incomplete and , at least in the case of MLV , a relatively small fraction of the viral Gag protein is needed to be capable of recruiting clathrin in order for virions to be nearly fully infectious . Conversely , in the case of M-PMV , the magnitude of the effect of clathrin binding site mutation on particle infectivity was nearly precisely recapitulated by clathrin depletion or sequestration . Moreover , the impaired infectivity of the clathrin recruitment-defective mutant was not further impaired reduced by clathrin depletion or sequestration . Thus , we can be quite confident in the case of M-PMV case that clathrin enhances infectivity primarily through the action of the DLISLD motif in Gag . Details of the mechanism by which clathrin enhances M-PMV infection remain to be investigated and are difficult at present due to the paucity of reagents for studying this virus , but M-PMV may make the most tractable system for future investigations . It is challenging , based on these data , to make definitive general conclusions as to the role of clathrin in retrovirus replication , because the ultimate outcome , in terms of defined lesions that occurred as a consequence of blocking clathrin recruitment , differed in each retrovirus studied herein ( Table S2 ) . Moreover , in some retroviruses Pol was responsible for clathrin recruitment , while in others Gag was responsible . Although conceptually unsatisfying , it may simply be the case that clathrin plays a completely different role in each retrovirus . However , for HIV-1 , SIVmac and MLV , a common theme was that clathrin appeared to influence proteolysis of the viral polyproteins . This would likely have different , potentially pleiotropic , consequences for different retroviruses , including infectivity and morphogenesis defects as well as apparent viral protein instability . We observed all of these outcomes in this study and each could be reasonably hypothesized to be the consequence of mis-regulation of viral polyprotein processing . Models that could potentially explain the aforementioned phenomena include the idea that clathrin contributes to the spatial organization of Gag and Pol proteins during particle assembly . As a homotrimeric Gag or Pol binding protein , clathrin could bind to multiple Gag or GagPol molecules simultaneously . Indeed , the affinity of clathrin for cellular clathrin adaptors , and likely , therefore , assembling Gag or Pol proteins , is dependent on their multimerization/polyvalency [12] , [13] . The association of clathrin with assembling Gag and Pol proteins could influence proteolysis of the viral proteins by positively or negatively influencing protease dimerization ( which is required for protease activity ) , by influencing substrate ( Gag or Pol ) mobility and/or accessibility to the protease , or by helping to retain or remove the products of partial or complete proteolysis from a nascent virion as particle assembly and proteolysis proceeds . Such a model can also be reconciled with the findings that the effects of clathrin appear quite variable , from overt effects on particle morphogenesis and viral protein stability to more subtle effects on virion infectiousness .
The proviral HIV-1 plasmid used throughout was pNL4-3 ( NIH AIDS Research and Reference Reagent Program , Catalog No . 114 ) . A protease defective variant , and other mutants thereof , were constructed by generating PCR products harboring mutations in PR or RT or IN , that were digested with SphI and EcoRI before subcloning . The SIVmac proviral plasmid was previously described [37] and based on SIVmac239 . An MLV proviral plasmid ( pNCS ) was a gift from Stephen Goff and the M-PMV proviral plasmid pSARM-4 was a gift from Eric Hunter . Overlap-extension PCR was used to generate HIV-1 , SIVmac , MLV , and MPMV mutants . For HIV-1 Gag , HIV-1 GagPol , SIVmac Gag , SIVmac GagProtease ( Gag-PR ) and SIVmac GagPol expression plasmids , wild-type or mutant encoding sequences were amplified by PCR , using primers that incorporated 5′EcoRI and 3′ NotI sites and inserted into the HIV-1-based expression plasmid pCRV1 [38] . In some cases an HA epitope was inserted , fused in frame at the C-terminus of the expressed protein into pCRV1 vector . The panel of SIVmac double mutants PTAP- ( PTAP/LTAL ) DLL- ( DLL/ALL ) , PTAP- ( PTAP/LTAL ) PY- ( PY/LS ) and DLL- ( DLL/ALL ) IN-N184L and triple mutants PTAP- ( PTAP/LTAL ) PY- ( PY/LS ) DLL- ( DLL/ALL ) were constructed by introducing additional point-mutation into pre-constructed single or double mutants , respectively . To generate chimeric HIV-1/SIVmac Gag proteins ( Figure 2B ) , overlap-extension PCR was performed with primers targeting the corresponding region of SIVmac or HIV-1 Gag and PCR products were inserted along with a C-terminal HA-epitope tag into pCRV1 . To insert mutant SIVmac p6 at the C-terminus of mutant HIV-1 GagPol , NotI sites were incorporated at both 5′ and 3′ end of the SIVmac p6-encoding PCR product which was inserted between HIV-1 Gag-Pol and the HA epitope in pCRV1 . Plasmids expressing codon-optimized SIVmac Gag , pCR3 . 1/SIVmac-Gag , as well as codon-optimized HIV-1 Gag , codon-optimized HIV-1 Gag-Pol and codon optimized GagGFP , were previously described [38] and mutations were made by overlap-extension PCR . To construct pCR3 . 1/SIVmacGag-IRES-eGFP expressing both SIVmac Gag and GFP , the IRES-eGFP region was amplified , using pIRES2-EGFP ( Clontech ) as template , digested by NotI and subcloned at the C-terminus of SIVmac Gag in WT and mutant pCR3 . 1/SIVmac-Gag expression plasmids . For MLV mutants , PCR products harboring D156LL- mutations ( DLL/ALL or DLL/DAA ) were digested using BsrGI and XhoI and inserted into pCAGGS-based MLV Gag or Gag-Pol expression vectors [39] or an N-tropic MLV GagPol expression plasmid [30] . For M-PMV mutants , PCR products bearing the DL129I130SLD/DAASLD mutation were digested by SmaI/SacI and inserted into pSARM-4 [40] or pSARM-X-eGFP as described in [41] . Plasmids expressing clathrin HC residues 1–363aa or 1–494aa with a C-terminal HA or Myc epitope were amplified by using full-length clathrin HC as a template and inserted into pCR3 . 1 ( Invitrogen ) . The plasmid expressing pCR3 . 1/ALIX , the HIV-1 proviral plasmid bearing YFP in the stalk region of matrix ( pNL4-3 MA-YFP ) and the XMRV ( xenotropic murine leukemia-related virus ) GagPol expression plasmid were described previously [42] , [43] , [44] . The plasmid expressing FLAG-AP180C was a gift from Lois Greene , a plasmid expressing DsRed-clathrin LC was a gift from Sanford Simon and pSIV-T1 was from Francois-Loic Cosset [45] . Monoclonal antibodies included anti-HA ( Covance ) , anti- Clathrin HC ( BD Transduction laboratories ) , anti-FLAG ( Sigma ) , anti-HIV capsid p24 ( 183-H12-5C ) , anti-Env ( 1D6 ) , anti-RT ( MAb21 ) ( all from NIH AIDS Research and Reference Reagent Program ) , anti-HIV IN ( a gift from Michael Malim ) , anti-MLV capsid p30 ( ATCC CRL-1912 R187 ) , anti-MLV p12 ( ATCC , CRL-1890 548 ) , and anti-M-PMV Gag ( a gift from Eric Hunter ) [46] . In addition , secondary antibodies included goat anti-mouse or anti rabbit IgG conjugated to horseradish peroxidase , or to Alexa Fluor 488 ( Invitrogen ) , IRDye® 800CW and IRDye® 680 ( LI-COR Biosciences ) . Adherent cell lines from human ( 293T , TE671 , TZM-bl ) were maintained in DMEM supplemented with 10% fetal calf serum and gentamycin . CD4+lymphoid cell lines CEMx174 , MT2 , MT4 and C8166 were grown in RPMI/10%FCS/antibiotics . In general , for transfection experiments in 293T cells , cells were seeded at a concentration of 1 . 5×105 cells/well ( 24-well ) or 3×105 cells/well ( 12-well ) or 2×106 ( 10-cm dish ) and transfected the following day using polyethylenimine ( PolySciences ) . To identify proteins incorporated into HIV-1 VLPs , 10 µg of an HIV-1 GagPol and Gag expression plasmids were transfected into 293T cells in 10 cm dishes . Particles in 30 ml of filtered supernatant were pelleted through 20% sucrose , resuspended in PBS , treated with subtilisin ( Sigma ) and separated on Optiprep gradients . Particulate material in each of 8 fractions was recovered by diluting the fractions with PBS and ultracentrifugation and analyzed by SDS-PAGE followed by Coomassie blue or silver staining . Bands of interest were excised and protein identification was done by the Rockefeller University Proteomics Center . To determine domains responsible for clathrin incorporation , 293T cells were transfected with 5 µg of HIV-1 or SIVmac GagPol or Gag expression plasmid along with 3 µg of clathrin heavy chain 1–494aa or clathrin heavy chain 1–363aa or 0 . 5 µg GFP-HA expression plasmid . Analysis of clathrin incorporation into other viruses was performed by transfection of 8 µg of MLV GagPol expression plasmid or 8 µg of M-PMV proviral plasmid , with or without 2 µg of clathrin HC 1–494aa or 2 µg of clathrin HC 1–363aa expression plasmids followed by ultracentrifugation and subtilisin treatment of pelleted virions . To measure SIVmac VLP release , cells were transfected with 150 ng of native sequence or codon-optimized SIVmac Gag expression plasmids , along with 300 ng of ALIX expression plasmid where indicated . In Figure 6H , 500 ng of codon optimized SIVmac Gag was cotransfected with 0 ng , 0 . 5 µg , 1 . 0 µg or 1 . 5 µg of AP180C expression vector . VLPs were harvested at 24 hrs post transfection , pelleted through 20% sucrose and subjected to Western blot analysis . To define the ALIX binding motif on SIVmac Gag , 500 ng of plasmids expressing wild-type or mutant SIVmac Gag proteins were cotransfected with 500 ng of pCR3 . 1/HA-ALIX into 293T cells . Cell lysates and VLPs were harvested and analyzed by Western blotting after 40 hrs . To generate infectious viruses bearing a GFP reporter , 293T cells were transfected with 200 ng of an HIV-1 GagPol expression plasmid and 200 ng of reporter vector pHRSIN-CSGW [47] ( for HIV-1 ) , 200 ng of an SIVmac GagPol expression plasmid and 200 ng of reporter vector pSIV-T1 ( for SIVmac ) , 200 ng of an MLV GagPol expression plasmid and 200 ng of reporter vector pCNCG [30] ( for MLV ) , or 400 ng of pSARM-X-eGFP ( for M-PMV ) , along with 100 ng of plasmid expressing the VSV-G envelope . To measure the effect of AP180C on infectiousness of viruses , 500 ng of empty vector or AP180C expression plasmid was added to the transfection mixture . To generate the VSV-G pseudotyped noninfectious VLPs used in the TRIM5 saturation assays in Figure 8I and Figure S11B , 10 µg of WT or DLL- mutant N-tropic MLV and 2 µg of VSV-G expression plasmid were used to transfect a 10 cm dish of 293T cells . To determine the effects of AP180C on HIV-1 or SIVmac viral protein expression level , 200 ng of HIV-1 proviral plasmid was cotransfected with 0 ng , 300 ng , 600 ng , 900 ng of AP180C expression plasmid ( Figure 4F ) , or 200 ng of SIVmac GagPol expression plasmid was cotransfected with 0 ng , 250 ng , 500 ng or 1000 ng of AP180C expression plasmid ( Figure 7C ) . To assess the ability of the MLV protease to process WT or mutant Gag proteins ( Figure S9 ) , 50 ng of WT or DLL/DAA mutant MLV GagPol expression plasmid were contransfected with increasing amounts ( 0 , 25 ng , 50 ng , 100 ng , 200 ng , 400 ng , 800 ng ) of plasmid expressing WT or DLL/DAA mutant MLV Gag . Alternatively , in Figure S10 , 6 µg of XMRV GagPol expression plasmid were cotransfected with 300 ng of WT or DLL/DAA MLV Gag expression plasmid . Cell lysates and VLPs were analyzed by Western blotting 48 hrs later . In all transfection experiments , the total amount of DNA was held constant within the experiment by supplementing transfection mixtures where necessary with empty expression vector . Transfected 293T cells were placed in fresh medium at 20 hrs post infection and virion containing cell supernatants were harvested and filtered ( 0 . 22 µm ) at 40 hrs post transfection . Infectious virus release ( for HIV-1 and SIVmac proviral plasmid derived virus ) was determined by inoculating TZM cells seeded in 96 well plates at 1 . 2×104 cells/well . At 48 hrs post infection , β-galactosidase activity was determined using GalactoStar reagent as per the manufacturer's instructions . Reporter viruses ( HIV-1 , SIVmac , Moloney MLV and M-PMV ) bearing a GFP indicator gene were generated by transient transfection of 293T cells along with VSV G envelope protein , in the absence or presence of AP180C as indicated above . TE671 cells were seeded one day prior to infection , inoculated with GFP reporter viruses and FACS analysis was carried out using Guava EasyCyte instrument ( Guava Technologies ) . TE671 were seeded at 2×104 cells/well in 24-well plates one day before infection . Cells were inoculated with a fixed dose of N-MLV or B-MLV GFP reporter virus , selected so that infection with the restricted virus in the absence of VLPs gave low , but measurable , levels of infection ( about 0 . 1% GFP-positive cells ) . Restriction-abrogating VLPs were serially diluted two-fold and added to the target cells simultaneously with the fixed dose of N-MLV in the presence of polybrene . Infection by the GFP reporter virus was measured 48 h later as described above . 293T cells were transfected with 60–100 pmol of a clathrin-specific RNA duplex ( SMART pool , Dharmacon ) or a control firefly luciferase duplex ( Dharmacon ) using Lipofectamine2000 ( Invitrogen ) according to manufacturer's instructions . Twenty-four hours post-transfection cells were harvested and replated . Forty-eight hours after the first transfection , cells were co-transfected with siRNA ( as above ) and 200 ng of proviral plasmid or , in the case of GFP-reporter viruses , 200 ng of GagPol expression plasmid , 200 ng of corresponding retroviral reporter plasmid along with 100 ng of VSV-G expression plasmid using Lipofectamine2000 ( Invitrogen ) . Virions and cells were harvested 48 hours later . Cell lysates and pelleted virions or VLPs ( recovered by centrifugation through 20% sucrose ) were separated on NuPage Novex 4–12% Bis-Tris Mini Gels ( Invitrogen ) . Proteins were blotted onto nitrocellulose membranes . Thereafter , the blots were probed with primary antibodies and a corresponding peroxidase conjugated secondary antibody and were developed with WestPico chemiluminescent detection reagents ( Pierce ) . Alternatively , blots were probed with antibodies as above , followed by secondary antibodies conjugated to IRDye 800CW or IRDye 680 . Fluorescent signals were detected and quantitated using Odyssey ( LI-COR Biosciences ) . 293T cells stably expressing DsRed-Clathrin light chain were seeded on 3 . 5-cm , glass-bottomed dishes coated with poly-L-Lysine ( Mattek ) . The following day , they were transfected with a plasmid expressing AP180C , using Lipofectamine 2000 . Cells were fixed 24 hrs later and observed by deconvolution microscopy using an Olympus IX70-based Deltavision microscopy suite as described previously [48] . To generate fluorescent VLPs for microscopic analysis , 293T cells stably expressing DsRed-clathrin LC were cotransfected with plasmids expressing codon-optimized HIV-1 Gag , or Gag-Pol along with a plasmid expressing codon optimized Gag-GFP at a ratio of 4∶1 or 8∶1 . Forty-eight hours post transfection the culture supernatants were pelleted through a 20% sucrose cushion . The resulting VLPs were resuspended in PBS , 0 . 22 µm filtered , and diluted 1∶1 with PBS containing 3% paraformaldehyde . They were fixed overnight onto poly-L-Lysine coated glass bottom dishes and , after permeabilization with 1% Trition X-100 and 0 . 5% SDS , the VLPs were stained with anti-HIV capsid antibody , and analyzed via microscopy ( Deltavision , Applied Precision ) . Alternatively , proviral plasmids pNL4-3 and pNL4-3 MA-YFP were transfected in a 1∶1 ratio into 293T expressing DsRed-clathrin LC , subjected to the above protocol but were visualized without immunostaining . For scanning electronic microscopy studies , 293T cells were transfected with plasmids expressing codon optimized pCR3 . 1/SIVmac Gag-IRES-eGFP cassette and inspected by fluorescent and scanning electron microscopy using a Hitachi S4700 field emission SEM ( University of Missouri Electron Microscopy Core Facility ) . For transmission electron microscopy , samples were fixed using paraformaldehyde and glutaraldehyde , postfixed using osmium tetroxide and then dehydrated and embedded in Embed-812 ( EMS ) . Following polymerization , approximately 65 nm sections were cut with an ultramicrotome and mounted on copper grids . Sections were stained with uranyl acetate followed by lead citrate and imaged using a FEI TECNAI G2 transmission electron microscope .
|
The assembly and maturation of infectious retroviruses is driven by two viral proteins , Gag and Pol . Additionally , a number of cellular proteins are found in retrovirus particles , many of which lack a known functional role . One such protein is clathrin , which normally mediates several physiological processes in cells and was previously thought to be only passively incorporated into virions . In this study we show that clathrin is actively , specifically and abundantly incorporated into retrovirus particles . In several cases , retroviral proteins encode peptide motifs that mimic those found in cellular adaptor proteins that are responsible for clathrin recruitment . The range of retroviruses into which clathrin is packaged includes human and simian immunodeficiency viruses as well as other murine and simian retroviruses . Manipulations that prevented clathrin incorporation into virions also caused a variety of defects in the genesis of infectious retroviruses , including viral protein destabilization , inhibition of particle assembly and release , and reduction in virion infectiousness . The precise nature of the defect varied according to which particular retrovirus was examined . Overall these studies suggest that clathrin is frequently employed by retroviruses to facilitate the accurate assembly of infectious virions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology"
] |
2011
|
Clathrin Facilitates the Morphogenesis of Retrovirus Particles
|
Neurons encode information in sequences of spikes , which are triggered when their membrane potential crosses a threshold . In vivo , the spiking threshold displays large variability suggesting that threshold dynamics have a profound influence on how the combined input of a neuron is encoded in the spiking . Threshold variability could be explained by adaptation to the membrane potential . However , it could also be the case that most threshold variability reflects noise and processes other than threshold adaptation . Here , we investigated threshold variation in auditory neurons responses recorded in vivo in barn owls . We found that spike threshold is quantitatively predicted by a model in which the threshold adapts , tracking the membrane potential at a short timescale . As a result , in these neurons , slow voltage fluctuations do not contribute to spiking because they are filtered by threshold adaptation . More importantly , these neurons can only respond to input spikes arriving together on a millisecond timescale . These results demonstrate that fast adaptation to the membrane potential captures spike threshold variability in vivo .
Neurons encode information in sequences of stereotypical action potentials or spikes . Spikes are all-or-none voltage deflections triggered when the membrane potential of a neuron crosses a threshold . In vivo , the spiking threshold , as measured as the voltage at the upstroke of spikes , varies with firing history and input properties . This phenomenon has been widely observed in the central nervous system , e . g . visual cortex [1] , [2] , auditory midbrain [3] , hippocampus [4] , somatosensory cortex [5] . It has been proposed that threshold variability measured in vivo reflects an adaptation of the spike threshold to the membrane potential , due to the inactivation of sodium channels [6]–[8] or the activation of potassium channels [9] , [10] . Threshold adaptation would have a profound influence on how the combined input of a neuron is encoded in the spiking output [5] , [6] , [11]–[14] , such as enhancing coincidence detection [1] , [2] , improving feature selectivity [5] and temporal coding in sensory neurons [15] . However , previous studies in vivo only reported correlations suggestive of threshold adaptation . Other authors have suggested that threshold variability observed in vivo could also reflect measurement artifacts because spikes are initiated at the axon initial segment but measured at the soma [16] . Threshold variability could also be due to channel noise [17] , slow changes in excitability [18] or modulation by axonal synapses [19] . More generally the voltage measured at the upstroke of spikes may be a poor estimate of the actual criterion for spiking ( which could depend on unobserved quantities ) . The goal of this study was to determine whether threshold variability observed in vivo is mainly due to threshold adaptation to the membrane potential , or to one of the alternative hypotheses . Unfortunately , this question cannot be entirely addressed in vitro , where inputs are better controlled . First , there are potential sources of threshold variability in vivo that do not exist in vitro; in particular , noise and synaptic inputs to the initial segment . Second , properties of Na channels are likely to be different in vivo . Indeed , Na channels can be modulated in various ways , including their peak conductance and both the time constant and voltage-dependence of inactivation [20] . Therefore , results in vitro may not readily extend to in vivo conditions . In this work we studied the dynamics of the spiking threshold in neurons of the barn owl's external nucleus of the inferior colliculus ( ICx ) in vivo . While the spatial tuning [21] , [22] and the underlying computations in ICx neurons have been investigated [23]–[25] , previous studies have shown wide variation in spiking threshold over the stimulus duration [3] . To understand this variability , we fitted a mechanistic model of spike threshold adaptation that generalizes a model based on sodium-channel inactivation [26] to intracellular recordings in vivo . The model is used to test whether it is possible to accurately predict spiking from the membrane potential history . If threshold variability is due to noise , then this prediction should fail; if it is due to factors other than adaptation ( for example phosphorylation of Na channels , or GABA inputs onto the initial segment ) , then the parameter values of the fitted model should depend on the stimulus . The model was able to predict spikes with high accuracy and to account for most observed variance in measured threshold . In addition , it allowed us to estimate the threshold at all times , including between spikes . We found that the spike threshold tracks the membrane potential at a shorter time scale than the membrane time constant . The “effective signal” for spike initiation is then best defined as the difference between threshold and membrane potential . Fast threshold adaptation has two major functional consequences: 1 ) the effective signal is less variable than the membrane potential , because low frequency components of the input are filtered out; 2 ) the neuron can only respond to inputs with dynamics faster than the adaptation timescale , an order of magnitude lower than the membrane time constant . These findings show that most threshold variability observed in vivo in these neurons can be explained by fast adaptation to the membrane potential .
Neurons of the barn owl's ICx are selective to sound direction [21] , by combining tuning to interaural time ( ITDs ) and intensity differences ( IIDs ) [22] , [23] . We recorded the membrane potential ( ) of ICx neurons in vivo while presenting 100 ms broadband sounds ( white noises filtered between 0 . 5 and 10 kHz ) through earphones , varying either ITD or IID ( Fig . 1a ) . Spike thresholds ( Fig . 1b ) were measured using the derivative , a procedure known to produce reliable estimates [27] . At spike initiation the derivative increases abruptly and the precise value of the criterion makes little difference to the estimated voltage ( Fig . 1c ) . The spike threshold was highly variable , spanning a range of about 8 mV ( σ = 3 . 1±1 . 1 mV ) . In fact , the distribution of spike thresholds was so large that it overlapped the distribution induced by the input ( Fig . 1d ) . As previously observed in this [3] and other areas [1] , [5] , [13] , spike threshold was positively correlated ( r = 0 . 75±0 . 1 , regression slope = 0 . 43±0 . 1 ) with the average preceding spikes ( Fig . 1e ) , and negatively correlated ( r = 0 . 61±0 . 1 , regression slope = −0 . 49±0 . 3 ms ) with the rate of depolarization before spikes ( Fig . 1f ) . We did not observe significant correlation between inter-spike interval ( ISI ) and spike threshold ( r = 0 . 2±0 . 2 , regression slope = −0 . 01±0 . 03 mV/ms , Fig . 1g ) , as was observed in a few other studies [4] , indicating that spike refractoriness is shorter than typical ISIs . These observations suggest that the spike threshold adapts to the dynamics . However , what we called “spike threshold” above is in fact only a measurement of the voltage at the upstroke of spikes . It could be that the relevant criterion for spiking is a quantity ( or set of quantities ) other than somatic voltage , and that the voltage at the upstroke of spikes is correlated with history but has no causal relationship therewith . To address this issue and demonstrate that the spiking criterion ( and not just the measured voltage at the upstroke of spikes ) adapts to the membrane potential , we used a generalization of a model of threshold adaptation based on sodium channel inactivation [8] to predict the occurrence and timing of spikes . Our goal was to predict spike trains , not the voltage at the upstroke of spikes . Although the model was derived from properties of sodium channels , we used it here as a phenomenological model of threshold adaptation , which may also be consistent with other intrinsic mechanisms ( see Discussion ) . This model consists of a differential equation describing the adaptation of the threshold to a function of , , with a time constant :A spike is predicted to occur when . More generally , the spike threshold is defined as the voltage value at which the neuron would spike if its membrane potential were instantaneously brought above it . Thus it is a threshold in the sense of an explicit spiking criterion , unlike the empirical measurements . The function , called the steady-state threshold , represents the value of the spike threshold when is clamped at a fixed value . This can be considered as a general first-order model of threshold adaptation . Theory based on the properties of sodium channels predicts that the steady-state threshold is constant below the half-inactivation voltage , and increases approximately linearly above it [8] . However , threshold adaptation can also result from activation of voltage-gated potassium channels [9] . Therefore , to be more general , we did not impose a constant threshold below . Instead , we used a smooth function with a different slope below and above the critical voltage , and a parameterized curvature ( Fig . 2a ) . Parameters characterizing the two slopes , the connecting point and the curvature were constrained by the data . Some threshold adaptation models also include an explicit effect of spikes on threshold [26] , [28] ( the threshold increases after each spike ) , but it did not appear useful in our case , as we observed no correlation between ISI and spike threshold . A straightforward approach would be to use this model to predict the value of spike threshold measured in the intracellular traces . However , as argued above , the measured somatic voltage at the upstroke of spikes may not correspond to the spike threshold , in the sense of a criterion for triggering a spike . For example , it has been argued that the relevant criterion should in fact be the voltage value at spike onset in the axon initial segment ( AIS ) , where spikes are initiated [16] , [29] . Even if the somatic voltage at the upstroke of spikes truly corresponded to the spike threshold , there would still be a methodological issue with optimizing the threshold model to predict that voltage . Indeed a trivial solution to the fitting problem is the threshold model defined by and ms: the “spike threshold” always equals the membrane potential , in particular at the upstroke of spikes . To avoid these problems , we instead used the threshold model to predict the occurrence of spikes and their precise timing based only on . The trivial solution mentioned above is a poor predictor of spikes since it would predict too many spikes . The voltage trace was thus passed through the model equation to produce a dynamic spike threshold ( Fig . 2b ) . Theory predicts that a spike should be produced when the voltage trace crosses threshold . The model can fail by producing spikes at the wrong time or by producing extra spikes . To account for both types of errors , we defined a stringent coincidence window ( δ = 84 µs ) and calculated the proportion of coincident spikes in both the recorded and predicted spike trains . We used the gamma factor , a normalized coincidence measure that has been used in a number of studies [30]–[32] . We optimized the model parameters to maximize on a given recording , and the model performance was then tested on different recordings in the same cell . We first checked that this optimization strategy was correct on different neuron models with an explicit adaptive threshold with a time constant of 3–5 ms ( Fig . 2c–e; see Materials and Methods ) . The first model had a fixed threshold ( Fig . 2c ) , the second an adaptive threshold with rectified-linear characteristics ( only adapts above ; Fig . 2d ) and the third a threshold that adapted linearly in the entire voltage range ( Fig . 2e ) . Note that there is a constant bias in the predicted threshold , corresponding to the sharpness of spike initiation in these neuron models ( i . e . , spikes start slightly above the threshold value because sodium channels open gradually ) [8] . Apart from this bias , both the steady-state threshold curve and the adaptation time constant were correctly estimated by the optimization procedure ( Fig . 2c–e , bottom curves ) . We also confirmed that the fitting procedure yielded expected results when the threshold time constant was an order of magnitude shorter than the membrane time constant ( Fig . 2f ) . Finally , we checked that the resulting parameters did not depend on the input statistics , by running the optimization procedure with input currents of different means and standard deviations on models with a short threshold time constant ( Fig . 2g ) and a short input autocorrelation time constant ( Fig . 2h ) . We then applied the fitting procedure to a biophysically detailed neuron model , in which spikes are initiated in the AIS and Na channel densities in the axon were measured with immunochemistry [7] . In this multicompartmental model , the value of the spike threshold measured at the soma can be accurately predicted from the value of ionic channel gating variables at the AIS [26] . We stimulated the model with fluctuating current , and we observed that there was a linear dependence between the measured value of the spike threshold and the logarithm of the Na inactivation variable h at the AIS ( Fig . 3a; slope −3 . 2 mV; r = −0 . 98 ) . We then fitted the threshold model to the voltage response of the model ( Fig . 3b ) . After optimization , we observed that the time-varying threshold of the fitted model closely tracked the spike threshold estimated from ionic channel gating variables ( which were hidden to the fitting procedure ) . At the spike times predicted by the fitted model , the corresponding predicted threshold was very close to the actual measured threshold ( Fig . 3c ) . The steady-state threshold curve matched the curve calculated from the Na inactivation function [8] , especially near the spike initiation region ( Fig . 3d ) . In the multicompartmental model , the time constant of Na inactivation is voltage-dependent , unlike in our simple threshold model ( Fig . 3e , green ) . However , the fitted threshold time constant matched the value of the inactivation time constant in the spike initiation region ( −50 to −40 mV; Fig . 3e , red ) . Finally , we found that the value of the Na inactivation variable h at the AIS could be estimated between spikes from the value of the spike threshold in the fitted model ( Fig . 3f; slope −0 . 22 mV−1; r = −0 . 97 ) . Those results show that our method can successfully predict the spike threshold and characterize the sodium inactivation properties at the AIS of a complex multicompartmental neuron model containing an axon and an extended dendritic tree . Taken together , these results show that our optimization strategy can indeed accurately characterize the properties of spike threshold adaptation . We then applied this technique on our recordings , where spike times were accurately predicted , with few false alarms and typical rectified-linear curves for the estimated steady-state threshold ( Fig . 4a ) . To emphasize the fact that we predict the threshold for spike initiation , and not simply the voltage at the upstroke of spikes , we show the voltage trace vs . dynamic threshold in Fig . 4b , where it can be seen that a spike is produced as soon as the identity line is crossed . Also , there are no crossings of the identity line between spikes . The absence of threshold crossings between spikes can be related to the sharpness of spike initiation [33] , due to the compartmentalization of spike initiation in the AIS [34] . This observation means that the value of in the model predicts whether a spike is initiated , rather than simply predicting the somatic voltage at spike onset . This implies that it is indeed possible to predict spikes using only the membrane potential at the soma , even though spikes are initiated in the axon initial segment . If threshold variability is due to ionic channel properties , then threshold parameters should depend on the cell and not on the experimental condition . On the contrary , if threshold variability were due to other factors such as synaptic input onto the axonal initial segment , we would expect these parameters values to be variable across conditions . Therefore we optimized the model parameters separately on each cell and sound-stimulation condition ( e . g . , one condition is varying the ITD with a fixed IID ) to check for stimulus dependency . As an additional check of robustness , we divided the entire set of recordings into subsets ( 2–8 ) with different ranges , and optimized the model parameters separately in each set . We then compared the parameter values obtained for the same cell but different recordings . We found little variation in the results across conditions in the same cell ( Fig . 5 and Figs . 6a , b , c ) . Consistent with theoretical predictions for Na channel inactivation [8] , the steady-state threshold was near-constant at low voltages and increased linearly with slope near 1 at high voltages ( Fig . 5 ) . The fact that the steady-threshold curve does not cross the diagonal ( Fig . 5 , dashed lines ) is consistent with large threshold variability [8] . In the entire set of cells ( n = 21 ) , we consistently found that the slope of the steady-state threshold curve was small at voltage smaller than ( Figs . 6a , mean ) and near 1 above ( Figs . 6b , mean ) , which is consistent with predictions based on sodium channel inactivation [8] . The mean critical voltage , , was −59±6 mV and the minimum threshold was = −61±6 mV . Although there is some uncertainty about absolute voltage in intracellular recordings with sharp electrodes , the values are within the range of half-inactivation voltages of Na channels [35] . The curvature of the steady-state threshold is determined by the model parameter ka = 7±2 mV , which is in the range of measured Na activation slopes [35] . Finally , the threshold-adaptation time constant was ( Fig . 6c ) . Although this may seem small , time constants tend to be short in the barn owl's auditory brainstem , which is specialized for fast temporal processing [6] , [13] , as also seen in the timescale of spikes in Fig . 1 ( see Discussion ) . In addition to the fact that threshold-adaptation time constants were similar across cells and recording conditions , the precise value of the model time constant was also important for predicting spikes . Fixing the time constant to a shorter or larger value than the optimal one significantly degraded the fitting quality ( Fig . 7 ) . A consistent observation is that above , the steady-state threshold always lies just a few mV above ( Fig . 5 , distance between solid curve and dashed line ) . Thus the condition for triggering spikes is not exceeding a fixed threshold , but rather a fast depolarization of a few mV . This property implies that when the neuron is slowly depolarized , it does not spike because the threshold increases at the same time . It can contribute in making the neuron respond with a single spike at the onset of a current step – but not necessarily because the reset may introduce fast variations in membrane potential . Electrophysiological properties of IC neurons are not known in the barn owl , but onset electrophysiological behavior has been observed in IC neurons of rodents , although not all neurons [36] . In the chick , neurons in Nucleus Laminaris , which project to IC , also respond to current steps by firing a single spike [37] . The optimized parameters varied across cells but not across stimulation protocols or ranges in the same cell ( Fig . 6a–c , blue error bars ) . The average distance between steady-state threshold curves obtained in the same cell for different conditions was an order of magnitude smaller than the average distance between steady-state threshold curves and the diagonal ( Fig . 6d ) . These findings indicate that there is little threshold adaptation acting on a slow timescale in these neurons . We then tested the optimized threshold models on recordings in the same cell that were not used for fitting the parameters , whether a different stimulation protocol or a different range , and we found that the models produced few false alarms ( 6 . 8% , Fig . 6e ) . Finally we tested whether at spike times , the value of the spike threshold variable in the model corresponded to the measured somatic voltage at the upstroke of spikes . We found that the model threshold could account for 89% of experimentally measured “spike threshold” variance on average ( Fig . 6f ) . This means that the measured somatic voltage at spike onset does in fact correspond to the spike threshold , in the sense of a criterion for triggering a spike . In addition , since this value can be accurately predicted by our model , this result implies that the measured spike threshold is in fact determined by the dynamics at the soma , rather than noise or external factors . Finally , it also implies that if there was stimulus-specific adaptation in these neurons as found in rats [38] , it did not act on spike threshold , since we did not include such phenomena in the model . We finally turn to the functional implications of spike-threshold adaptation . Since the spike threshold adapts to , any voltage fluctuations that are slower than threshold adaptation should not have an impact on output spiking . This is captured by the concept of ‘effective signal’ ( ES ) illustrated in Fig . 8 . The ES is the difference between the and the dynamic spike threshold ( Fig . 8a ) . A spike is produced when the ES exceeds a fixed threshold ( 0 mV ) . Therefore , the dynamics with threshold adaptation is equivalent to the ES dynamics with a fixed threshold . In the ES , voltage variability is greatly reduced , dropping from σ = 4 . 4 mV in the to σ = 1 . 6 mV in the ES for this recording ( Fig . 8b ) . This occurs because slow voltage fluctuations are filtered out by threshold adaptation . This becomes clear when we compute the autocorrelation of the voltage traces ( Fig . 8c ) . We found that the half-height width ( HHW ) of the autocorrelation was 4 . 6 ms . This value corresponds to a membrane time constant of 3 . 3 ms for white noise input ( HHW/ ( 2 . log 2 ) ) ; in this case it may also reflect the timescale of synaptic currents . In contrast , the HHW of the autocorrelation of the ES is only 0 . 5 ms , which is in the order of magnitude of the threshold time constant . Because of threshold adaptation , postsynaptic potentials ( PSPs ) are effectively shortened . Specifically , the exponential decay of PSPs disappears from the ES , making the effective PSP shorter ( Fig . 8d ) . In all cells , voltage variability is greatly reduced by threshold adaptation: from about σ = 5 . 1±1 . 0 mV in the to σ = 2 . 2±0 . 8 mV in the ES ( Fig . 8e ) . Since threshold adaptation has little effect on the peak size of a fast PSP ( Fig . 8d ) , the ratio between PSP size and background voltage variability is effectively increased . In the same way , HHW is reduced from 4 . 7±1 . 4 ms to 1 . 7±1 . 5 ms ( Fig . 8f ) . This means that the integration time window of these neurons is about three times shorter than inferred from the membrane potential alone , making the neuron sensitive to input coincidences at a millisecond timescale .
In vivo , the spiking threshold is highly variable , typically spanning a range of about 10 mV . This phenomenon has been observed in many areas of the nervous system: visual cortex [1] , [2] , auditory midbrain [3] , hippocampus [4] , somatosensory cortex [5] , neocortex [10] , and prefrontal cortex [7] . Spike threshold has been found positively correlated with average membrane potential [2] , [7] and inversely correlated with the preceding rate of depolarization [1] , [2] , [5] , [12] . These observations are consistent with the hypothesis that the spike threshold adapts to the membrane potential , because of inactivation of sodium channels [1] , [2] , [5] , [8] , [26] and/or activation of low-voltage activated potassium channels ( Kv1 ) [9] , [10] , [26] . However , these observations could also result in part or entirely from one or several of the following alternative causes: Empirical support for threshold adaptation and for these alternative hypotheses comes from in vitro studies , and therefore it is not known whether and to what extent they may explain in vivo observations . Indeed , there are potential sources of threshold variability in vivo that do not exist in vitro ( noise , synaptic inputs to the initial segment ) , and Na channels can be modulated in various ways , including their peak conductance and both the time constant and voltage-dependence of inactivation [20] . To distinguish between these hypotheses , we applied a predictive approach to in vivo recordings , which does not rely on measuring the somatic voltage at spike onset . Instead , the threshold model is evaluated on the basis of its ability to predict the occurrence of spikes from the previous membrane potential . This approach addresses the concern that criteria based on spike shape at the soma to measure “threshold” might inaccurately assess the actual criterion for triggering a spike . In these data , the threshold model accounted for 89% of measured spike threshold variance . Therefore , most observed variability was due to deterministic processes , which ruled out hypothesis ( a ) . It confirms theoretical considerations showing that ion channel stochasticity should imply a positive correlation between rate of depolarization and spike threshold , contrary to our and previous experimental observations [8] . According to hypothesis ( b ) , spikes are actually initiated at a fixed voltage threshold , but it appears variable because it is not measured at the initiation site ( in the axon ) . Our results discard this possibility because the threshold model is optimized to predict the occurrence of spikes , not the measured voltage at spike onset at the soma . It indeed predicts the occurrence and precise timing of spikes very accurately , and with very few false alarms . Therefore , the variability of measured somatic voltage at spike onset did reflect the variability of spike threshold in these recordings ( see also Fig . 4b ) . It confirms theoretical considerations showing that variability due to hypothesis ( b ) should also imply a positive correlation between rate of depolarization and spike threshold [8] . To address hypothesis ( c ) , we fitted the threshold model in the same cell but in different experimental conditions ( either different ranges of or different stimulus conditions ) . If threshold variability were due to other processes that are not directly determined by ( e . g . synaptic input to the AIS or intrinsic plasticity ) , then we would expect the fitting process to yield different parameters values depending on context . In contrast , parameter values of the model were very robust across different conditions for the same cell , and variable between cells . These results make hypothesis ( c ) implausible in our recordings . On the basis of single-compartment biophysical models , it has been proposed that the total synaptic conductance may also modulate the spike threshold in a logarithmic way , by opposing the Na current [26] . Our results would only be consistent with this hypothesis if total synaptic conductance were constant in all conditions ( all stimuli and all mean ) . Although it seems unlikely , we cannot entirely rule out this possibility . Recent theoretical analysis taking into account the axonal initiation of spikes indicates that the total synaptic conductance at the soma should have negligible impact on spike threshold because spike initiation is compartmentalized [34] ( i . e . , only channels expressed at the AIS can directly modulate the spike threshold ) . Therefore , our results discard all the alternative hypotheses mentioned above , and demonstrate that threshold variability reflects deterministic adaptation of the spike threshold to the somatic membrane potential . Adaptation of spike threshold points to voltage-gated ion channels expressed in the AIS . Spike initiation is due to Na channels of the Nav1 . 6 subtype expressed in the distal part of the AIS [40] . These channels are partially inactivated at rest , and therefore voltage changes should substantially modulate the spike threshold by changing the proportion of available channels for spike initiation . The threshold model used in this study derives from a theoretical analysis of the biophysical properties of Na channels [8] , [26] . This analysis accurately predicted the spike threshold in a multicompartmental model of a cortical neuron with measured channel densities in the AIS [26] . The theory predicts that 1 ) the spike threshold is constant in the hyperpolarized range because Na channels are not inactivated , 2 ) the spike threshold follows the membrane potential in the depolarized range because activation and inactivation curves have similar slopes [8] , 3 ) the transition between the two regimes occurs at around half-inactivation voltage . Our results confirm these predictions . The time constant of threshold adaptation may seem surprisingly low , about 250 µs . In Hodgkin-Huxley models , this adaptation time constant reflects the time constant of the underlying ionic channel mechanism ( inactivation of Na channels or activation of K channels ) . Na channel inactivation time constants for subthreshold voltages are generally found to be on the order of the ms in vitro , in the cortex and hippocampus [41] . However , there is evidence that the time constant of inactivation can be modulated [20] , and that it depends on functional constraints , such as energetic efficiency [41] . In the electric organ of the electric fish , it has been found the inactivation time constants of Na and K channels are co-regulated , and correlate with the frequency of electrical discharges [42] . In this particular context , Na inactivation time constant varied between 500 µs and 3 ms ( Fig . 7 ) . Therefore it seems possible that this time constant is also short in a nucleus involved in the processing of sounds with frequencies of several kHz . The fact that spikes are shorter than 500 µs ( Fig . 1b ) in our recordings is an indication that it may indeed be the case . Low-voltage activated potassium channels ( Kv1 ) are also expressed at high density in the AIS [43] , [44] . Activation of Kv1 channels by depolarization can also raise the threshold , and therefore , Kv1 channels can produce threshold adaptation with similar qualitative properties as Na channel inactivation [26] . A few in vitro studies show that pharmacologically blocking Kv1 channels can abolish threshold variability [9] . This could be because Kv1 channels are responsible for threshold adaptation , or because blocking these channels lowers the spike threshold so that spikes are initiated before Na channels can inactivate ( this happens in Fig . 3 if threshold curves are shifted down and intersect the diagonal ) . It is possible that the residual threshold adaptation seen in the hyperpolarized range ( Fig . 5 ) is due to Kv1 channels . Clearly distinguishing between Na inactivation and Kv1 activation might require dual recordings in the soma and AIS , sodium imaging or pharmacological manipulations . Our results were obtained with in vivo intracellular recordings in the barn owl's inferior colliculus , and one may wonder to what extent they may generalize to other areas . The detailed statistics of threshold variability are similar to previous observations in cortical neurons [1] , [5] , both qualitatively and quantitatively , except perhaps for the depolarization rates , which tend to be larger in our recordings ( Fig . 1f ) . The mechanisms of spike initiation are also widely shared across the nervous system [40] , [44] . Therefore it is reasonable to expect that our findings are generally valid . However , it is likely that the time constant of threshold adaptation ( which was only a few hundred of microseconds in our study ) is larger in other areas . Indeed auditory neurons in subcortical areas are known to display faster kinetics than in other areas , not only in the barn owl but also in mammals [45] , [46] . Another likely difference is that in some in vivo studies , spike threshold was found to strongly depend on the time since the previous spike [4] , [28] . This is not contradictory with the model , which displays this phenomenon when the adaptation time constant is larger than the typical interspike-interval . Finally , in pyramidal cells of the cortex , and also in hippocampus neurons , the AIS is targeted by GABAergic neurons named Chandelier cells [19] . Their action could potentially modulate the spike threshold depending on local network activity ( for instance on the phase relative to theta oscillations in the hippocampus [47] ) , in a way that is not determined by the cell's Vm at the soma ( hypothesis ( c ) ) . Our results show that threshold variability is mainly due to deterministic features of the input , rather than noise . Given the extent of this variability ( more than 10 mV ) , this finding has major implications for the input-output properties of neurons . It implies that the relevant time-dependent variable is not so much the membrane potential , but rather its distance to a dynamic threshold , which we called the “effective signal” . Our method allowed us to estimate the spike threshold not only at spike times but also continuously between spikes , and thus to estimate the effective signal . We found that a large part of the variability appearing in the voltage trace vanishes in the effective signal , because slow variations of the membrane potential are filtered out by threshold adaptation , leaving only variations that are faster than threshold adaptation . Secondly , we found that the effective signal varies on a shorter timescale than the membrane potential . It implies that the temporal window of integration is shorter than expected from the membrane time constant , and closer to the threshold time constant . These findings confirm previous suggestions that threshold variability enhances coincidence detection properties of cortical neurons [1] , [5] , and corroborate observations that spikes tend to be preceded by fast depolarizations in cortical neurons in vivo [48] . Taken together , these findings demonstrate the causal link between membrane potential dynamics and spike threshold variability in vivo . In elucidating the deterministic nature of threshold , this work shows that threshold adaptation makes neurons selective to fast input variations and remarkably insensitive to slow ones .
The protocol #20110502 for this study followed the National Institutes for Health Guide for the Care and Use of Laboratory Animals and was approved by the Institutional Animal Care and Use Committee of California Institute of Technology . Data were obtained from in vivo intracellular recordings of 21 ICx neurons in 14 anesthetized adult barn owls , as described previously [24] , [24 , 49] . Sharp glass electrodes ( 40–80 MΩ ) filled with 2M potassium acetate were used for recording . All experiments were performed in a double-walled sound-attenuating chamber . Acoustic stimuli were digitally synthesized and delivered through earphones . Sound stimuli consisted of broadband-noise bursts ( 0 . 5 to 10 kHz , 100 ms in duration and 5 ms linear rise/fall times , 30 dB above threshold ) presented once per second . Earphone assemblies containing a speaker and a calibrated microphone were inserted into the ears and gaps were sealed with silicone material . The earphones were calibrated at the beginning of each experiment to correct for speaker irregularities . Intracellular recordings were stored at 24 kHz sampling rate . Measured spike threshold is defined as the voltage at the onset of action potentials . For each spike , the onset is defined as the first time preceding the peak when the first derivative crosses a fixed criterion , 25 mV/ms . On the phase plot ( Fig . 1c ) , it corresponds to a voltage value that is only crossed when a spike is produced . The precise value is not critical for model fitting because we predict the timing of spikes rather than the voltage at spike onset . Fig . 1e shows the mean computed in 5-ms window preceding a spike . Fig . 1d shows the rate of depolarization over 1 . 5 ms preceding a spike . The dynamic threshold depends only on , and is determined by [26]:Where is the time constant of the threshold dynamics . is the steady-state threshold ( Fig . 2a ) :where is the slope on the left side of the knee . The slope on the right side is . The curvature C ( Fig . 2a ) is indirectly determined by , , , , , , , were the parameters to optimize . A spike is produced when exceeds and is followed by a refractory period of 0 . 5 ms . If threshold modulation is due to sodium channel inactivation , the theoretical prediction [8] corresponds to . Given a trace and its corresponding spike onsets ( described above ) , we want to find the parameter values of the adaptive threshold model that maximize the similarity between predicted and recorded spike trains . This similarity is quantified using the gamma factor ( ) [30] , [31] , a normalized measure of coincidence between spike trains within a temporal window : is the mean firing rate of the experimental recording , is the number of coincidences between the predicted and recorded spike trains computed within a time window , and denote the number of spikes in the recorded and predicted spike train , respectively . is the expected number of coincidences generated by a Poisson process with rate . The first term in brackets is a normalization factor so that the maximum of is 1 . means that there are no more coincidences than expected by chance whereas means that the model prediction is perfect , at temporal resolution . To perform the optimization , an evolution algorithm ( CMAES ) [50] was implemented on Graphical Processing Units ( GPU ) [32] , [51] using the Playdoh optimization toolbox [52] . A spiking neuron model with an adaptive threshold was used to generate the membrane voltages of Fig . 2c–h . All other voltage traces are intracellular recordings . The model is based on the exponential integrate and fire [53] . is governed by a differential equation that includes a leak current and an exponential term describing sodium current activation at spike initiation:where is the membrane time constant , is the reversal potential of the leak current , characterizes the sharpness of the initiation . except in Fig . 2 f . , g . and h . where . is the membrane resistance . The membrane voltage diverges quickly once it exceeds the threshold , it is then reset to −70 mV , and a refractory period of 0 . 8 ms follows ( in practice , spikes are detected when ) . In Fig . 2c–f , the input current I is an Ornstein-Uhlenbeck process with mean 40 pA , standard deviation 120 pA , and time constant 3 ms . In Fig . 2g–h , the optimization is performed on a set of currents with mean between 20 and 200 pA and standard deviation between 50 and 400 pA , selecting those eliciting at least 20 spikes and a firing rate lower than 200 Hz . Current time constant was 3 ms in Fig . 2g and 0 . 5 ms in Fig . 2h . The exponential model accurately captures the dynamics of the sodium current near spike initiation [28] , while allowing sharp spike initiation . We used this , rather than a Hodgkin-Huxley model , because spike initiation is unrealistically shallow in a single-compartment Hodgkin-Huxley model and spike onsets are not well defined [33] , [34] . Multicompartmental models can display sharp spike initiation [7] but the threshold is not explicitly defined , a problem to test the predictive power of a threshold model . We assume that threshold dynamics are governed by the differential equation given in section “Adaptive threshold model” . The model has a constant threshold in Fig . 2c ( , , mV , mV and ms ) , rectified threshold in Fig . 2d ( same except mV ) , linear threshold in Fig . 2e ( , , mV , mV and ms ) , rectified threshold in Fig . 2f , but with fast threshold adaptation ( ms ) . In Fig . 3 , we used a biophysically detailed multicompartmental model of a cortical neuron based on immunochemistry measurements , in which spikes are initiated in the axonal initial segment [7] . It was stimulated at the soma with fluctuating current as described above , with mean 0 . 7 nA , standard deviation 0 . 2 nA and time constant 10 ms . The spike threshold is estimated from ionic channel gating variables as described in [26] ( Fig . 3b , green ) . All simulations except for the multicompartmental model were performed using the Brian simulator [54] with a sampling frequency of 42 kHz . The multicompartmental model was simulated with Neuron [55] . For each cell , the voltage traces were grouped in subsets . A subset is a set of traces sharing common conditions . The first type of condition used to characterize subsets is the binaural protocol used . For instance , the first subset can be the set of traces recorded when varying the ITD , another when varying IID , and another when varying average binaural intensity ( ABI ) . Depending on the cell , there were two or three recording protocols used , resulting in two or three subsets . The second type of condition is the mean during stimulation . For each cell , responses to all sounds are ordered by mean . Each subset is then constructed incrementally by adding consecutive traces until there are at least 120 spikes in the subset . This makes 2–8 subsets per cell . The prediction performance is quantified using two metrics . The false alarm rate ( FA ) , reported as a percentage , is defined as the number estimated spikes that are not coincident with recorded spikes divided by the total number of recorded spikes . The explained variance ( EV ) quantifies the prediction quality of the voltage at spike onset:withwhere is the voltage at spike onset in the recorded trace and is the predicted voltage at spike onset . These two metrics were always used on recordings not used for fitting the model ( different binaural protocol or different mean ) . For each cell , we calculate the average distance between steady-state functions obtained for different conditions ( Fig . 6d ) using the following formula:Where and are respectively the maximum and minimum sub-threshold voltages in the trace under consideration , and n is the number of conditions . For comparison , we also report the average distance to the diagonal :
|
Neurons spike when their membrane potential exceeds a threshold value , but this value has been shown to be variable in the same neuron recorded in vivo . This variability could reflect noise , or deterministic processes that make the threshold vary with the membrane potential . The second alternative would have important functional consequences . Here , we show that threshold variability is a genuine feature of neurons , which reflects adaptation to the membrane potential at a short timescale , with little contribution from noise . This demonstrates that a deterministic model can predict spikes based only on the membrane potential .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"neuroscience",
"neuroscience",
"biology",
"and",
"life",
"sciences",
"computational",
"biology"
] |
2014
|
Spike-Threshold Adaptation Predicted by Membrane Potential Dynamics In Vivo
|
Determining the total number of charged residues corresponding to a given value of net charge for peptides and proteins in gas phase is crucial for the interpretation of mass-spectrometry data , yet it is far from being understood . Here we show that a novel computational protocol based on force field and massive density functional calculations is able to reproduce the experimental facets of well investigated systems , such as angiotensin II , bradykinin , and tryptophan-cage . The protocol takes into account all of the possible protomers compatible with a given charge state . Our calculations predict that the low charge states are zwitterions , because the stabilization due to intramolecular hydrogen bonding and salt-bridges can compensate for the thermodynamic penalty deriving from deprotonation of acid residues . In contrast , high charge states may or may not be zwitterions because internal solvation might not compensate for the energy cost of charge separation .
Predicting the structural properties of proteins in the gas phase is crucial to interpret mass spectrometry data , yet this is far from being understood [1]–[10] . So far , it has been established that ( i ) compact structures acquire smaller net charges than unfolded ones [11]–[13] , ( ii ) secondary and tertiary structure elements play a crucial role for protein fragmentation [14]–[23] , and ( iii ) hydrogen bonds ( H-bonds ) and salt-bridges [3] , [24] , [25] may stabilize the structures . However , how desolvation impacts on structural facets of proteins [2] , [3] , [8] , [26]–[30] , peptides [3] , [31]–[35] and even single amino acids [36]–[45] is matter of a vivid debate . A key point is the presence of charge separation . Whilst amino acids exist mostly in their zwitterionic form in the aqueous solution [31] , [36] , [40] , [46] , conflicting assumptions and conclusions have been drawn for the same molecules in vacuo [47]–[50] . For peptides and proteins , the key issue of the charge state of ionizable groups , presumably different from that in solution , is even less clear [2] , [51]–[53] . One line of thought assumes neutral acidic functions for proteins analyzed in positive-ion mode ( i . e . , generating and detecting positively charged ions ) and neutral basic sites in negative-ion mode . In other words , the number of ionized groups is generally assumed to be equal to the net charge of the protein ion [54]–[56] . Electrostatic energy calculations based on this supposition fail to reproduce experimental values of apparent gas-phase basicity ( GPB ) for folded protein ions [57] . The GPB of a basic species B is defined as the negative of the free-energy change , , for the gas-phase protonation reactionIf B is the conjugate base of an acid AH , then , where GA is the gas-phase acidity of AH . Analogously , the proton affinity is defined as the negative of the protonation enthalpy , PA = − . In contrast , an increasing number of experimental [16] , [24] , [25] , [58]–[61] and theoretical [62] , [63] investigations carried out on peptides and small proteins indicate that zwitterionic states may survive in the absence of solvent if the structural features allow for adequate intramolecular solvation [64]–[67] . Recent ultraviolet photo-dissociation [16] and fluorescence [25] , [61] experiments indicate the presence of stabilizing salt-bridge motifs in small biomolecules . Salt bridges exist also in protonated , gas-phase serine dimers [24] and have been predicted for arginine dimers [63] , [68] , [69] . These interactions add to other stabilizing contributions such as hydrogen bonds [3] , [24] , [25] . Molecular dynamics ( MD ) simulations on a minimalistic lattice model of a zwitterionic system [1] turned out to reproduce the experimental observation that compact structures acquire smaller net charges than unfolded ones [11]–[13] . On the basis of these simulations , it has been also proposed that steric and electrostatic shielding of charged residues in compact conformations are the major factors responsible for this structural facet . Energy calculations [2] , [7] , [52] and measurements [51] , [70] on several well characterized proteins in their experimentally observed , most populated charge state suggest that the presence of zwitterions depends on the specific protein structure [2] , [51] . Deprotonated aspartic and glutamic residues persist in the most abundant , positively charged protomer of insulin , the C-terminal domain of the ribosomal protein L7/L12 and ubiquitin , but not in tryptophan-cage and lysozyme [2] . Prompted by the current lack of understanding of the charge state of protein ions in vacuo , here we have carried out an exhaustive energy analysis on three systems largely studied in the gas phase both experimentally [16] , [25] , [59] , [60] , [71]–[77] and theoretically [52] , [62] , [63] , [72] . These are the 8-residue peptide angiotensin II ( AN ) [74]–[76] and the 9-residue peptide bradykinin ( BK ) [16] , [59] , [62] , [63] , [71]–[73] , as well as the tryptophan-cage ( Trp-cage ) [16] , [25] , [52] , [60] , [77] protein . The latter is a 20-residue mini-protein with a well defined secondary and tertiary structure in aqueous solution at ambient conditions . It consists of an -helix and a compact hydrophobic core formed by a Trp side chain from the -helix , surrounded by several hydrophobic residues ( two prolines and one tyrosine ) [78] . A complete exploration of the protomer space ( i . e . , all of the possible charge configurations compatible with a given charge state ) of these biomolecules is performed coupling force field–based molecular dynamics and density functional theory ( DFT ) calculations . In contrast to previous computational studies [25] , [31] , [52] , [62] , [63] , [79] , [80] , all of the charge states generated by ionized and/or neutral basic ( R , K , H , Q , N-terminus ) and acidic groups ( E , D , C-terminus ) , featuring more than one protomer , are taken into account . A computational protocol apt to this task has been developed , allowing for an exhaustive exploration of the conformational space of each protomer . Based on such protocol , we suggest that low-charge states are likely zwitterions . In those cases , H-bonds and salt-bridges stabilize largely zwitterionic states , considerably reducing the differences in the apparent GPB between basic residues and the conjugated base of acidic residues . At high net charge , instead , non-zwitterion states are most likely .
The sequences of BK , AN and Trp-cage are RPPGFSPFR , DRVYIHPF , NLYIQWLKDGGPSSGRPPPS , respectively . For each system , the following protonation sites were considered: , , N- and C-term for BK; , , , N- and C-term for AN ( in the neutral state can be protonated either in or , both tautomers were considered ) ; , , , , and N- and C-term for Trp-cage . In the latter , protonation of was considered for the and charge states on the basis of experimental evidences [19] . BK and AN have no determined secondary structure and all of the calculations started with an all-trans backbone and side-chain conformation . Instead , the Trp-cage initial structure was obtained by a 20-ns MD simulation in aqueous solution at ambient conditions based on the NMR structure number 1 deposited in the protein data bank ( PDB code: 1L2Y ) [78] ( see Text S1 ) . The most probable protonation state in water [78] was chosen . For the chosen set of protonation sites , all of the charge states which feature more than one protomer were taken into account . For these charge states , all of the possible protomers were considered , for a total of 100 protomers ( see Tables 1 , 2 , and 3 ) . OPLS/AA force field-based [81] , [82] , constant-temperature MD calculations and geometry optimizations were carried out . The cutoff of electrostatics and van der Waals interactions was fixed at 0 . 7nm . In the MD simulations , the equations of motion were numerically integrated with a time step of 1 . 5 fs . All the hydrogen-bond lengths were kept fixed using the LINCS [83] algorithm . The temperature was controlled by the Nosé-Hoover thermostat [84] . The results of force field based MD simulations depend critically on the charge state used . Therefore , we performed a simulation for each protonation state . Specifically 8-ns MD simulations at high-temperature ( 700K for AN and BK , 350K for Trp-cage ) were performed for each protonation state . The chosen temperatures were selected after several careful tests . In particular , for Trp-cage , a temperature of 350K turns out to allow for an exhaustive sampling of the side chains conformations without disrupting , in the relatively short simulation time , the secondary structure . The resulting trajectories were split into 5-ps , non overlapping time windows . For each window , the geometry of the lowest-energy MD conformation was optimized by a conjugated gradient scheme up to 0 . 1 kJ/molÅ residual force on any atom . This simulated annealing-like procedure yielded for each protomer a large set of conformations . The geometry of structures within 100 kJ/mol ( 60 kJ/mol for Trp-cage ) from the lowest-energy force field conformer were refined at the ab initio level ( see Section “Identifying relevant protomers of a given charge state” ) . With this criterion , 60 conformers ( 35 for Trp-cage ) , were randomly selected from equally spaced energy windows , one from each window , and re-optimized at DFT/BLYP level of theory . The GROMACS [85] software package was used for all MD calculations . Quantum-chemical geometry optimizations were performed within the framework of DFT . The Becke exchange [86] and Lee-Yang-Parr [87] correlation functionals ( BLYP ) were used in a hybrid Gaussian and plane wave approach [88] . The wave function was optimized by using an orbital transformation technique [89] and analytic Goedecker-Teter-Hutter [90] , [91] pseudopotentials ( PP ) . The TZV2P Gaussian basis set was used for valence electrons of all atoms , while the auxiliary electron density was expanded in plane waves up to a cutoff of 280 Ry . The interaction between periodic images in the reciprocal space was removed according to the decoupling scheme presented in [92] . The calculations were carried out with the CP2K code [89] , [93] , [94] , which has been shown to be very efficient for these systems . The adopted DFT scheme was validated against more accurate ( and more expensive ) quantum chemistry methods . First , the relative energy of canonical and zwitterionic arginine conformers calculated with the present scheme agrees well with that obtained from all-electrons B3LYP , MP2 , and CCSD calculations ( see Table 2 in Text S1 ) . Second , all of the 14 protomers of AN with total charge underwent all-electrons , single-point energy evaluations at DFT/B3LYP level with the 6–311++G ( d , p ) basis set using the Gaussian03 code [95] ( Angiotensin II was chosen because it is the smallest of the three molecules studied and , in particular , the charge state was considered because it presents the largest set of protomers , and it is , therefore , a good benchmark case ) . These and the previous calculations provided the same energy ranking ( see Table 3 in Text S1 ) . A final concern for using DFT for non-covalent systems is the underestimation of dispersion forces [96] , [97] . This flaw of the current GGA functionals might influence the conformational energy , especially in the case of large molecular assemblies like those considered here . To quantify this error an estimate of the dispersion energy was performed for the DFT optimized structures using the OPLS/AA force field . The results of this calculation ( see Tables 4 , 5 and 6 in Text S1 ) indicate that the dispersion energy is not expected to change qualitatively the DFT energy ranking of protomers .
A standard procedure to identify the relevant protomers is currently lacking , even for peptides with more than a few amino-acids . On the one hand , the high complexity of the conformational space hampers an exhaustive search based on first-principle quantum chemistry ( such as DFT ) of the minimum-energy conformers . On the other hand , force field–based calculations [62] , [63] , [98] , [99] , or semiempirical quantum chemical methods [52] , may not be accurate enough . For instance , Merck molecular force field [100] energies have been shown to correlate poorly with those calculated at the DFT/B3LYP level [62] , [63] . In addition , the energies calculated by force fields do not take into account higher-order effects , which may play a role in our systems . DFT can , instead , take such effects into account . However , if the empirically calculated conformer is much higher in energy than another ( say with a greater than few hundreds of kJ/mol ) , it will be highly probable that the same ranking holds at the ab initio level ( see Figure 1 in Text S1 ) . Here , we seek such value by performing MD simulations based on the OPLS/AA , which offers the most complete set of base/conjugate acid pairs . The calculations on the three systems in vacuo provided several hundreds conformations , which then underwent DFT/BLYP [86] , [87] geometry optimizations . Such quantum chemical scheme is extremely efficient for large molecules , as those investigated here [101] , [102] . We found that less than 5% of the ab initio conformers located within 10kJ/mol from the energy minimum fall more than = 100kJ/mol ( 60kJ/mol for Trp-cage ) above the OPLS/AA minimum ( see Figure 2 in Text S1 ) . Exploiting this fact , we used the ensuing protocol to identify the lowest-energy minimum for each charge state for each peptide: ( i ) generation of conformers for all possible protomers by OPLS/AA MD and simulated annealing-like calculations; ( ii ) elimination of conformers whose energy is larger than from the absolute minimum; ( iii ) DFT/BLYP geometry optimization of the conformers within ; ( iv ) ranking of the conformers based on their DFT energies . Errors of this protocol are associated with ( i ) the accuracy of the DFT approach , ( ii ) limitations of sampling and ( iii ) absence of entropy contributions . This points are discussed in the following . We therefore conclude that the ranking obtained with our protocol provides a reliable identification of the most stable protomers . We discuss here the salient structural data of the low-energy protomers identified with the protocol outlined above for each system and for every charge state that features , according to our choice of ionizable residues , more than one protomer . More details and additional observations can be found in Text S1 . Structural data for each protomer of the considered charge states are reported in Tables 1 , 2 , and 3 . Our calculations suggest that most of the low-charge states are zwitterions , whilst high charge states might not . We now analyze the key factors for the stabilization of these two different states .
A computational protocol aimed at identifying the most stable species of angiotensin II , bradykinin , and tryptophan-cage has been developed and may be easily extended to other systems of similar size . The protocol provides results fully consistent with the experimental data . The results suggest that most of the low-charge states are zwitterions . Intramolecular interactions can stabilize zwitterionic states considerably , by reducing the differences in apparent GPB between basic residues and the conjugated base of acidic residues Based on a combined structural and energetic analysis , we suggest that salt-bridges provide a key energetic stabilization , in agreement with previous findings [3] , [38] , [51] , [63] , [116] . Indeed , the stabilization due to salt bridging might be such to reduce enormously the GPB of the biomolecules considered in the present study ( up to 900 kJ/mol ) . H-bonding also has an important role in promoting charge separation . As a result , networks are formed where two ( or more ) salt bridges are clustered together , whenever it is possible . Thus , we further corroborate the hypothesis that deprotonated carboxylate groups can be maintained in gas-phase peptide and protein ions produced by electrospray in positive-ion mode ( and , vice-versa , protonated basic groups in negative-ion mode ) [1] , [2] , [30] , [38] , [39] , [62] , [63] , [111] . On the other hand , the formation of zwitterionic species in high charge states requires the protonation of residues with progressively lower GPB , which is accompanied by a large thermodynamic penalty that might not be compensated by internal solvation .
|
In the last two decades mass spectrometry has given an impressive contribution to biochemistry , protein science , proteomics and structural biology . This technique offers powerful insights into protein structure and dynamics along with useful information on the role of solvent in protein stability as it is able to preserve non-covalent interactions and globular structures during the proteins' flight inside the mass spectrometer . Unfortunately , the key issue of the charge state of ionizable groups , presumably different from that in solution , has not been elucidated yet . So far conflicting assumptions and conclusions have been drawn by several groups . In the present work a very accurate structural and energetic analysis of the protonation state of two peptides and a small protein in the gas phase was performed . Results suggest that internal solvation can stabilize charge separation with the formation of zwitterionic states .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"biophysics/protein",
"chemistry",
"and",
"proteomics",
"biochemistry/protein",
"chemistry",
"biophysics/theory",
"and",
"simulation"
] |
2010
|
On the Zwitterionic Nature of Gas-Phase Peptides and Protein Ions
|
There is an urgent need for an improved diagnostic assay for typhoid fever . In this current study , we compared the recently developed TPTest ( Typhoid and Paratyphoid Test ) with the Widal test , blood culture , and two commonly used commercially available kits , Tubex and Typhidot . For analysis , we categorized 92 Bangladeshi patients with suspected enteric fever into four groups: S . Typhi bacteremic patients ( n = 28 ) ; patients with a fourfold change in Widal test from day 0 to convalescent period ( n = 7 ) ; patients with Widal titer ≥1:320 ( n = 13 ) at either acute or convalescent stage of disease; and patients suspected with enteric fever , but with a negative blood culture and Widal titer ( n = 44 ) . We also tested healthy endemic zone controls ( n = 20 ) and Bangladeshi patients with other febrile illnesses ( n = 15 ) . Sample size was based on convenience to facilitate preliminary analysis . Of 28 S . Typhi bacteremic patients , 28 ( 100% ) , 21 ( 75% ) and 18 ( 64% ) patients were positive by TPTest , Tubex and Typhidot , respectively . In healthy endemic zone controls , the TPTest method was negative in all , whereas Tubex and Typhidot were positive in 3 ( 15% ) and 5 ( 25% ) , respectively . We then estimated sensitivity and specificity of all diagnostic tests using Bayesian latent class modeling . The sensitivity of TPTest , Tubex and Typhidot were estimated at 96 . 0% ( 95% CI: 87 . 1%-99 . 8% ) , 60 . 2% ( 95% CI: 49 . 3%-71 . 2% ) , and 59 . 6% ( 95% CI: 50 . 1%-69 . 3% ) , respectively . Specificity was estimated at 96 . 6% ( 90 . 7%-99 . 2% ) for TPTest , 89 . 9% ( 79 . 6%-96 . 8% ) for Tubex , and 80 . 0% ( 67 . 7%-89 . 7% ) for Typhidot . These results suggest that the TPTest is highly sensitive and specific in diagnosing individuals with typhoid fever in a typhoid endemic setting , outperforming currently available and commonly used alternatives .
Typhoid and paratyphoid fevers ( collectively referred to as enteric fever ) are caused by Salmonella enterica serovar Typhi ( S . Typhi ) and serovar Paratyphi A , B , C [1] . Enteric fever causes high morbidity and mortality worldwide [2 , 3] . The majority of cases of enteric fever are caused by S . Typhi , with approximately 22 million cases of typhoid fever occurring annually , resulting in over 100 , 000 deaths globally each year [2 , 4] . In endemic areas , the burden is highest in young children . The incidence of typhoid fever in slum residents in Dhaka , Bangladesh is approximately 2 . 0 episodes/1000 persons per year with a higher incidence found in children aged < 5 years of age ( 10 . 5/1000 person per year ) than in older persons ( 0 . 9/1000 person per year ) [5] . Clinical diagnosis of often enteric fever is difficult due to its non-specific nature ( non-localizing febrile illness [6] . Unfortunately , accurate diagnosis of typhoid fever is problematic [7] . Several diagnostic approaches are commonly used , including microbiologic culturing of blood , and serologic assays such as the Widal or antigen-specific assays . All of these approaches suffer from poor sensitivity and/or poor specificity , especially in areas of the world endemic for enteric fever [7] . Many individuals with suspected typhoid fever are just empirically treated with antimicrobial agents , a clinical approach that drives development of microbial resistance , leaves individuals with other diagnoses without the correct treatment , and unnecessarily exposes patients to adverse effects of antibiotics . Although microbiologic culturing of bone marrow culture is considered a gold standard for diagnosing individuals with typhoid fever [8] , it is clinically impractical due to its invasive nature [8] . As such , there is a pressing need for an accurate diagnostic assay for typhoid fever . We have previously described development of an immunodiagnostic assay for enteric fever based on detection of anti-Salmonella enterica antibodies secreted by activated lymphocytes in the peripheral blood of acutely infected patients [9–11] . This assay , the TPTest ( Typhoid and Paratyphoid Test ) , measures S . Typhi membrane preparation ( MP ) -specific IgA responses in peripheral blood mononuclear cell culture secretions [10 , 11] . Initial pilot analyses in Bangladesh have demonstrated high sensitivities ( 100% ) and specificities ( 78%-97% ) , depending on the definition used , in identifying patients with enteric fever [10 , 11] . In this current study , we performed a direct comparison of the TPTest to blood culture , Widal analyses , and two common serologic assays , Tubex and Typhidot . We compared performance of the assays in 92 children and adults with suspected enteric fever in Dhaka Bangladesh , as well as in healthy endemic-zone controls , and a cohort of individuals febrile with non-typhoidal illnesses . Due to the absence of an acceptable gold standard , we used a Bayesian latent class modeling approach to estimate sensitivity and specificity of the various diagnostic approaches compared in this study [12 , 13] .
In total , we enrolled 127 participants in this study , including 92 participants who were clinically suspected of having enteric fever . Enrolment criteria for being a suspected enteric fever case included being 1–59 years of age , non-pregnant , having fever of ≥ 39°C for 3–7 days duration , and lacking an obvious alternative diagnosis . We collected 3–5 mL of venous blood for microbiologic culturing at the time of clinical presentation , and an additional 3 mL of blood for serologic analyses at clinical presentation and 7–28 days later . We also enrolled 20 healthy controls who also reside in Dhaka , Bangladesh , an area endemic for enteric fever , as well as 15 study participants who were febrile with non-typhoidal illness ( visceral leishmaniasis and tuberculosis ) . We performed microbiological culturing of venous blood using a BacT/Alert automated system , sub-culturing positive bottles on MacConkey agar , blood agar , and chocolate agar plates , and identifying colonies using standard biochemical tests and agglutination test with Salmonella-specific antisera ( Denka Seiken Co . , LTD , Tokyo , Japan ) . Antimicrobial susceptibility testing of isolates was performed using the disc diffusion method following a modified Kirby-Bauer technique [14] . We calculated the sensitivity and specificity with 95% confidence interval of the diagnostic methods using OpenEpi version 3 . We then estimated the sensitivity and specificity of each of the diagnostics using a Bayesian framework with latent class models . For prior information , we assumed that the sensitivity of culture was 40–80% ( 95% confidence interval ) and specificity was 100% [7] . We used prior estimates of sensitivity of Tubex and Typhidot from a recent meta-analysis; mean ( 69% ) and 95% confidence interval ( 45–85% ) were estimated for Tubex in that analysis , but not Typhidot [19] . We therefore assumed that a range of previously established estimates ( 56–84% ) reflected the 95% confidence interval for Typhidot . For prior estimates on specificity among all tests except culture , we used data from healthy controls and individuals with fever and confirmed alternative etiologies . We used a Gibbs sampler to sample from conditional parameter distributions using 100 , 000 Monte Carlo iterations; we discarded the first 50 , 000 , and used the remainder for inference . Multiple chains were run and results examined to ensure convergence . This study was approved by the research review and the ethical review committees of the icddr , b , and Institutional Review Board of the Massachusetts General Hospital . Written informed consent was obtained from all adult participants ≥18–59 years of age , and from parents or guardians of children 1–17 years of age .
Of the 92 study participants with suspected enteric fever , 48 ( 52% ) were male ( Table 1 ) . The median age was 6 years and 3 months , with a range of 1 to 46 years . The median temperature at enrollment was 39 . 2°C . The mean duration of fever before enrollment was 4 days . Reported symptoms and signs included headache ( 79% ) , abdominal pain ( 55% ) , constipation ( 30 ) , coated tongue ( 51% ) , diarrhea ( 13% ) , vomiting ( 9% ) , non-specific rash ( 9% ) , and rose spot ( 7% ) . S . Typhi was isolated from the peripheral blood of 28 ( 30% ) of the 92 suspected enteric fever cases patients; of these culture positive patients11 were male ( 39% ) and 17 were female ( 61% ) . Resistance to antimicrobial agents was common . Of the 28 S . Typhi isolates resistance to ampicillin , trimethoprim-sulfamethoxazole , chloramphenicol , ciprofloxacin , and nalidixic acid was 15 ( 54% ) , 15 ( 54% ) , 17 ( 61% ) , 9 ( 32% ) , and 27 ( 93% ) , respectively . All isolates remained susceptible to cefixime , ceftriaxone and gentamicin . No S . Paratyphi were isolated by blood culture during the course of this study . Among the 92 study participants with suspected enteric fever , seven were positive for a 4-fold change of Widal titre from acute to convalescent phase of illness , and 13 patients had a titre of ≥ 1:320 . None of the patients with a 4-fold change had a titre ≥ 1:320 , and no one with a positive Widal ( defined as either a 4-fold change or single titer ≥ 1:320 ) had a positive blood culture . Based on the results of blood culture and Widal testing , we divided our 92 study participants with suspected enteric fever into four groups for further analysis: Group I- patients with a positive blood culture for S . Typhi ( n = 28 ) ; Group II- patients with a fourfold change of Widal titre ( n = 7 ) ; Group III- patients with Widal titre ≥1:320 ( n = 13 ) ; and Group IV- suspected patients with a negative blood culture and a negative Widal test ( n = 44 ) . Interestingly , none of the 92 patients met criteria for more than one cohort . For comparison , we also created a Group V- healthy endemic zone controls ( n = 20 ) , and a Group VI- patients with other febrile illnesses ( visceral leishmaniasis and tuberculosis ) ( n = 15 ) . Out of 28 S . Typhi bacteremic patients , 28 ( 100% ) , 21 ( 75% ) and 18 ( 64% ) patients were positive by TPTest , Tubex and Typhidot , respectively ( Table 2; Fig 1; Fig 2 ) . For the four-fold change of Widal titre , the TPTest , Tubex and Typhidot were positive in 7 ( 100% ) , 6 ( 86% ) , 5 ( 71% ) , respectively . In case of patients with Widal titre ≥1:320 , the TPTest , Tubex and Typhidot were positive in 9 ( 69% ) , 5 ( 38% ) , 3 ( 23% ) , respectively . The TPTest method was negative for all healthy controls , whereas Tubex and Typhidot were positive in 3 ( 15% ) and 5 ( 25% ) , respectively . When considering patients with other febrile illnesses , the TPTest , Tubex and Typhidot were negative in 14 ( 93 . 3% ) , 14 ( 93 . 3% ) , and 13 ( 86 . 7% ) patients , respectively . Among 44 participants who were negative by blood culture and Widal test , but clinically diagnosed as enteric fever , 24 ( 55% ) , 9 ( 20% ) , and 15 ( 34% ) were positive by TPTest , Tubex and Typhidot , respectively . To establish upper and lower limits of sensitivity and specificity of the various assays , we used a number of definitions of patient cohorts . If we considered only blood culture-confirmed cases as positive and only patients with other febrile illnesses and healthy controls as negative for enteric fever , the estimated sensitivity and specificity was 100% ( 95% CI: 87 . 9%-100% ) and 97 . 1% ( 95% CI: 85 . 5%-99 . 5% ) for the TPTest; 75% ( 95% CI: 56 . 6%-87 . 3% ) and 88 . 6% ( 95% CI: 74 . 1%-95 . 5% ) for the Tubex assay; and 64 . 3% ( 95% CI: 45 . 8%-79 . 3% ) and 80% ( 95% CI: 64 . 1%-90% ) for the Typhidot assay , respectively . Alternatively , if we considered all blood culture-confirmed and Widal positive cases as positive , and patients with other febrile illnesses and healthy controls as negative for enteric fever , the sensitivity and specificity of the TPTest was 91 . 7% ( 95% CI: 80 . 5%-96 . 7% ) and 97 . 1% ( 95% CI: 85 . 5%-99 . 5% ) ; of Tubex , 66 . 7% ( 95% CI: 52 . 5%-78 . 3% ) and 88 . 6% ( 95% CI: 74 . 1%-95 . 5% ) ; and of Typhidot , 54 . 2% ( 95% CI: 40 . 3%-67 . 4% ) and 80% ( 95% CI: 64 . 1%-90% ) , respectively . If we consider all blood culture-confirmed and Widal positive cases as positive , and patients with other febrile illnesses , healthy controls , and individuals with suspected enteric fever but a negative blood culture and Widal as negative for enteric fever , the sensitivity and specificity of the TPTest became 91 . 7% ( 95% CI: 80 . 5%- 96 . 7% ) and 68 . 4% ( 95% CI: 57 . 5%-77 . 6% ) ; of Tubex , 66 . 7% ( 95% CI: 52 . 5%-78 . 3% ) and 83 . 5% ( 95% CI: 73 . 9%-90 . 1% ) ; and of Typhidot , 54 . 2% ( 95% CI: 40 . 3%-67 . 4% ) and 72 . 2% ( 95% CI: 61 . 4%-80 . 8% ) , respectively . In a latent class model in which sensitivity and specificity of all the diagnostic methods were estimated simultaneously , the sensitivity of TPTest was estimated at 96 . 0% ( 95% CI: 87 . 1%-99 . 8% ) , Tubex was 60 . 2% ( 95% CI: 49 . 3%-71 . 2% ) , and Typhidot was 59 . 6% ( 95% CI: 50 . 1%-69 . 3% ) ( Table 3 ) . Specificity was estimated at 96 . 6% ( 90 . 7%-99 . 2% ) for the TPTest , 89 . 9% ( 79 . 6%-96 . 8% ) for Tubex , and 80 . 0% ( 67 . 7%-89 . 7% ) for Typhidot . The Widal test had a low sensitivity when using a single high titer ( 14 . 9% ) or four-fold rise in titers ( 12 . 6% ) , but had excellent specificity ( 86 . 3% and 100 . 0% , respectively ) . Blood culture had intermediate sensitivity at 51 . 8% ( 41 . 2%-62 . 9% ) , but 100% specificity .
Enteric fever remains an important cause of morbidity and mortality worldwide , especially in infrastructure-limited countries including Bangladesh [2 , 3] . The clinical diagnosis of enteric fever is difficult because the symptoms and signs of enteric fever are similar to those of many other febrile illnesses [6] . Isolation of S . Typhi and S . Paratyphi from microbiologic culturing of bone marrow is considered a gold standard for the confirmation of enteric fever . However , the procedure is not clinically practical , especially when considering young children who bear a large component of the enteric fever burden in endemic areas [8] . Microbiologic culturing of blood is thus often used as an alternative diagnostic option when laboratory capacity is available . Unfortunately , the sensitivity of blood culturing is only 40–80% , reflecting in part the low burden of organisms in blood , and often prior use of antimicrobial agents [7] . Results require 2–7 days , but do provide a confirmed diagnosis and an antimicrobial susceptibility profile [20 , 21] . It is important to note that in our current study , antimicrobial resistance among the S . Typhi isolates was very common , with only one oral agent ( cefixime ) having uniform effective anti-microbial activity . These results underscore the need for an improved diagnostic assay , so that antimicrobial agents that still have activity can be appropriately targeted for use . The Widal assay has also been available for decades and can be performed using venous blood; unfortunately , the assay has low sensitivity and specificity , especially in endemic zones , and optimally requires comparison of samples drawn at the acute and convalescent stage of illness [22 , 23] . Nucleic acid amplification assays , such as PCR and LAMP-based assays show promise , but their utility has been hampered in clinical situations by low organisms load and presence of inhibitors in peripheral blood , reagent and equipment expense , and often lack of technical expertise in areas endemic for enteric fever , although such assays may have higher sensitivity than blood culture [7] . As such , to supplement these assays , many clinicians and public health studies rely on other commercially available assays , such as the Tubex and Typhidot immunodiagnostic assays . These assays have the advantage of not requiring extensive laboratory capacity or training , and can be performed on a small volume of venous blood collected at the acute stage of illness . These assays , however , have been limited by less than optimal sensitivity and specificity [19] . It is for this reason that we evaluated the TPTest under field conditions in Bangladesh , comparing results to the other standard used enteric fever diagnostic assays . Our results are quite encouraging . We analyzed performance using two approaches . First , by stratifying the patients by blood culture and Widal reactivity . Second , recognizing the limitations of the absence of a true gold standard , we used a Bayesian latent class modeling approach to estimate sensitivity and performance , comparing across the five imperfect tests . Using the first approach , the sensitivity and specificity of each test reflects the definition of a true positive or negative used in each analysis . However , in each analysis , the TPTest performed with higher sensitivity and specificity than the Tubex and Typhidot assays . It should also be noted that a number of the patients in our study with suspected enteric fever but who had a negative by blood culture and Widal assay may indeed still have had enteric fever , due to the low sensitivity of these assays . To address this limitation , we analyzed these data using latent class models , a form of statistical analysis that infers an unmeasured , true prevalence , based on the test characteristics and the results of multiple imperfect diagnostics . This approach thereby enables estimation of sensitivity and specificity of diagnostics in the absence of a gold standard . Using such a Bayesian latent class framework thus enabled us to estimate the sensitivity of the TPTest , Tubex , Typhidot and Widal assays among all febrile patients with suspected typhoid fever , rather than restricting our primary analysis to blood culture positivity , which only captures 40–80% of cases [7] . Under this approach , we estimated a very high sensitivity ( 96% ) of the TPTest , in comparison with blood culture , Tubex , Typhidot , and Widal tests [13] . These findings were consistent regardless of model specification . Specificity was high as well , but varied according to prior information about TPTest , Tubex and Typhidot utilized in the model . Specificity estimates for Tubex and Typhidot appeared to be consistent with estimates from a recent meta-analysis , serving as one robustness check [19] . As such , the TPTest appears quite promising . It appears to offer higher sensitivity and specificity than other commonly used assays . At present , the TPTest requires moderate laboratory capacity , results are now available in 24–48 hours , and antimicrobial susceptibility profiles are not provided . However , simplified adaptions are in development , including removal of the requirement of PBMC separation , and CO2 during incubation , and removal of the need for an ELISA read out [9–11] . Our study has a number of limitations . It did not include bone marrow aspiration as a gold standard; however , our inclusion of the latent class modeling does allow us to estimate performance using a range of imperfect assays . Our study did not include any patients with confirmed S . Paratyphi A bacteremia , so we cannot comment on test performance in such patients in this current study , although we have previously reported that the TPTest detects both S . Typhi and S . Paratyphi A infected patients [9–11] . Our study is also relatively small , limiting both its context , as well as our ability to assess the impact of age and other factors . It was also of note that in our study , there was no overlap of patients with a positive blood culture and a 4-fold change in Widal titer or high absolute titer . We hypothesize that this may in part reflect the relatively small sample size as well as the deficiencies of the Widal assay in this typhoid-endemic area . In addition , we modeled the diagnostic tests as conditionally independent , but in reality the serologic tests measure immune responses they may be correlated independently from disease . Because of uncertainty in these independent correlations between the multiple serologies , we did not incorporate them in the model , but further characterizing these correlations could improve accuracy of sensitivity and specificity estimates in the future . Despite these various limitations , our study describes a preliminary comparison of the most commonly used diagnostic assays for enteric fever with the evolving TPTest technology , and our results strongly support the continued development of this diagnostic approach .
|
We compared the performance of the recently developed TPTest ( Typhoid and Paratyphoid Test ) with the Widal test , blood culture , and commercially available kits , Tubex and Typhidot in diagnosing the patients with typhoid fever . There is no acceptable gold standard; therefore , we estimated the sensitivity and specificity of these diagnostic methods with Bayesian latent class modeling . We found that the sensitivity and specificity of the TPTest is higher than other commonly used methods for diagnosis of typhoid fever .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"microbial",
"cultures",
"pathogens",
"biological",
"cultures",
"microbiology",
"salmonella",
"typhi",
"bacterial",
"diseases",
"signs",
"and",
"symptoms",
"enterobacteriaceae",
"immunologic",
"techniques",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"widal",
"test",
"infectious",
"diseases",
"serology",
"medical",
"microbiology",
"microbial",
"pathogens",
"immunoassays",
"salmonella",
"hematology",
"typhoid",
"diagnostic",
"medicine",
"blood",
"anatomy",
"fevers",
"physiology",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2016
|
Comparison of the Performance of the TPTest, Tubex, Typhidot and Widal Immunodiagnostic Assays and Blood Cultures in Detecting Patients with Typhoid Fever in Bangladesh, Including Using a Bayesian Latent Class Modeling Approach
|
Chlamydia trachomatis , the causative agent of trachoma and sexually transmitted infections , employs a type III secretion ( T3S ) system to deliver effector proteins into host epithelial cells to establish a replicative vacuole . Aside from the phosphoprotein TARP , a Chlamydia effector that promotes actin re-arrangements , very few factors mediating bacterial entry and early inclusion establishment have been characterized . Like many T3S effectors , TARP requires a chaperone ( Slc1 ) for efficient translocation into host cells . In this study , we defined proteins that associate with Slc1 in invasive C . trachomatis elementary bodies ( EB ) by immunoprecipitation coupled with mass spectrometry . We identified Ct875 , a new Slc1 client protein and T3S effector , which we renamed TepP ( Translocated early phosphoprotein ) . We provide evidence that T3S effectors form large molecular weight complexes with Scl1 in vitro and that Slc1 enhances their T3S-dependent secretion in a heterologous Yersinia T3S system . We demonstrate that TepP is translocated early during bacterial entry into epithelial cells and is phosphorylated at tyrosine residues by host kinases . However , TepP phosphorylation occurs later than TARP , which together with the finding that Slc1 preferentially engages TARP in EBs leads us to postulate that these effectors are translocated into the host cell at different stages during C . trachomatis invasion . TepP co-immunoprecipitated with the scaffolding proteins CrkI-II during infection and Crk was recruited to EBs at entry sites where it remained associated with nascent inclusions . Importantly , C . trachomatis mutants lacking TepP failed to recruit CrkI-II to inclusions , providing genetic confirmation of a direct role for this effector in the recruitment of a host factor . Finally , endocervical epithelial cells infected with a tepP mutant showed altered expression of a subset of genes associated with innate immune responses . We propose a model wherein TepP acts downstream of TARP to recruit scaffolding proteins at entry sites to initiate and amplify signaling cascades important for the regulation of innate immune responses to Chlamydia .
The gram-negative bacterium Chlamydia trachomatis is the causative agent of trachoma , the leading cause of infectious blindness worldwide , and the major cause of bacterial sexually transmitted infections ( STI ) in the developed world [1] . Approximately 2 . 9 million cases of visual impairment and ∼1 . 17 million cases of blindness are attributed to ocular infections by C . trachomatis [2] . In the US , an estimated 2 . 8 million cases of Chlamydia STIs occur annually , imposing a significant burden on the public health system [3] . Chlamydiae have a biphasic life cycle that alternates between two unique developmental forms , the environmentally stable , infectious elementary body ( EB ) , and the replicative , but non-infectious , reticulate body ( RB ) [4] . Infection starts with the attachment of EBs to host cell membranes . After inducing its own internalization , C . trachomatis rapidly modifies its endocytic vacuole to avoid fusion with lysosomes [5] and migrates to a perinuclear region of the cell [6] , [7] where it undergoes a developmental transition to the RB form . Bacterial replication occurs within a membrane-bound vacuole called an inclusion , and mid-to-late in the infectious cycle , bacterial cell replication becomes asynchronous with RBs transitioning back to the EB form ( for reviews , see [4] , [8] ) . In late stages of infection , the inclusion occupies the bulk of the host cell cytoplasmic space and EBs are released to infect adjacent cells by cell lysis or extrusion of the inclusion [9] . Like many gram-negative bacterial pathogens , Chlamydia uses a type III secretion ( T3S ) system to deliver modulators ( effector proteins ) into their target host cell ( reviewed in [10] ) . These effectors interfere with diverse host cellular processes including signaling , cytoskeletal rearrangements , and vesicle trafficking to enhance bacterial entry , establish a replicative niche and evade innate immunity [11] . The temporal manner in which effectors are secreted is likely to be important for the proper manipulation of host cell functions . For instance , the Translocated Actin Recruiting Protein ( TARP ) is delivered into the host cytoplasm within 5 min of bacterial attachment [12] . TARP facilitates invasion by mediating actin re-arrangements through the direct nucleation of F-actin polymerization and the recruitment of Rac-specific guanine nucleotide exchange factors [13] , [14] . Another T3S effector , Ct694 , is delivered into host cells during Chlamydia entry where it engages the cytoskeletal organizing protein AHNAK [15] , although the timing of Ct694 translocation in relation to TARP is unknown . After entry , a new set of T3S effectors are synthesized and translocated to the inclusion membrane ( Inc proteins ) [16] where they mediate the recruitment of SNAREs , 14-3-3β , Rab proteins , signaling molecules and lipid transporters [17]–[20] . T3S effectors share a relatively unstructured , poorly conserved 20–30 amino acid signal sequence at their N-terminus that is often sufficient to mediate their broad secretion by T3S systems [21] . However , many effectors often require ancillary chaperone proteins for efficient translocation into target cells . Depending on the type of their client protein cargo , T3S chaperones can be divided into three classes ( reviewed in [22] ) . Class III chaperones prevent pre-oligomerization of needle components in the bacterial cytoplasm before secretion; class II chaperones stabilize translocators by binding their hydrophobic regions , and class I chaperones stabilize and/or enhance effector secretion . Depending on the number of effectors with which they can associate , class I chaperones are further divided into class IA and IB [23] . Class IA chaperones are specific for single effectors and the genes encoding the chaperone-effector pair are often co-transcribed or adjacent to each other on the bacterial genome . Class IB chaperones associate with more than one effector and the genes encoding these chaperones are often unlinked from that of its cognate cargo . Many common features are shared by class I chaperones , including small size ( 15∼20 kDa ) and acidic isoelectric points , a stable homodimer conformation , and ATP-independent chaperoning function [23] . Class I chaperones contain a basic core consisting of three α-helices and five β-sheets [24]; however , the degree of conservation at the amino acid level is limited , making the identification of new T3S chaperones difficult . From amino acid sequence analysis , C . trachomatis encodes at least six putative T3S chaperones: Slc1 ( Ct043 ) , Scc1 ( Ct088 ) , Scc2 ( Ct576 ) , Scc3 ( Ct862 ) , Ct274 and Scc4 ( Ct663 ) [25] , [26] . Several studies have validated their function as chaperones and have identified the substrates they engaged . For instance , Slc1 from C . trachomatis interacts with TARP and enhances its translocation/secretion in a heterologous Yersinia T3S system [27] , [28] . Scc1 and Scc4 from C . pneumoniae enhance the secretion of CopN whereas Scc3 inhibits CopN secretion [29] . Additional T3S chaperones have been defined functionally . For example , Mcsc ( Multiple Cargo Secretion Chaperone ) , which does not share any obvious sequence homology with known chaperones , was identified based on its ability to bind and stabilize the Inc proteins Cap1 and Ct618 [30] . A quantitative proteomic study indicated that the EB form of C . trachomatis is equipped with a complete set of T3 secretion components and is pre-packed with an abundant arsenal of putative T3S effectors and chaperones [31] . In addition to TARP , Ct694 and Ct695 , over 50 Chlamydia-specific hypothetical proteins lacking signal peptides ( ∼7% of the molar mass of the EB form ) were also identified . Since multiple effectors are likely translocated during invasion , a significant number of these Chlamydia-specific proteins may function as effectors with important roles early in infection . Interestingly , the EB form also contains a full complement of T3S chaperones , with Slc1 and Mcsc being the most abundant ( Fig . 1A ) [31] . Given their abundance and that genes encoding these chaperones are genetically unlinked from their effector protein cargos , we postulated that Slc1 and Mcsc may engage additional effectors in EBs . In this study , we immunoprecipited Slc1 and Mcsc from Chlamydia EB lysates and identified co-purifying proteins by mass spectrometry . In this manner , we determined that Slc1 is in complex with multiple T3S effectors and that co-expression of Slc1 enhanced the secretion of these effectors in Yersinia pestis . In the process , we identified and characterized a new effector , TepP ( Translocated early phosphoprotein – Ct875 ) , which is tyrosine-phosphorylated upon translocation into host cells . Given that the majority of TARP , but not TepP , is pre-complexed with Slc1 within EBs , and that TepP phosphorylation at Tyr residues occurs later with respect to TARP phosphorylation , we postulate that Slc1 helps impart a hierarchy to effector translocation during Chlamydia entry . We further show that phosphorylated TepP associates with host scaffolding proteins Crk-I and Crk-II and that the recruitment of Crk proteins to nascent inclusions is absent in cells infected with a Chlamydia mutant harboring a tepP null allele and is restored in complemented strain . In addition , endocervical epithelial cells infected with this mutant exhibited transcriptional changes in a subset of innate immunity-related genes . We propose a model wherein TepP acts downstream of TARP to recruit scaffolding proteins that initiate and amplify signaling cascades important for establishing a replicative niche for Chlamydia within the infected host .
A quantitative proteomics analysis of C . trachomatis indicated that around 2% of the EB total mass is composed of T3S chaperones , with Slc1 , Mcsc and Scc2 accounting for over 99% of all known T3S chaperones ( Fig . 1A ) [31] . Slc1 forms complexes with TARP and enhances TARP translocation into HeLa cells when both proteins are co-expressed in Yersinia enterocolitica [27] , [28] . In EBs , the majority of TARP is in complex with Slc1 , and both TARP and Slc1 are present at similar molar concentrations [31] . However , since slc1 is located ∼500 kb from tarP , it is unlikely that Slc1 constitutes a TARP-specific Class IA T3S chaperone . Furthermore , because putative T3S effectors in EBs are present at a 10 fold molar excess over the three major T3S chaperones [31] , we hypothesized that chaperones like Slc1 and Mcsc must bind multiple effectors . We further hypothesized that these chaperones bind their cargos in a hierarchical manner and thus impart coherence to effector secretion . To test this premise we determined the compendium of EB proteins that associate with Slc1 and Mcsc . We immunoprecipitated ( IP ) Slc1 and Mcsc under native conditions from EB lysates and identified all proteins that associated with each individual chaperone by LC-MS/MS ( liquid chromatography-tandem mass spectrometry ) . As previously reported [27] , [31] TARP was one of the major proteins that co-purified with Slc1 . In addition , two predicted T3S effectors , Ct694 and Ct695 , and two hypothetical proteins , Ct365 and Ct875 co-purified with Slc1 ( Fig . 1B ) . As described in the sections below , we renamed Ct875 as TepP to reflect the new functions defined in this study for this protein . CdsD , a T3S apparatus component , and proteins involved in metabolism were also specifically detected in the Slc1 IP samples ( Table S1 ) . Interestingly , both Mcsc and Slc1 mutually co-IP each other . Since these chaperones are not predicted to form heterodimers [27] , this result suggests that Slc1 and Mcsc homodimers may co-chaperone the same cargo , presumably Ct365 , which was detected in both IP samples . We validated the LC-MS/MS results by immunoblot analysis of the IP materials and immunodepleted flow-through samples with antibodies specific to Slc1 and the identified co-precipitating proteins ( Fig . 1C ) . We detected TARP , Ct694 , Mcsc , CdsD and Ct875/TepP in samples immunoprecipitated with anti-Slc1 antisera but not with pre-immune antisera . In contrast , MOMP , a very abundant EB protein , was not detected in the Slc1 IP samples , highlighting the specificity of the immunoisolations . As we had previously observed , immunodepletion of Slc1 from EB lysates led to a co-depletion of TARP [31] . Similarly , Ct694 was also substantially depleted , indicating that these two effectors within EBs are largely present in Sc1-containing complexes . The other co-purifying targets were not efficiently co-depleted which suggest that the complexes were either unstable or present in sub-stoichiometric amounts . We next tested if the major interactions identified for Slc1 , especially those with potential T3S effectors , could be recapitulated in vitro with recombinant proteins . The most abundant proteins that associated with Slc1 are Ct875/TepP , TARP , Ct694 , Ct695 and Ct365 . We expressed these proteins as GST-tagged fusion proteins together with untagged Slc1 in E . coli and tested whether Slc1 would co-purify on glutathione sepharose beads . Because , we could not express the predicted inclusion membrane protein Ct365 [32] in E . coli , this potential effector was not studied further ( data not shown ) . Pull downs of GST-tagged TARP , Ct694 , Ct695 and Ct875/TepP with glutathione beads led to the co-isolation of Slc1 . This binding was specific as neither GST nor GST-tagged Ct288 , another inclusion membrane protein [33] that is present in EBs , led to Slc1 co-purification ( Fig . 2A ) . T3S chaperones form stable complexes with a predicted 2∶1 stoichiometry of chaperone to effector protein [34] . To test if such larger complexes could form , we co-expressed in E . coli untagged Slc1 and hexahistidine-tagged full length TARP , Ct694 , Ct695 , or Ct875/TepP and isolated protein complexes on nickel resins . The bound material was eluted from the resin with imidazole and analyzed by gel filtration chromatography . All four proteins formed stable complexes with Slc1 , with apparent molecular weights above 150 kDa , which are larger than that expected for 2∶1 chaperone effector complexes ( Fig . 2B ) . Although gel filtration cannot always accurately predict molecular sizes for non-globular proteins [35] , given that TARP forms hexamers and that the oligomerization domain is distinct from the Slc1 binding domain [14] , [27] , we postulate that size discrepancies between Slc1/TARP and the other Slc1/effector complexes represent the formation of higher order oligomeric forms . Based on the gel filtration results we speculated that Slc1 , as reported for TARP [27] , would act as T3S chaperone and enhance the secretion of Ct694 , Ct695 , and Ct875/TepP . To test this premise , we reconstituted chaperone-assisted T3S in Y . pestis as previously described [29] . Effectors were expressed in Y . pestis as either untagged ( Ct694 and Ct875 ) or FLAG-tagged ( Ct695 ) proteins under the control of an arabinose-inducible promoter in the presence of Slc1 or Mcsc . Upon induction of T3S by calcium chelation , Ct694 and Ct695 were secreted into the supernatants as previously reported [15] . The secretion of Ct694 and Ct695 was enhanced two to three fold when Slc1 , but not Mcsc , was co-expressed , suggesting that Slc1 functions as a bona fide T3S chaperone specific for these proteins ( Fig . 2C ) . Consistent with the known behavior of most T3S chaperones , neither Slc1 nor Mcsc were secreted into supernatants ( data not shown ) . The role of Slc1 in TARP secretion in this system could not be determined as expression of full length TARP in the presence of Slc1 hindered Y . pestis T3S ( data not shown ) . Full length Ct875/TepP was also a target of secretion by the Y . pestis T3S system , even though Ct875/TepP is not predicted to have a T3S signal [21] , [31] , [36] , and its secretion was enhanced by Slc1 . Ct875/TepP is the most abundant Chlamydia-specific hypothetical protein pre-packed in EBs [31] . We speculated that this protein is secreted early during EB attachment to cells and that it plays a role in invasion and/or establishment of the nascent inclusion . Consistent with this , an analysis of Ct875/TepP localization by immunofluorescence microscopy indicated a close association of Ct875/TepP with EBs by 2 hours post infection ( hpi ) in HeLa cells ( Fig . 3A ) . However , the localization of Ct875/TepP was distinct from that of chlamydial LPS , suggesting translocation of Ct875/TepP from the EB at attachment sites ( Fig . 3A ) . This localization pattern is similar to what has been described for TARP and Ct694 [15] , and is clearly distinct from that of Slc1 which is mostly localized to the bacterial cytosol . Overall , these findings provide experimental evidence that Ct875/TepP is a new T3S effector translocated early during EB entry into epithelial cells . Approximately 5–6 distinct proteins are tyrosine-phosphorylated early upon infection with C . trachomatis [37] , [38] . Most of these proteins were originally assumed to be host proteins that were phosphorylated as a consequence of the activation of multiple kinase signaling pathways by Chlamydia , until TARP was identified as one of these major tyrosine-phosphorylated proteins [12] . Because Ct875/TepP has a molecular weight of 65 kDa , similar to one of the major tyrosine-phosphorylated protein detected within 15 minutes of C . trachomatis LGV-L2 infection [37] ( Fig . 3B ) , we considered the possibility that Ct875/TepP is also tyrosine-phosphorylated . Consistent with this , analysis by indirect immunofluorescence microscopy indicated that anti-Ct875/TepP and anti-phosphotyrosine signals co-localized around bacteria at 2 hpi ( Fig . 3C ) . Next , we determined that Ct875/TepP immunoprecipitated from infected HeLa cell lysates , but not from EBs lysates , was detected by anti-phosphotyrosine antibodies in immunoblots , suggesting that Ct875/TepP is tyrosine-phosphorylated upon association with host cells ( Fig . 3D ) . Phosphorylation of TepP provided us an opportunity to monitor the kinetics of its translocation into cells . EB infections were synchronized at 4°C and shifted to 37°C for various time intervals within a 1 h span . Infected samples were lysed , split in two and TARP and TepP were immunoprecipitated . Total TARP , TepP and tyrosine-phosphorylated protein in the IP material was monitored with specific antibodies . We observed that tyrosine phosphorylation of TepP was delayed with respect to TARP by ∼10 min ( Fig . 3E ) , suggesting either that kinases responsible for TepP phosphorylation are recruited with delayed kinetics to EB entry sites or that TARP translocation from EBs precedes TepP . However , TepP might not be the major ∼65–70 kDa phosphoprotein since the TepP band only partly overlapped with the major ∼65–70 kDa phosphotyrosine protein bands in dual labeling immunoblots ( Fig . 3F ) . We proceeded to map the phosphorylation sites on translocated Ct875/TepP by mass spectrometry . Infections were scaled up and Ct875/TepP was IP at 4 hpi in the presence of phosphatase inhibitors and processed for phospho-proteomics analysis . Four phospho-peptides were detected in the sample: two contained phosphorylated tyrosine ( Y43 and Y496 ) and two phosphorylated serine residues ( S410 and S415 ) ( Fig . S1 , Fig . 4A ) . Because the peptide spanning Y496 is part of a tandem repeat ( ASDYDLPR ) , it is unclear if one or both tyrosine residues ( Y496 and Y504 ) are phosphorylated . Overall , these results indicate that Ct875/TepP is phosphorylated at multiple residues during infection and thus we renamed this protein as TepP for translocated early phosphoprotein . Protein tyrosine phosphorylation plays an important role in signal transduction by mediating the recruitment of proteins containing Src homology 2 ( SH2 ) and/or phosphotyrosine binding ( PTB ) domains to target proteins [39] . Analysis of the TepP phosphopeptides using NetPhorest ( http://netphorest . info/ ) [40] indicated the presence of a consensus pYxxP binding site ( ASDYDLPR ) for the scaffolding protein Crk [41] . Crk is an adaptor protein that mediates phosphorylation-mediated regulation of cytoskeletal dynamics , cell adhesion and migration , phagocytosis and tumorogenesis [42] . Mammals express Crk-I and Crk-II , two alternatively spliced forms of CRK , and the Crk-like protein , Crk-L . Crk proteins contain SH2 and SH3 domains , which mediate binding to phosphorylated tyrosine residues and proline-rich domains , respectively [42] . To test if TepP translocated from EBs associated with Crk proteins , we performed a time course of infection and immunoprecipitated TepP in the presence of phosphatase inhibitors . Chlamydia infection did not alter the steady state levels of CrkI-II ( Fig . S2 ) ; whereas increasing amounts of CrkI and CrkII co-IP with TepP as the infection progressed ( Fig . 4B ) , indicating that these proteins are part of a complex . By immunofluorescence microscopy , we further observed endogenous Crk recruited to bacteria as early as 1 hpi and that this association continued past 8 hours , a point at which nascent inclusions have migrated to the microtubule organizing center ( Fig . 4C ) . To test if the association of TepP with Crk is direct , we next examined if recombinant TepP could bind to Crk in vitro . We found that the binding between TepP and GST-Crk was significantly enhanced by in vitro phosphorylation of TepP and reduced after phosphatase treatment ( Fig . S3 , Text S1 ) , suggesting that TepP phosphorylation is important for the recruitment of Crk . Together , these data are consistent with a model wherein TepP , after being secreted into host cells , is tyrosine-phosphorylated and recruits Crk to nascent inclusions , presumably through interactions with its SH2 domain . Although Chlamydia spp . lacks a system for performing targeted gene disruptions , chemically-derived mutants can be readily generated and loss-of-function mutations can be identified by whole genome sequencing [43] . Using such an approach , our group has recently generated and sequenced a comprehensive collection of ethyl methyl sulfonate ( EMS ) -derived mutants of C . trachomatis LGV-L2 ( unpublished results ) . We identified 6 strains with mutations in TepP , including strain CTL2-M062 , bearing a G to A transversion that led to a premature stop codon at amino acid 103 ( W103* ) ( Fig . 5A ) . By immunoblot analysis , EBs from CTL2-M062 lacked any detectable TepP , suggesting that the truncated form of TepP was either unstable or not properly expressed ( Fig . 5B ) . Interestingly , even though TepP is not the predominant ∼65–70 kDa tyrosine-phosphorylated protein observed early during C . trachomatis infection , these phosphorylated species were no longer detected when cells were infected with the TepP-deficient CTL2-M062 mutant ( Fig . 5C ) ; whereas the levels of the phospho-TARP band ( ∼150 kDa ) were unaltered . These results indicate that TepP may induce the phosphorylation of additional host and or secreted bacterial proteins at the early stages of infection . Next , we determined whether Crk association with nascent inclusions was dependent on TepP . HeLa cells were infected with either wild type or the TepP-deficient C . trachomatis mutant and the co-localization of bacteria with Crk was assessed by immunofluorescence microscopy . Crk was no longer associated with intracellular CTL2-M062 by 8 hpi ( Fig . 5D ) . The original TepP deficient isolate carried an additional 11 EMS-induced single nucleotide variants ( SNVs ) ( Fig . 5A; Table S3 ) . To exclude the possibility that these SNVs contributed to the lack of Crk recruitment to nascent inclusions , we generated recombinant strains between the TepP deficient isolate ( generated in a RifR background ) and a spectinomycin ( SpcR ) LGV-L2 strain by co-infection of cells as previously described [43] . Double drug resistant recombinants were plaque purified and screened by allele-specific PCR to determine the relative segregation of mutations present in the original TepP-deficient isolate . A total of 200 recombinants were screened and four co-isogenic strains were identified where various assortment of mutations were present ( Fig . S5 ) . The marked reduction of the ∼65–70 kDa signal in phosphotyrosine immunoblots was only observed in the strain lacking TepP ( Fig . 5E ) . Similarly , Crk recruitment to nascent inclusions was only impaired in strains encoding the TepPW103* variant ( Fig . 5F ) , providing further genetic support to the notion that TepP is directly responsible for the recruitment of Crk to inclusions . Finally , to provide final genetic confirmation of the role of TepP in mediating these events , we took advantage of recent developments in DNA transformation in C . trachomatis [44] . We transformed the recombinant TepPW103* strain with a C . trachomatis-E . coli shuttle plasmid expressing the tepP gene under the control of its endogenous promoter or the empty vector control and selected for stable transformants in the presence of penicillin . The transformed strains expressed TepP to a similar level as wild type strains ( Fig . 6A ) . An immunoblot analysis of infected cells showed the restoration of the ∼65–70 kDa phosphotyrosine immunosignals in TepP complemented strains but not in strains transformed with the empty vector ( Fig . 6B ) . In addition , Crk recruitment to nascent Chlamydia inclusions was restored ( Fig . 6C ) . Both results , the restoration of protein phosphorylation and Crk recruitment to nascent inclusions , were observed when infecting A2EN cells , a newly derived endocervical epithelial cell line [45] ( data not shown ) . Taken together , our results confirm that TepP is the major contributor , either directly or indirectly , to the tyrosine-phosphorylation of multiple proteins early during C . trachomatis infection and the subsequent recruitment of Crk to nascent inclusions . Although the CTL2-M062 had a ∼15 fold growth defect compared to the wild type parental strain , we could not unambiguously link this growth defect to strains bearing the tepP mutant allele ( data not shown ) . To address whether TepP played a role in promoting C . trachomatis replication or infectivity , we compared the ability of the tepP mutants transformed with the tepP expressing plasmid or plasmid alone to generate infectious progeny . We observed no differences in the generation of EBs among these strains in either HeLa cells ( Fig . S6D and Text S1 ) or in A2EN cells ( data not shown ) . Similarly , silencing of Crk , a major TepP-binding partner , with siRNAs in HeLa cells did not impact C . trachomatis replication ( Fig . S6A–C and Text S1 ) . These observations suggest that TepP is not essential for C . trachomatis replication in cell culture models of infection , although we cannot rule out a role for TepP in promoting invasion or intracellular survival in other cell types or in the context of intact tissues and an active immune system . However , given the observation that TepP recruits Crk , a scaffolding protein important in cell signaling , we hypothesized that the presence of TepP should lead to defined transcriptional responses by the infected cell . We compared the global transcriptional profile of mock infected A2EN cells or A2EN cells infected with the tepP mutant and its complemented counterpart for 4 h . A microarray analysis revealed 33 genes displaying greater than 1 . 5 fold changes in gene expression levels ( Fig . 7A and Table S4 ) . A Gene Ontology ( GO ) analysis of these differentially expressed genes revealed an enrichment for genes with immunity-related functions ( data not shown ) . We next validated by quantitative-PCR , the expression of five of these genes: IL-6 and CXCL3 , and MAP3k8 , IFIT1 and IFIT2 , whose transcription decrease and increase , respectively , depending on the presence of TepP ( Fig . 7B ) . Interestingly , the fold-change for IFIT1 and IFIT2 changed from 2 fold to more than 10 fold by 8 hpi; whereas the level of fold-change for MAP3k8 did not ( Fig . 7C ) . These data suggests that one of the functions of TepP is to modulate gene expression in the early stages of infection , presumably to impact the type and magnitude of the ensuing innate immune response .
Approximately 5–10% of the C . trachomatis genome encodes for proteins with putative T3S signals [21] , [36] . These potential effector proteins are presumably translocated into host epithelial cells at various stages of infection to mediate epithelial cell invasion , establishment of a protected replicative vacuole , and evasion of innate immune responses ( reviewed in [11] ) . Because Chlamydia cannot be readily manipulated with molecular genetic tools , most approaches to identify effectors have been indirect and relied on heterologous expression systems [46] . To date , only two effectors that are secreted in the early stage of infection , TARP and Ct694 , have been experimentally validated in C . trachomatis [12] , [15] . Given the complexity of interactions with the host cell during Chlamydia entry and nascent inclusion development , we expected additional effectors to be secreted upon the association of the EB with its target epithelial cell and that most of these effectors are pre-loaded in EBs . Indeed , a quantitative analysis of the C . trachomatis EB proteome [31] suggested the presence of at least 20 abundant hypothetical proteins with putative T3S signals . EBs are also pre-loaded with most of the predicted T3S chaperones encoded by the Chlamydia genome . However , putative T3S effectors are in molar excess to that of available T3S chaperones , which led us to speculate that these chaperones may engage multiple effectors . The two most abundant T3S chaperones in EBs are the TARP chaperone Slc1 [27] , [31] and Mcsc , which engages at least two inclusion membrane proteins during the RB stage [30] . Because the genes encoding Slc1 and Mcsc are unlinked from that of their effectors , we considered the possibility that these chaperones engage multiple effectors in EBs . To test this premise we decided to identify the set of proteins that stably associate with Slc1 and Mcsc at the EB stage by IP coupled to mass spectrometry . We only found a limited number of interactions for Mcsc in EBs . In contrast , Slc1 engaged at least four new substrates of the T3S system: Ct365 , Ct694 , Ct695 and Ct875/TepP , a protein not previously thought to harbor a T3S signal based on early prediction algorithms [21] , [36] . Consistent with our findings , Mota's group recently showed that Slc1 can interact with Ct694 and Ct695 in vitro and enhances their secretion in Y . enterocolitica [28] . In addition to their traditional role in escorting cargo for secretion , T3S chaperones can regulate additional cellular functions . For instance , SicA , a Salmonella T3S chaperone for the effector SipA , directly interacts with the transcriptional activator InvF , and this complex activates the expression of T3S genes [47] . Similarly , SycD ( LcrH ) , the chaperone for the translocator protein YopD , functions with YopD to represses Yop synthesis in Yersinia pseudotuberculosis [48] . In Chlamydia , the chaperone Scc4 ( Ct663 ) negatively regulates σ66–dependent transcription by directly interacting with both σ66 and β subunits of RNA polymerase [49] . At least one T3S chaperone , the Shigella Spa15 protein , is secreted and plays a role in preventing apoptosis of the infected host cell [50] . Slc1 is unlikely to be a T3S cargo as it was not secreted by Yersinia ( data not shown ) and Slc1 localization was restricted to EBs early in infection ( Fig . 3A ) . Nevertheless , we identified several additional proteins that specifically co-IP with Slc1 , including the late transcription unit B protein ( ltuB , Ct080 ) , a serine protease ( HtrA ) , a putative aminopeptidase ( PepA ) and a transaldolase ( TalB ) . While not experimentally confirmed , it is unlikely that any of these proteins constitute T3S effectors as computational programs aimed at identifying T3S signals give these proteins very low confidence for T3S ( www . effectors . org ) . At present , it is unclear whether these interactions are direct . However , one could envision scenarios wherein Slc1 could interact with non-effector proteins to signal the successful entry into a host cell after it is no longer physical bound to an effector . Such a mechanism might be especially important for an intracellular pathogen like Chlamydia , where physical engagement of the T3S apparatus may need to be coupled to rapid metabolic adaptations and post-translational modifications that drive successful colonization of the target cell . T3S chaperones enhance the secretion of its bound cargo by stabilizing effectors in the bacterial cytoplasm and maintaining them in a secretion competent state [51] . In addition , T3S chaperones can also prioritize the secretion of effectors . For instance , YopE , a Yersinia T3S effector , lacking a SycE chaperone binding site , is secreted in the absence of other T3S effectors . However , YopE secretion is severely impaired when other effectors are allowed to compete with it , indicating that additional secretion signals are either unmasked or conferred by the T3S chaperone [52] . Similarly , CesT from enteropathogenic Escherichia coli ( EPEC ) , binds to 9∼10 T3S effectors but one of the cargos , Tir , is secreted first and Tir secretion is important for secretion of the remaining effectors [53] . Indeed , a real-time analysis of CesT-dependent traslocation revealed a distinct order in the translocation of EPEC effectors and the effector-chaperone interaction is suggested to be one of the factors influencing translocation efficiency [54] . Based on these findings , we hypothesized that T3S chaperones would play a prominent role in establishing a hierarchy to the secretion of effectors by Chlamydia EBs . The most abundant T3S chaperone in EBs is Slc1 , which , like CesT , engages multiple effectors . TARP , one of Slc1's cargos [27] , is secreted within 5 min upon EB attachment to epithelial cells [12] . Interestingly , the majority of TARP in EBs is found pre-complexed with Slc1 ( Fig . 1C ) [31] , implying that this pre-engagement with its chaperone could prime TARP for rapid secretion . Similarly , a significant proportion of Ct694 within EBs is pre-complexed with Slc1 ( Fig . 1C ) , suggesting that Ct694 may be also secreted very early during infection , possibly at the same time as TARP . In contrast , only a minor portion of TepP was present in complexes with Slc1 in EBs , even though TepP is more abundant than TARP and Ct694 [31] . Given these observations , we considered a model wherein T3S substrates pre-bound by Slc1 in EBs , such as TARP and Ct694 , will be delivered first , followed by TepP and potentially other effectors which do not exist as pre-formed effector-chaperone complexes in EBs . Consistent with this model , TepP was tyrosine-phosphorylated later than TARP ( Fig . 3E ) . Although we cannot exclude the possibility that TepP-specific tyrosine kinases are recruited later than those that phosphorylate TARP , our preliminary results indicate that TepP is phosphorylated by kinases that also target TARP ( data not shown ) , making this possibility less likely . Overall , the delayed phosphorylation of TepP , coupled with the relative abundance of Slc1-TARP as compared to Slc1-TepP complexes within EBs lead us to propose that Slc1 imparts a hierarchy to the translocation of effectors during Chlamydia invasion ( Fig . 8 ) . Our findings indicate that TepP is a novel C . trachomatis effector that is targeted for tyrosine phosphorylation upon delivery into epithelial cells . This protein is highly conserved in all serovars of C . trachomatis , including ocular , genital and LGV strains . It shares 49% identity with TC_0268 , ortholog in C . muridarum and less than 25% identity with potential orthologs in other Chlamydiae spp ( http://www . uniprot . org/blast ) . TepP is phosphorylated at both tyrosine and serine residues ( Fig . 4A ) , with a phosphotyrosine residue mapping to the peptide ASDYDLPR , which is repeated in tandem between amino acids 496–504 . This phosphorylation site matches the pYxxP consensus binding site for the host adaptor proteins Crk-I and Crk-II [41] . Indeed , we found that both Crk proteins co-IP with endogenous TepP during the early stages of Chlamydia infection ( Fig . 4B ) . This interaction is likely to be direct and dependent on tyrosine phosphorylation since recombinant TepP only interacts with GST-CrK upon in vitro phosphorylation ( Fig . S3 ) . Consistent with these observations , CrkI-II associated with EBs at entry sites and nascent inclusions ( Fig . 4C ) . Importantly , the lack of CrkI-II association with nascent inclusions in a tepP null mutant strain and restoration after complementation provided genetic confirmation that TepP is required for Crk recruitment . CrkI-II are well-characterized scaffolding proteins that organize cytoskeletal rearrangement and signal transduction events [42] . The SH2 domain of Crk interacts with tyrosine-phosphorylated proteins . The SH3 domain , in turn , can interact with multiple proteins including guanine nucleotide exchange factors ( GEFs ) such as C3G , Sos1 and Dock180 , and c-Abl and PI3K ( p85 subunit ) , a protein and lipid kinase ( reviewed in [42] ) . The manipulation of Crk function by bacterial effectors is not unprecedented . In Helicobactor pylori , CagA , a Type IV secretion effector , is tyrosine-phosphorylated upon translocation and interacts with Crk [55] . This interaction then triggers downstream signaling events of Crk , including SoS1/H-Ras/Raf1 , C3G/Rap1/B-Raf and Dock180/ELMO pathway , resulting in CagA-specific cell responses such as epithelial cell scattering and cell-cell dissociation [55] . In Pseudomonas aeruginosa , exoenzyme T ( ExoT ) ADP-ribosylates CrkI and CrkII thus preventing the binding of the Crk SH2 domain to focal adhesion complex proteins , paxillin and p130cas [56] . This leads to uncoupling of integrin signaling , actin depolymerization , and potentially contributes to the anti-phagocytic activity of this opportunistic pathogen . The function of Crk in C . trachomatis infection is less clear . Because the Crk-binding partners Sos1 and Dock180 activate Rac1 [57] , [58] and Rac1 activation is partially required for C . trachomatis invasion [59] , it is possible that TepP , in addition to TARP [13] , contributes to Rac1 activation at bacteria entry sites through its recruitment of Crk . However , since TepP is not essential for bacterial entry or establishment of the replicative vacuole within epithelial cells ( Fig . S6D ) , we see a limited role for this effector in the activation of these pathways . Similarly , RNAi-mediated silencing of CrkI-II in HeLa cells did not affect the ability of bacteria to enter and replicate in these cells ( Fig . S6A–C ) . These findings are in contrast to observations made in Drosophila S2 cells where Crk was identified as a potential host factor important for C . muridarum growth [60] , and suggest possible differences in cell lines or Chlamydia strains used . At this stage we do not know the full compendium of proteins recruited to nascent inclusions via TepP and Crk-mediated protein scaffolding . However , it is clear from the number of proteins that are tyrosine-phosphorylated in a TepP-dependent manner that TepP contributes to multiple signaling events during Chlamydia invasion . Since signaling events often lead to changes in gene transcription , we performed microarray analysis to compare the global transcriptional response of A2EN cells to infection with the tepP mutant versus its complemented counterpart . Interestingly , many immune-related genes showed a TepP-dependent activation . Among them , the most striking changes were observed for IFIT1 and IFIT2 with a fold-change more than 10 fold by 8 hpi ( Fig . 7C ) . IFIT1 and IFIT2 belong to a family of interferon-induced protein with tetratricopeptide repeats ( IFITs ) that have a well-established role in host anti-viral defense . IFIT proteins repress the translation of viral genes by binding to eIF3 , a translation initiation factor , and virus RNA bearing 5′-triphosphate , leading to suppression of virus replication [61]–[63] . In addition , ectopic overexpression of IFIT2 promotes cell death by a mitochondrial pathway , revealing another potential anti-viral mechanism for these proteins [64] . The role of IFIT proteins in bacterial infections is less clear . In murine macrophages , overexpression of IFIT2 represses lipopolysaccharide ( LPS ) induced TNF- α and IL-6 expression [65] . In contrast , IFIT2−/− mice display reduced TNF- α and IL-6 in serum and LPS-mediated lethality in an endotoxic shock model , suggesting IFIT2 is a critical mediator for the secretion of LPS-induced proinflammatory cytokines [66] . In C . trachomatis , both IFIT1 and IFIT2 genes were reported to be up-regulated more than 10 fold in HeLa 229 cells by 16 hpi [67] . Our study shows that in the absence of TepP , the transcriptional level of IFIT1 and IFIT2 is similar to the level of uninfected A2EN cells , implying that TepP-mediated signaling regulates the expression of these genes during C . trachomatis infection . IFIT1 and IFIT2 can be induced directly by type I interferons [68] , or indirectly after activation of host pattern recognition receptors [69] . We are currently investigating what TepP-mediated signal transduction pathways may mediate the expression of IFITs and other genes early in infection and the consequences of these transcriptional events in colonization of the host . Taken together , our data suggests that TepP acts as a scaffolding protein that upon tyrosine phosphorylation , recruits additional scaffolding proteins like Crk , which in turn recruit more proteins to nascent inclusions , presumably to help establish an early niche for replication within the host ( Fig . 8 ) . In addition , the transcriptional response of host cells to infection with tepP mutants suggest that there is a distinct gene expression program that is dependent on TepP-mediated signaling events . We hypothesize that the compendium of genes activated and repressed in a TepP-dependent manner are important in establishing an immune environment within infected tissues that is more conducive to C . trachomatis colonization , survival and/or dissemination .
Chlamydia trachomatis biovar LGV , serotype L2 , strain 434/Bu was propagated in HeLa cells or Vero cells maintained in Dulbecco's Modified Eagle Medium ( Sigma-Aldrich , St . Louis , Missouri , USA ) supplemented with 10% fetal bovine serum ( FBS ) ( Mediatech , Manassas , Virginia , USA ) . Human endocervical epithelial A2EN cells [45] were maintained in keratinocyte-SFM medium ( Gibco , Life Technologies corp . , Grand Island , NY , USA ) supplemented with 10% FBS , 0 . 5 ng/mL human recombinant epidermal growth factor and 50 µg/mL bovine pituitary extract . EBs were purified by density gradient centrifugation using Omnipaque 350 ( GE Healthcare , Princeton , New Jersey , USA ) as previously described [31] . All recombinant protein expression was performed in Escherichia coli strain BL21 ( DE3 ) . The Y . pestis KIM8-E ( Δail ) strain [70] used in this study is avirulent and is excluded from the National Select Agent Registry due to the lack of the 102-kb pgm locus [71] . In addition , this strain carries deletions of the yopE , sycE and ail genes and has been cured of the plasminogen activator ( Pla ) -encoding pPCP1 plasmid . All experiments with Y . pestis KIM8-E ( Δail ) were reviewed and approved by the Institutional Biosafety Committee at the University of Miami . All reagents used are of analytical grade . Antibody generation was performed as previously described [30] . Briefly , recombinant proteins: His-tagged Slc1 ( Ct043 ) and Mcsc ( Ct260 ) , and GST-tagged TepP ( Ct875 ) were purified on affinity resins and used to immunize New Zealand White rabbits . Anti-GST antibodies were removed by pre-incubation with recombinant GST , and anti-TepP antibodies were affinity purified with TepP recombinant protein . E . coli BL21 ( DE3 ) was co-transformed with a pET24d vector expressing Slc1 and pGEX vector alone or pGEX expressing individual GST-tagged test protein . Protein expression was induced using 0 . 5 mM isopropyl-1-thio-β-d-galactopyranoside ( IPTG ) for 3 hours at 37°C . Cells were pelleted and lysed by sonication in binding buffer ( 1% Triton X-100 , 20 mM Tris , 150 mM NaCl , 1 mM EDTA , 1 mM PMSF , pH 7 . 4 ) , and GST–tagged proteins were isolated from the supernatant using glutathione-Sepharose beads ( GE Healthcare , Pittsburgh , PA , USA ) . After 4 h incubation , beads were washed 3 times with binding buffer , followed by 3 washes with washing buffer ( binding buffer-0 . 2% Triton X-100 , 300 mM NaCl ) . Bound proteins were solubilized in 1X Laemmli sample buffer ( 20 mM Tris-HCl , pH 6 . 8 , 1% SDS , 5% Glycerol , 10 mM DTT , 0 . 01% bromophenol blue ) and resolved by SDS PAGE , followed by immunoblot analysis with anti-Slc1 and –GST antibodies . Gel filtration chromatography was performed on proteins expressed in E . coli using a bi-cistronic vector to co-express untagged Slc1 , with His-tagged test proteins . The protein complexes were first purified using a Nickel resin ( GE Healthcare , Pittsburgh , PA , USA ) , eluted with 500 mM imidazole and applied to a Superdex 200 gel filtration column ( GE Healthcare , Pittsburgh , PA , USA ) for analysis . Fractions from each sample were collected and analyzed by immunoblot with anti-His and anti-Slc1 antibodies . Size markers , alcohol dehydrogenase ( 150 kDa ) , Conalbumin ( 75 kDa ) and Carbonic Anhydrase ( 29 kDa ) were purchased from GE and Sigma . Yersinia pestis KIM8-E ( Δail ) was co-transformed with the compatible plasmids pBAD33 and pBAD24 ( Table S2 ) either alone or in combination with pBAD plasmids encoding T3S chaperones and putative effectors . Transformed bacteria were grown overnight at 27°C in TMH media supplemented with ampicillin and chloramphenicol . Overnight cultures were used to inoculate fresh TMH media containing 2 . 5 mM calcium or without calcium at OD620 = 0 . 2 and incubated for 1 hour at 27°C . L-arabinose was added to 0 . 2% final concentration to induce chaperone and effector proteins expression and the culture temperature was shifted to 37°C for 5 hours . Bacteria were harvested by centrifugation and cell pellets were separated from culture supernatants . Proteins in the supernatant fractions were precipitated with 10% ( v/v ) trichloroacetic acid , resuspended in 1X SDS-PAGE sample buffer and normalized to the final OD620 of the respective culture . Approximately 5×104 HeLa cells/well were seeded onto glass coverslips placed in a 24 well plate . The following day , cells were incubated with LGV-L2 EBs at an MOI of 20 and infections were synchronized by centrifugation ( 3000 rpm for 30 min ) at 10°C followed by transferring the plates to a 37°C , 5% CO2 humidified incubator . At the indicated time points , the coverslips were fixed either with 100% methanol on ice for 15 min or with 3% formaldehyde/0 . 025% glutaraldehyde at room temperature for 20 min . Cells were then permeabilized with 0 . 2% Triton in phosphate buffer saline solution ( PBS ) , blocked with 3% BSA in PBS for 30 min and stained with antibodies against TepP ( 1∶10 ) , LPS ( 1∶250 ) ( EV1-H1 ) , MOMP ( 1∶250 ) ( gift from K . A . fields ) , phosphotyrosine ( 1∶100 ) ( Cell signaling #9411 ) , Slc1 ( 1∶100 ) , Crk ( 1∶10 ) ( BD Transduction Laboratories 610035 clone 22 ) . DAPI ( Invitrogen Life Technologies , Carlsbad , California , USA ) was used for staining nucleic acids . To permeabilize EBs at very early time points , 0 . 005% SDS in PBS was used . The images were further deconvolved using Huygens software ( Scientific Volume Imaging , Hilversum , Netherlands ) . For short-term infections , ∼105 HeLa cells were seeded per well in 24-well plates the day before experiment . Cells were incubated with LGV-L2 or its mutant derivatives at a MOI of 100 and infections were synchronized by centrifuging at 3000 rpm for 30 min at 10°C , followed by a shift to 37°C . Samples were collected at indicated time points by washing the well once with PBS followed by adding 90 µL 2X Laemmli sample buffer . For IPs of effector proteins , ∼5×106 HeLa cells were pre-seeded into a 10-cm cell culture dish the day before infections and cells were incubated with LGV-L2 or its mutant derivatives at a MOI of 100 . Infections were synchronized by pre-incubation for 1 hour at 4°C in Hanks balanced salt solution , followed by the addition of DMEM media pre-warmed to 37°C . At the indicated time points , infected cells were washed two times with ice-cold PBS and lyzed in 1 mL Pierce IP lysis buffer supplemented with 1 mM PMSF , protease inhibitor cocktail and Halt phosphatase inhibitor ( Pierce , Rockford , Illinois , USA ) . After sonication and high speed centrifugation to remove insoluble debris , antibodies against TARP [74] or TepP were added to cell lysates and incubated for 3 hours at 4°C . Protein A resin was added to isolate immunocomplexes . After washing thoroughly with Pierce IP lysis buffer , bound proteins were resolved by SDS PAGE and subjected to immunoblot analysis . IP . Two 15-cm cell culture dishes of confluent HeLa cells were infected with LGVL-2 at a MOI of 100 as described above . Another two dishes of confluent HeLa cells were mock-infected . Four hours post treatment , cells were collected , washed twice with PBS and lyzed as previously described [75] to generate cytosolic and nuclear/EB fraction . After clarifying the cytosolic fraction by high speed centrifugation , antibodies against TepP were added to the infected and control cell lysates and incubated for 3 hours at 4°C with continuous mixing . Immunocomplexes were purified with protein A resin and bound proteins were solubilized in SDS sample buffer . A strain containing a G309A ( W103* ) null allele in tepP was initially identified by whole genome sequencing of a collection of ethyl methyl sulfonate ( EMS ) -mutagenized and plaque-purified C . trachomatis LGV-L2 strains generated as previously described [43] ( Bastidas R . J . and Valdivia R . H . unpublished results ) . Strain CTL2-M062 harboring tepP G309A was identified from a pool of 20 mutants by Sanger sequencing of the tepP locus ( CTL0255 ) . CTL2-MO62 was recovered in sucrose-phosphate-glutamate ( SPG ) buffer ( 0 . 22 M sucrose , 0 . 01 M potassium phosphate , 0 . 005 M L-glutamic acid , pH 7 . 0 ) after hypotonic lysis of a monolayer of infected HeLa cells grown in a 10-cm2 cell culture dish . Lysates were sonicated ( 2×10 seconds in ice water ) and bacterial cells were spun down at 14 , 000 rpm , for 15 minutes at 4°C , and resuspended in 1X DNAse I buffer ( New England Biolabs , Ipswich , MA , USA ) . Cell suspensions were treated with 4 Units of DNAse 1 ( New England Biolabs , Ipswich , MA , USA ) for 1 hour at 37°C to deplete contaminating host DNA . Following a wash with PBS buffer , total DNA was isolated with a DNA isolation kit ( DNeasy tissue and blood kit , Qiagen , Valencia , CA , USA ) as described by the manufacturer . For whole genome sequencing , 1 µg of CTL2-M062 enriched DNA was fragmented with an Adaptive Focused Acoustics S220 instrument ( Covaris , Inc . Woburn , MA , USA ) , and DNA sequencing libraries were prepared with a library construction kit ( TruSeq DNA Sample Preparation Kit v2 , Illumina , Inc . San Diego , CA , USA ) according to the manufacturer's instructions . Libraries were sequenced in a MiSeq DNA Sequencing Platform ( Illumina , Inc . San Diego , CA , USA ) at the Duke University IGSP DNA Sequencing Core facility . Genome assembly and single nucleotide variant ( SNV ) identification was performed with Geneious Software version 6 ( Biomatters - http://www . geneious . com/ ) . The C . trachomatis L2 434/Bu genome ( GenBank no . NC_010287 ) was used as reference sequence . All non-synonymous SNVs identified in CTL2-M062 ( see Table S3 ) were independently verified by Sanger sequencing . Chlamydia recombinants were generated as previously described [43] . Briefly , confluent Vero cells grown on a 24-well plate were co-infected with CTL2-MO62 ( Rifampin resistant , RifR ) and a Spectinomycin resistant mapping strain ( SpcR ) at a ratio of 2∶4 and recombinant progenies were selected from among plaques that formed in the presence of Rif ( 200 ng/mL ) and Spc ( 200 µg/mL ) . Plaque-purified recombinants were further expanded in Vero cells and genotyped with SNV specific primers . The gene encoding TepP and its predicted promoter region ( 300 bp upstream of the TepP start codon ) was amplified by PCR from LGV-L2 genomic DNA and inserted into the E . coli-Chlamydia shuttle vector p2TK2-SW2 [78] . The recombinant TepPW103* strains ( G1 ) were transformed with either empty vector or vector harboring wild type TepP as previously described with some modifications [44] . Briefly , around 4×106 Vero cells treated with transformation buffer ( 10 mM Tris pH 7 . 4 in 50 mM CaCl2 ) were infected with G1 pre-incubated with >6 µg plasmid in transformation buffer . Transformed Chlamydia was selected under 1U of penicillin G ( Sigma P3032 ) for several passages . After initial selections , transformed Chlamydia was maintained in the presence of 10U penicillin G . TepP expression was confirmed by immunoblot analysis . Approximately 0 . 8×106 A2EN cells were seeded per well in 6-well plates the day before experiment . Duplicate cell samples were mock infected , or infected with the tepP mutant strain transformed with empty vector or the vector harboring wild type tepP at an MOI of 10 . Infections were synchronized by centrifugation at 3000 rpm for 30 min at 10°C , followed by an immediate shift to 37°C with pre-warmed cell culture media . Samples were collected at 4 hpi using QIAGEN RNeasy Plus Mini Kit ( QIAGEN , Valencia , CA , USA ) as described by the manufacturer . RNA integrity was assessed with Agilent 2100 Bioanalyzer G2939A ( Agilent Technologies , Santa Clara , CA , USA ) and quantified with a Nanodrop 8000 spectrophotometer ( Thermo Scientific/Nanodrop , Wilmington , DE , USA ) . Hybridization targets were prepared with MessageAmp Premier RNA Amplification Kit ( Applied Biosystems/Ambion , Austin , TX , USA ) from total RNA , hybridized to GeneChip Human Genome U133A 2 . 0 arrays in Affymetrix GeneChip hybridization oven 645 , washed in Affymetrix GeneChip Fluidics Station 450 and scanned with Affymetrix GeneChip Scanner 7G according to standard Affymetrix GeneChip Hybridization , Wash , and Stain protocols ( Affymetrix , Santa Clara , CA , USA ) . Data analysis was performed using Partek Genomic Suite 6 . 6 ( Partek Inc . , Saint Louis , MO , USA ) . Q-PCR was performed using Power SYBR Green RNA-to-CT 1-Step Kit and StepOne Real-Time PCR system as described by the manufacturer ( Applied Biosystems , Grand Island , NY , USA ) . Primers specific for IL-6 , CXCL3 , MAP3k8 , IFIT1 , IFIT2 , and Actin were designed using Roche Universal Probe Library ( http://www . roche-applied-science . com ) and are listed in Table S5 . Primers against Chlamydia 16S rRNA were used to quantify the amount of bacteria in each sample ( Table S5 ) . The use of rabbits for the generation of antisera ( Protocol A301-11-12 ) was approved by the Duke University Office of Animal Welfare Assurance ( OAWA ) after review by the IACUC committee . The IACUC ensures compliance of this protocol with the U . S Animal Welfare Act , Guide for Care and Use of Laboratory Animals and Public Health Service Policy on Humane Care and Use of Laboratory Animals .
|
Chlamydia trachomatis is an obligate intracellular bacterial pathogen that causes a range of human diseases of significant public health importance . To create a suitable replicative niche within its host , Chlamydia delivers effector proteins across mammalian membranes via a syringe-like apparatus termed a Type III secretion ( T3S ) system . The lack of a robust system for the molecular genetic manipulation of these pathogens has hindered progress in identifying and characterizing T3S effectors . In this study , we took a mass spectrometry-based approach to identify Chlamydia effector proteins based on their interaction with Slc1 , an abundant T3S chaperone . We identified a previously uncharacterized protein , Ct875/TepP , as a new T3S effector and determined that TepP is phosphorylated upon translocation into host cells , leading to the recruitment of the host scaffolding protein Crk and presumably manipulating Crk-dependent signaling functions . Finally , we provide genetic confirmation of the role of TepP in recruiting Crk and in modulating the expression of genes involved in innate immune responses to Chlamydia . This study is the first example of genetic validation of the function of a T3S effector in Chlamydia and a new example of a bacterial effector that directly co-opts the oncoprotein Crk to modulate host cell signaling events .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"mutagenesis",
"signal",
"transduction",
"gram",
"negative",
"genetic",
"mutation",
"tyrosine",
"kinase",
"signaling",
"cascade",
"genetics",
"microbial",
"pathogens",
"host-pathogen",
"interaction",
"biology",
"microbiology",
"molecular",
"cell",
"biology",
"bacterial",
"pathogens",
"signaling",
"cascades"
] |
2014
|
The Chlamydia trachomatis Type III Secretion Chaperone Slc1 Engages Multiple Early Effectors, Including TepP, a Tyrosine-phosphorylated Protein Required for the Recruitment of CrkI-II to Nascent Inclusions and Innate Immune Signaling
|
Dengue virus transmission occurs in both epidemic and endemic cycles across tropical and sub-tropical regions of the world . Incidence is particularly high in much of Southeast Asia , where hyperendemic transmission plagues both urban and rural populations . However , endemicity has not been established in some areas with climates that may not support year-round viral transmission . An understanding of how dengue viruses ( DENV ) enter these environments and whether the viruses persist in inapparent local transmission cycles is central to understanding how dengue emerges in areas at the margins of endemic transmission . Dengue is highly endemic in tropical southern Vietnam , while increasingly large seasonal epidemics have occurred in northern Viet Nam over the last decade . We have investigated the spread of DENV-1 throughout Vietnam to determine the routes by which the virus enters northern and central regions of the country . Phylogeographic analysis of 1 , 765 envelope ( E ) gene sequences from Southeast Asia revealed frequent movement of DENV between neighboring human populations and strong local clustering of viral lineages . Long-distance migration of DENV between human population centers also occurred regularly and on short time-scales , indicating human-mediated viral invasion into northern Vietnam . Human populations in southern Vietnam were found to be the primary source of DENV circulating throughout the country , while central and northern Vietnam acted as sink populations , likely due to reduced connectedness to other populations in the case of the central regions and to the influence of temperature variability on DENV replication and vector survival and competence in the north . Finally , phylogeographic analyses suggested that viral movement follows a gravity model and indicates that population immunity and physical and economic connections between populations may play important roles in shaping patterns of DENV transmission .
Dengue viruses ( DENV ) are single-stranded , positive-sense , mosquito-borne RNA viruses ( family Flaviviridae ) , within which considerable genetic diversity is present on both global and local scales [1] . All four serotypes ( DENV-1 to DENV-4 ) are capable of infecting humans and result in a spectrum of clinical outcomes , ranging from asymptomatic to severe disease . In the 1960s , fewer than ten countries reported a total of ∼15 , 000 dengue cases to WHO annually [2] , [3] . Rapid urbanization and global travel have fueled the global spread and establishment of DENV populations across the tropics and sub-tropics , and recent estimates suggest that approximately 96 million symptomatic cases now occur in more than 120 countries every year [4] . Endemic DENV transmission occurs throughout the tropics , while sub-tropical regions experience epidemics of varying size that subside with the shift to winter temperatures . Little to no detectable transmission occurs in these sub-tropical areas during cooler months and , in some areas , DENV transmission remains at very low levels for several years following an epidemic [5]–[7] . Phylogenetic studies of DENV from southern China and Taiwan suggest that seasonal epidemics often originate with DENV-infected travelers arriving from endemic areas of Southeast Asia [8]–[13] . Due to a lack of long-term virological surveillance in regions where dengue outbreaks are sporadic , it is not clear whether DENV lineages persist across seasons in these environments . In addition , there has been no investigation of the processes by which DENV are introduced and become established in human populations that are closely linked to endemic areas by land and air travel but sit at the geographic and climatic margins of endemic transmission . A better understanding of the invasion of DENV into these environments would provide important insights into the expansion of the geographic range of dengue and its emergence in sub-tropical environments . Human populations in southern Viet Nam have supported hyperendemic DENV transmission since at least the early 1960s and now experience fairly stable endemic transmission [14]–[19] . In contrast , dengue is considered to be ‘emerging’ in northern Viet Nam , where the annual incidence has increased over the previous decade [5] . While the clinical burden of dengue across Southeast Asia is generally found in children less than 15 years of age , 85% of dengue cases reported in Hanoi ( the capital of Viet Nam , located in the north of the country ) occur in adults [20] . This suggests later exposure to DENV , reduced population immunity , and thus a lower force of infection or a lack of hyperendemic transmission in northern Viet Nam relative to southern Viet Nam [5] , [21] , [22] . Climate is likely a major factor in these geographical differences in incidence and transmission intensities , as cool winter temperatures in northern Viet Nam may reduce mosquito breeding and survival and increase extrinsic incubation times such that year-round endemic transmission is inhibited [23]–[26] . Revealing how DENV move between geographic localities is integral to understanding dengue epidemiology in both endemic and epidemic areas . While the molecular epidemiology of DENV has been investigated in southern Viet Nam and neighboring countries [17] , [18] , [27]–[30] , the DENV populations of northern and central Viet Nam have not been described . Moreover , after a decade of increasing DENV activity , it is not clear whether year-round autochthonous transmission of local lineages of DENV occurs in northern Vietnam , or if viral populations die out there in the winter months and are reseeded from endemic areas ( particularly central and southern Viet Nam ) every year . Understanding the epidemiology and evolution of this emerging pathogen at the margins of transmission will provide valuable insights into the process by which DENV transitions from epidemic to endemic transmission , and may reveal factors that influence pathogen emergence in human populations . The aim of this study was to investigate the spread of DENV-1 throughout Viet Nam over the course of a decade and to determine the routes by which viral populations enter northern and central Viet Nam . For this , we utilized a large data set ( n = 1 , 765 sequences ) of DENV-1 envelope gene sequences collected from Vietnamese hospitals and studies across Southeast Asia , where Genotype I has been the dominant circulating DENV-1 lineage since at least 1980 . With these data , we investigated the movement of this lineage into Viet Nam and addressed the following questions: ( i ) Does highly endemic southern Viet Nam act as a source population for DENV circulating in other parts of the country ? ( ii ) Do DENV populations persist over multiple seasons in central and northern Viet Nam ? ( iii ) What factors determine the patterns of dispersal of DENV lineages to new environments across Viet Nam and within Southeast Asia ?
Dengue viruses were recovered from suspected dengue patients presenting to hospitals across Viet Nam as part of routine diagnostic serology . The envelope ( E ) genes of 60 isolates recovered from north and central Viet Nam and one from southern Viet Nam were sequenced as described previously [31] , [32] and have been submitted to Genbank ( submission numbers 1608348 and 1608374 ) . Twenty additional DENV-1 viruses were collected from cases presenting to The National Hospital for Tropical Diseases in Hanoi and sequenced using standard Sanger sequencing methods . These have been assigned GenBank accession numbers HQ591537-HQ591556 . The geographic and temporal distribution of all Vietnamese sequences is shown in Table S1 . DENV-1 E gene sequences from northern and central Viet Nam were combined with full-length DENV-1 E gene sequences catalogued in GenBank to comprise all DENV-1 sequences from across Asia for which the year and country of sampling were known . Nucleotide alignments of 1765 full-length DENV-1 E gene sequences ( 1485 nt ) , including the 80 isolates from northern and central Viet Nam , were manually constructed using Se-AL [33] . To infer phylogenetic relationships for the complete data set of DENV-1 sequences and identify geographic regions with phylogenetic links to northern and central Viet Nam , we utilized the maximum likelihood ( ML ) method available in PhyML , incorporating a GTR model of nucleotide substitution with gamma-distributed rate variation among sites and a heuristic SPR branch-swapping search algorithm [34] . This initial analysis indicated that all northern and central Vietnamese DENV-1 sequences belong to a Southeast Asian subset of Genotype I , comprising viral sequences from Thailand , Cambodia , and southern Viet Nam , as well as a maritime Southeast Asian lineage based in Singapore , Malaysia , and Indonesia . Within the maritime Southeast Asian lineage , these isolates were most closely related to Singaporean viruses . A second alignment was then constructed using 80 DENV-1 E gene sequences from northern and central Viet Nam and 625 unique E gene sequences of viruses isolated from the surrounding regions between 1997 and 2009 and for which the exact date of sampling was known ( Cambodia , Thailand , Singapore , southern Viet Nam ) . Small numbers of sequences and a lack of exact sampling dates for viral sequences from other countries in maritime Southeast Asia prevented us from including sequences from these countries in the analysis . Thus , viral , geographic , and epidemiological data from Singapore were used to represent the maritime Southeast Asian clade in all analyses . Phylogenetic analyses were undertaken using the Bayesian Markov Chain Monte Carlo ( MCMC ) method implemented in BEAST ( v1 . 6 . 2 ) , incorporating the date of sampling [35] and utilizing a codon-structured SDR06 model of substitution , a relaxed molecular clock as in [18] , and a Bayesian skyline prior ( BSP; 5 piecewise constant groups ) . The MCMC chain was run for 100 million iterations , with sub-sampling every 10 , 000 iterations . All parameters reached convergence as assessed visually using Tracer ( v . 1 . 5 ) . The initial 10% of the chain was removed as burn-in , and maximum clade credibility ( MCC ) trees were summarized using TreeAnnotator ( v . 1 . 6 . 2 ) . To investigate the routes of invasion of DENV-1 into northern and central Viet Nam , the geographic areas of Viet Nam were categorized using ( i ) a ‘Local geographic model’ – categorized by government-defined regions ( MKD: Mekong Delta , HCM: Ho Chi Minh City , SE: Southeast , CHL: Central Highlands , SCC: South Central Coast , NCC: North Central Coast , RRD: Red River Delta ) , and by country outside of Viet Nam ( KH: Cambodia , SG: Singapore , TH: Thailand ) , and ( ii ) a ‘Regional geographic model’ – categorized by larger regions ( North: RRD and NCC , Central: CHL and SCC , South: SE , HCM and MKD ) and elsewhere by country as in the previous scheme . We inferred rates of viral migration between locations using an asymmetric model of discrete diffusion across Southeast Asia and within Viet Nam [36] . Posterior distributions of trees were estimated under a phylogenetic model using the MCMC method implemented in BEAST ( v1 . 6 . 2 ) using BEAGLE [35] , [37] . This model incorporated the date of sampling and a relaxed molecular clock , Bayesian skyline prior , and the SRD06 codon position model as described above . The MCMC chain was run for 100 million iterations , with sub-sampling every 10 , 000 iterations , and all parameters reached convergence . The initial 10% of the chain was removed as burn-in , and Maximum Clade Credibility ( MCC ) trees including ancestral location-state reconstructions were summarized using TreeAnnotator ( v . 1 . 6 . 2 ) . The expected number of location state transitions conditional on the location-related sequence data was determined using Markov Jump counts , summarized per branch and for the complete evolutionary history . Markov Jump counts of the expected number of geographic state transitions along branches provide a quantitative measure of gene flow between regions , representing successful viral introduction from one region to another , and are not heavily influenced by single isolate introductions [38] , [39] . Finally , parsimony score ( PS ) and association index ( AI ) tests were utilized to assess the extent of geographic structure across all trees using the Bayesian Tip-association Significance Testing ( BaTS ) program [40] based on the posterior distribution of trees generated in the BEAST analysis described above . To account for potential sampling biases in space and time , posterior distributions were also estimated as above for ten data sets that were subsampled randomly , with replacement , to include no more than 50 sequences per region ( 338 sequences total ) for each of the ten major geographic regions in the data set ( Cambodia , Singapore , Thailand , and Viet Nam: Mekong Delta , Ho Chi Minh City , Southeast , Central Highlands , South Central Coast , North Central Coast , Red River Delta ) and for ten data sets that were subsampled such that no more than five samples per year were chosen randomly from each of the locations indicated above ( 205 sequences ) . In addition , six major clades containing northern and central Vietnamese isolates were identified in trees inferred from the full data set and were analyzed individually using the asymmetric discrete diffusion model to assess potential differences in the spatial and temporal patterns of diffusion among them , to identify routes of invasion into northern and central Viet Nam , and to determine whether lineages were maintained over multiple years in these newly sampled populations . To test hypotheses related to viral dispersal and establishment , we set rate priors to specific values to construct a series of ( asymmetric ) phylogeographic models that might reflect the epidemiology and dispersal of DENV in Viet Nam . These models were analyzed using both the full and subsampled data sets for the Regional and Local geographic models . These models were: ( i ) a geographic diffusion model that assumes equal rates of viral migration between all regions of interest ( Model 1 , Equal Rates ) , ( ii ) a model based on the physical distance separating the populations in question using the inverse of the Euclidian distance between the centroids of the largest cities in each region ( Model 2 , Distance ) , ( iii ) a population size-based model , utilizing the census population estimate of the largest city from which sequences were collected in each region as representative of the influence of that region in attracting migration from other locations ( Model 3 , Population ) [41]–[44] , and ( iv ) a previously described gravity model incorporating geographic distances ( calculated as in Model 2 ) and population size data from both the recipient and donor locations , in which distance and population size act as repelling and attracting forces , respectively ( Model 4: Gravity Model ) [17] , [45] . Prior human immunity to DENV is likely to play an important role in the ability of viruses to invade populations [46] . Similarly , transmission intensity , as reflected in the average/median age of infection [21] , will vary with time and hence influence patterns of spatial spread . To incorporate these factors into our phylogeographic models , we determined the ratio of the mean age of reported dengue cases in the recipient population to that in the donor population , and utilized this as a crude measure of the likelihood of viral invasion [5] , [22] , [47]–[51] . We refer to this ratio as the Relative Endemicity factor ( REf ) ( Model 5 , REf ) . Due to a lack of age-specific case data from local populations , Regional estimates were extrapolated to all locations within the same geographic region for the Local model . In Model 6 , we integrated this immunity measure as a proportionality constant in gravity model calculations ( REf+Gravity Model ) . Finally , we investigated the effects of sample size on phylogeographic inference using a rate matrix based on the sample size of the donor population ( Model 7 , Sample Size ) . Model priors were normalized ( mean one and unit variance ) and incorporated into asymmetric matrices that allow for directional rates to vary between individual location pairs . A posterior simulation-based analogue Akaike's information criterion through MCMC ( AICM ) was implemented using likelihoods specific to the geographic model priors , and marginal log likelihood estimates for each model were compared to determine the best fit model to the data in hand [52] , [53] . Sequencing of de-identified viruses collected in this study was undertaken under Human Research Ethics Approval 0700000910 from the Queensland University of Technology . Viruses collected at The National Hospital of Tropical Diseases in Hanoi , Viet Nam were from patients enrolled in a study approved by the scientific and ethical committees at the National Hospital of Tropical Diseases and The Oxford University Tropical Research Ethics Committee ( OXTREC ) [54] . Patients provided written informed consent to participate in this study .
We determined the E gene sequences of DENV-1 isolates collected from 81 Vietnamese dengue patients and combined these with Southeast Asian DENV-1 E gene sequences collected between 1997 and 2009 . The full data set included 46 sequences from northern Viet Nam ( 1998–2009 ) , 34 sequences from central Viet Nam ( 2004–2009 ) , and 461 sequences from southern Viet Nam ( 2003–2008 ) , and 70 , 63 , and 31 envelope gene sequences from Thailand ( 1997–2007 ) , Cambodia ( 2000–2008 ) , and Singapore ( 2003–2008 ) , respectively . All viruses were collected subsequent to a clade replacement event that occurred within the DENV-1 population in Thailand in the mid-1990s that has been attributed to enhanced transmission capacity within the vector [55] . This appears to have been the primary DENV-1 lineage circulating in mainland Southeast Asia for at least a decade , although the lack of samples from Cambodia and Viet Nam in early years prevents investigation of the means by which this lineage initially spread through the region . Thailand , Cambodia , and Viet Nam all experienced the co-circulation and maintenance of multiple lineages for several years and importation of novel viruses from other countries ( Figure 1 ) . Our phylogeographic analyses provided no support for Cambodia or Viet Nam acting as a source of recent DENV-1 lineages circulating in Thailand [mean Markov Jump counts ( 95% highest posterior density [HPD] ) : KH to TH , 0 . 16 ( 0 , 1 ) ; South VN to TH , 0 . 04 ( 0 , 0 ) ; Central VN to TH , 0 . 08 ( 0 , 1 ) ; North VN to TH , 0 . 54 ( 0 , 3 ) ] , while moderate support was provided for migration routes from Thailand to Cambodia and from Cambodia to southern Viet Nam ( Table 1 , Table S2 ) . Due to differences in sampling densities over time , conclusions on the geographic origins of some lineages cannot be made . However , lineages in which contemporaneous sequences are present in both Cambodia and Viet Nam ( Clades 1 , 4 , 5 , and 6 ) strongly support a Cambodian origin of Vietnamese DENV-1 populations ( Figure 1 ) . Although frequent movement of viruses between locations was observed between 1990 and 2009 , very strong clustering by country and sub-national region within Viet Nam indicates that gene flow is much higher within the defined geographic areas than between them [Regional analysis – full phylogeny: Association Index , AI = 0 . 06 ( 0 . 05 , 0 . 08 ) , Parsimony Score , PS = 0 . 16 ( 0 . 15 , 0 . 17 ) , subsampled trees: averaged AI: 0 . 12 ( 0 . 10 , 0 . 15 ) , averaged PS = 0 . 25 ( 0 . 24 , 0 . 27 ) ; Local analysis – full phylogeny: AI = 0 . 19 ( 0 . 17 , 0 . 21 ) , PS = 0 . 27 ( 0 . 25 , 0 . 28 ) , subsampled trees: averaged AI: 0 . 25 ( 0 . 22 , 0 . 30 ) , averaged PS = 0 . 38 ( 0 . 35 , 0 . 40 ) ] . A minimum of eight distinct lineages of DENV-1 Genotype 1 entered Viet Nam between 1990 and 2007 and persisted until at least 2007–2009 . While all viral diversity captured within the country was represented in samples from the south , viral populations in the northern and central regions were less diverse . Nearly all viruses isolated from northern and central Viet Nam clustered within the diversity of the south with the exception of viruses isolated in northern Viet Nam prior to 2003 and a single divergent virus collected in the Central Highlands in 2004 ( basal to Clades 5 and 7 ) ; notably , very few sequences were available from southern Viet Nam during this time period ( nine in 2003 , two in 2004 ) . While strong support was found for a Cambodian origin of most Vietnamese lineages , Clades 7 and 2 may have been introduced from Singapore ( or elsewhere in maritime Southeast Asia ) and Thailand , respectively . Clade 7 contained two examples of importation of novel lineages of the maritime Southeast Asian clade into northern Viet Nam followed by localized , short-term transmission during a single year and apparent fade-out with the onset of winter ( Table 2 ) , when temperatures are low and vector populations are expected to be reduced . These analyses also suggested migration of viruses from maritime Southeast Asia into Thailand in 2004 followed by sustained co-circulation with indigenous Thai DENV-1 through the 2007 dengue season ( Figure 1 ) . A previous analysis of DENV-1 within southern Viet Nam indicated that Clade 2 became established there in 2002 [18] . The addition of isolates from northern Viet Nam in this study showed a considerably longer history of this lineage in the country , beginning in the late 1980s/early 1990s . Ancestral state reconstruction suggested that this lineage migrated from Thailand into northern Viet Nam , but Markov Jump counts at the basal node are quite low ( Regional model: 0 . 18 , Local model: 0 . 40 ) due to the existence of only a few sequences from viruses recovered during the early period of invasion and long branches between lineages . This was investigated further in clade-specific analyses and is discussed below . Of six viral lineages involved in transmission within the central and northern regions of Viet Nam , five showed invasion and dispersal throughout the country and frequent movement between areas of interest . To investigate the spread of DENV-1 genotype 1 within Viet Nam , we estimated Markov Jump counts between locations across the full phylogeny and in subsampled data sets , and analyzed specific viral clades to investigate fine-scale spatial and temporal patterns of dispersal . Using the Regional asymmetric phylogeographic migration model , we determined that southern Viet Nam was the likely source for Vietnamese viruses in Clades 1 , 4 , 5 , and 6 and for Clade 2 viruses isolated after 2002 ( Table 1 ) . No support was found for viral migration between the central and northern regions of the country [mean Markov Jump counts across the full phylogeny ( 95% HPD ) : central VN to northern VN , 0 . 23 ( 0 , 1 ) ; northern VN to central VN , 0 . 26 ( 0 , 1 ) ] , although the small numbers of sequences obtained from these regions may have obscured any such links if sampling was not representative of the full diversity in these locations . Finer scale spatial analysis using the Local model showed that DENV tended to move between neighboring areas , but also implicated Ho Chi Minh City as the primary source population for the entire country . Significant migration was detected from this densely populated urban area into the surrounding Mekong Delta and Southeast regions , as well as to the South Central Coast and the distant Red River Delta ( Table 1 , Figure S1 ) . These relationships were consistent across most clades and within sub-sampled data sets ( Table S2 ) , which suggested that inferred relationships are not an artifact of dense sampling in the south . The Mekong Delta and Southeast regions also acted as secondary centers of viral diversity within the country . The Mekong Delta region is the inferred entry site of Clade 1 into Viet Nam . Long-term transmission of these viruses as well as sub-lineages in Clades 2 and 4 also occurred in the Mekong Delta . However , viruses only disseminated from this region into areas with which it shares borders , namely Ho Chi Minh City and the Southeast ( Table 1 , Figure S1 ) . The Mekong Delta acted as a significant source for these populations across the full phylogeny based largely on several migration events and subsequent establishments in Ho Chi Minh City in Clade 1 ( Table 1 , Figure 2 ) . Notably , no links were detected between the DENV populations in the Mekong Delta and central or northern viral populations . In contrast , the Southeast region did not appear to play a significant role in maintaining diversity within the south and instead generally acted as an acceptor of viruses from Ho Chi Minh City and , to a lesser extent , from the Mekong Delta . While limited sampling could have obscured a role for this region as a source population in the south , analysis of the full data set revealed that the area is a significant source of viral populations appearing in the Red River Delta ( Table 1 ) , and branch-specific Markov Jump counts suggest this area as an independent source of established DENV populations in the Red River Delta and North Central Coast in Clades 4 and 6 ( Figure S1 ) . To determine whether distinct viral populations persisted in northern and central Viet Nam over multiple seasons , we investigated the timing of invasion and establishment of novel viral sub-lineages ( that is , viral clusters of two or more sequences originating from the same location as inferred by ancestral state reconstruction ) and the potential co-circulation of distinct lineages in these areas . Viral invasion and establishment were detected in northern Viet Nam in 1990 , 2004 , 2008 , and 2009 , and in central Viet Nam in 2003 , 2006 , and 2008 . Except in Clade 2 , these data suggest that invading lineages in northern Viet Nam did not persist in the region over multiple dengue seasons ( Table 2 ) . The times to most recent common ancestry ( TMRCAs ) of viral clusters in northern Viet Nam suggested that viruses are imported to the region throughout the year , although most invasion events occurred in the middle of the year via viral migration from the south ( Ho Chi Minh City and the Southeast ) . This period coincides with seasonal increases in the number of dengue cases throughout the country and in Hanoi [5] , when viral migration and establishment are more likely due to high levels of DENV transmission in the south and suitable climate conditions for the vector in the north . In contrast to the north , DENV from six central Vietnamese transmission clusters ( one each in 2003 , 2006 , and four clusters in 2008 ) suggested that seasonal invasion in this region occurred during the ‘dengue season’ in the second half of the year ( Table 2 ) , with uninterrupted transmission often maintained for multiple years . Among these persistent lineages , one central Vietnamese sub-lineage in Clade 2 became established in the Central Highlands around 2003 with viruses from this lineage later isolated in the South Central Coast region , where it was maintained into the 2009 dengue season . The concurrent invasion and co-circulation of multiple clades was common in the South Central Coast region ( two in 2006 , Clades 2 and 4; five in 2008–2009 , Clades 1 , 2 and 4 ) . To determine how viral transmission routes within Southeast Asia were influenced by population and geographic factors , we compared the fit of a variety of phylogeographic models related to human population size , distance , and DENV transmission intensities to the spatially- and temporally-related E gene sequence data . Although slightly different results were obtained for the Local and Regional models ( Figure 3 , Table 3 ) , phylogeographic models that utilized simple distance- and population-based gravity model priors generally showed a good fit to the data relative to the equal rates and single factor models ( distance , population ) , as indicated by lower values of marginal log likelihood AICM estimates . However , the incorporation of the Relative Endemicity factor ( REf ) , which reflects DENV transmission intensities in both recipient and donor populations , further improved the fit of gravity models to the data under the Regional geographic scheme . The REf alone showed a consistently good fit to the data relative to other single factor models , and the best overall fit was shown by the REf+Gravity Model ( Model 6 ) both for the full phylogeny and randomly subsampled data sets ( 50 per location ) . Importantly , the performance of the Sample Size model ( Model 7 ) was not significantly better than other models of viral movement in the full or subsampled data sets . This indicates that spatial bias in sampling was not the most important factor determining the patterns of viral migration observed here .
This study documents the dispersal of DENV-1 to populations across Viet Nam and provides evidence that viral populations are regularly introduced into northern Viet Nam from external populations but do not establish endemic cycles of transmission . Strong clustering at all spatial scales indicated that the viral diversity present in a given area is determined primarily by local gene flow . Frequent movement between neighboring locations under the Local geographic model may reflect a combination of human movement and vector dispersal in the movement of DENV between human populations in close proximity , as observed at smaller geographic scales [17] , [18] , [56] . However , many of the medium- and long-distance migration events observed here occur on a timescale of less than one year , suggesting that human-borne virus migration drives the long-range dispersal of DENV from south to north . Our results also show that the DENV population in sub-tropical northern Viet Nam is characterized by regular seasonal invasions of lineages from the highly endemic south , with no detectable persistence into the following dengue season . Regardless of the time of year in which invasion occurred , our phylogenetic data suggest that invading lineages experience severe seasonal bottlenecks and regular fade-out in northern Viet Nam at the end of each year , when temperatures in much of the north drop below those considered optimal for survival of the vector and efficient transmission of the virus by Aedes aegypti [23]–[26] . However , the short-term transmission of strains of DENV in the first half of the year , which are suggested to have been introduced into the north during the cold , dry winter , indicate that these seasonal conditions are not sufficient to completely block transmission – perhaps due to residual , indoor breeding of A . aegypti mosquitoes . In contrast , strains of DENV introduced into central Viet Nam establish cycles of transmission extending over multiple years . Clade 2 represents one possible exception to the observations above . Our analysis suggests that this lineage entered through the north and became established in the North Central Coast , where it may have persisted for over a decade prior to its invasion and establishment in the south . Although the possibility of long-term transmission in northern Viet Nam cannot be excluded , the lack of contemporaneous sequence data from the south and the distant relationships between these basal sequences result in a lack of resolution at this early time point and thus low support for an inferred ancestral location of the lineage . Additionally , a number of other sequences sampled from early time points in the north ( Clade 4 ) also fall at basal positions in the phylogeny , although generally within the diversity of southern Viet Nam . Thus , the relationships among these northern sequences do not necessarily indicate their persistence . Instead , they may represent multiple importations from DENV populations in the south in the 1990s that experienced a significant bottleneck in the early 2000s , prior to the entry and rapid establishment of Clades 1 , 4 , 5 , and 6 . If the processes of DENV invasion in Viet Nam were similar to those observed at more recent time points when sampling was conducted across the country , we would expect that these northern sequences would fall into southern lineages that circulated prior to the bulk of our sampling . However , the lack of any signature of intermediate diversity suggesting the presence of this lineage in the highly sampled south prior to 2002 makes it difficult to test this hypothesis . Among the more recently sampled sequences , there are a number of interesting patterns of viral dispersal within Viet Nam . Ho Chi Minh City and the Mekong Delta experience high transmission intensities and were highly sampled relative to the other populations ( Table S1 ) . Previous studies indicated that Ho Chi Minh City acted as a source of DENV diversity in the south [17] , [18] . Here , we show that the role of the city as a primary source population extends across the entire country . The Mekong Delta , in contrast , was a source of DENV for populations only within the south . While fewer samples were available for the Southeast , this region appeared to be a significant source of viruses circulating in the north ( Red River Delta and North Central Coast ) . The Southeast region has high population densities adjacent to HCMC and is the site of increasing numbers of industrial parks and emerging regional economic centers [57] . Economic migration to HCMC and the Southeast is especially common among young adults from the Mekong Delta , Central Coast regions , and the Red River Delta [58] . Importantly , young adults from the north may be dengue-naïve due to limited exposure to DENV , and may be at high risk of infection and illness after arriving in the southern industrial areas [59] . Movement of both migrants and short-term travelers between the economically important regions of Ho Chi Minh City , the Southeast , and Hanoi provides ample opportunity for the movement of DENV and a range of other important human pathogens . Improvements in road quality and accessibility to long-distance travel by air , land , and water over the last few decades have likely resulted in an increase in human movement throughout the country , and our results suggest that these movements may be partially responsible for changes in DENV activity in endemic and non-endemic regions . However , a lack of viral movement between the Southeast region and central Viet Nam , while similarly connected by human migration [58] , may reflect ( i ) higher levels of immunity in the center of the country such that viral establishment is rare even when viral importation is frequent , ( ii ) differences in human movement between central Viet Nam and other regions , or ( iii ) insufficient sampling of this area . Indeed , a lower average age of dengue cases in this region and the finding that central Viet Nam maintains local lineages over multiple dengue seasons suggests that levels of population immunity here may be higher than in the north . Previous studies in southern Viet Nam have suggested that viruses move through the area along somewhat predictable human migration routes based on estimates of physical and economic connectedness [17] , [18] . Here , we compared a variety of epidemiological models reflecting patterns of human and mosquito movement on a much larger scale than the previous studies and showed that models incorporating patterns of human migration fit the data relatively well . The simple gravity model showed the best fit to the data under the Local geographic model and suggests that the movement of DENV between locations can be explained by the connectedness of the human populations at this scale . This is reflected in the frequency of medium- and long-distance viral migration events between population centers in our phylogenies . However , the addition of even a crude factor related to population immunity and transmission intensities ( REf ) in the recipient and donor populations improved the fit of the gravity model in the Regional geographic analysis . This discrepancy between the Local and Regional model is not surprising given that REf estimates were based on limited data and extrapolated to all locations within a region , and thus may not reflect complex heterogeneities in immunity and transmission at finer spatial scales . The finding that the REf performs well against other factors indicates that transmission intensities in both the recipient and donor populations play an important role in shaping the likelihood of viral invasion . Notably , however , the co-circulation and frequent invasion of DENV-1 lineages across all populations suggests that large susceptible populations exist across the region , even in areas of high transmission intensity . In this case , the role of immunity may be limited relative to the number of infections occurring in a given area over time . We acknowledge that the REf estimation is a greatly over-simplified indicator of relative population immunity and transmission intensities , and is based on extrapolation of the mean age of infection from multiple studies that used diverse methods and surveillance data sources . These data may not be directly comparable , and it is clear that this crude estimation technique and the underlying data should be refined . Although we believe that REf calculations in part integrate issues related to vector biology , reflecting a lower force of infection in Singapore ( where aggressive vector control efforts may limit transmission ) and in northern Viet Nam ( where winter temperatures likely limit vector density and competence ) , the lack of models that explicitly integrate vector biology is a weakness of this analysis . The availability of data on vector densities and species in the locations considered is limited , and the spatial and temporal scales of these analyses do not easily lend themselves to explicit consideration of vector dynamics in our models . As analytical methods and the scale of viral sampling improve , the use of phylogeographic methods that model the impact of factors such as vector density and vector competence on the dynamics of DENV in locations where such data are systematically collected may offer greater insight into the processes that mediate viral migration and establishment in new regions . Additionally , a better understanding of true human movements in the region ( including air , land , and water routes ) may allow us to further elucidate the relative roles of human movement , vector species and competence , and population immunity in the dispersal and persistence of DENV in these environments . Phylogeographic inference may be strongly affected by uneven sampling in space and time [60] . Here , we employed multiple methods to control for both spatial and temporal biases in our data , and these consistently upheld most of the migration links between locations as inferred in our full DENV-1 data set . However , even under our subsampling schemes , the data were biased toward sequences from the south ( 2006–2008 ) and the small number of sequences obtained from areas in northern and central Viet Nam may have been insufficient to capture long-term persistence of rare lineages co-circulating with the dominant invading viruses . Additionally , differences in sampling over time hindered inference at deep locations in the tree and prevented us from determining the origins of Clade 2 , a possible long-term northern DENV lineage . Importantly , the spatial sampling bias inherent to these data ( 85% of our Vietnamese sequences are from the south ) reflects the reality of the dengue burden in the country , where 85% of all reported dengue cases occur in the south [20] . It is difficult to identify an appropriate sampling scheme for phylogeographic analysis , particularly given that large reductions in the number of samples from the south have the potential to greatly reduce the overall diversity of the data set . It remains to be determined whether more even sampling in all locations across time would yield different or more robust results , as this type of sampling could bias the analysis to represent an unrealistic epidemiological scenario . As studies of viral phylogeography become more common in diverse environments , it is important that appropriate methods of systematic sampling ( and resampling ) are developed to optimize inference under varied epidemiological and evolutionary scenarios . Given recent increases in DENV incidence in northern Viet Nam and its geographic position at the margins of endemic transmission , additional sampling in this area is clearly warranted . Results here suggest that human populations that are connected to dengue-endemic regions may be at constant risk of DENV invasion if effective vector species are present , and that aggressive vector control measures may be necessary to prevent epidemics , even in sub-tropical and temperate regions with little to no history of DENV activity . Greater understanding of the processes by which DENV invades sub-tropical northern Viet Nam and the potential of this area to maintain long-term autochthonous viral transmission would yield important information relevant to sub-tropical and temperate areas at risk of DENV invasion worldwide .
|
Reports from sub-tropical regions of the world suggest a growing risk of introduction and establishment of dengue viruses ( DENV ) in new locales . Recent dengue epidemics in northern Viet Nam present an opportunity to study how DENV invades and spreads in these environments . The proximity of this region to tropical areas experiencing year-round endemic DENV transmission makes it an ideal site for studying the effects of human population movement and climate on DENV emergence . We performed a phylogenetic analysis using DENV-1 envelope gene sequences from Southeast Asia . We show that DENV are regularly imported into northern and central Viet Nam from southern Vietnam , and that increasingly large seasonal epidemics in the north are caused by newly introduced viruses each year . While tropical Vietnam maintains localized virus populations for multiple years , cool winter temperatures in sub-tropical northern Viet Nam may reduce mosquito populations and virus replication to levels that are not conducive to year-round DENV transmission . Finally , we found that the dispersal of DENV across the region is well-described using human movement and immunity data , and believe that increased epidemiological , entomological , and virological surveillance are needed to understand the processes by which endemic DENV transmission becomes established in new populations .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2013
|
Dengue Virus in Sub-tropical Northern and Central Viet Nam: Population Immunity and Climate Shape Patterns of Viral Invasion and Maintenance
|
In multi-cellular organisms , tissue homeostasis is maintained by an exquisite balance between stem cell proliferation and differentiation . This equilibrium can be achieved either at the single cell level ( a . k . a . cell asymmetry ) , where stem cells follow strict asymmetric divisions , or the population level ( a . k . a . population asymmetry ) , where gains and losses in individual stem cell lineages are randomly distributed , but the net effect is homeostasis . In the mature mouse intestinal crypt , previous evidence has revealed a pattern of population asymmetry through predominantly symmetric divisions of stem cells . In this work , using population genetic theory together with previously published crypt single-cell data obtained at different mouse life stages , we reveal a strikingly dynamic pattern of stem cell homeostatic control . We find that single-cell asymmetric divisions are gradually replaced by stochastic population-level asymmetry as the mouse matures to adulthood . This lifelong process has important developmental and evolutionary implications in understanding how adult tissues maintain their homeostasis integrating the trade-off between intrinsic and extrinsic regulations .
Development and tissue homeostasis of multi-cellular organisms is an extraordinary cellular orchestra starting from a single zygote [1] . Cascades of cell divisions generate and subsequently maintain a great diversity of cells in an organism [2] . This life-long balance is strictly controlled and maintained through a rigid cellular hierarchy , where the stem cells lie at the apex of the division cascades [3] . Stem cells are a group of cells with a dual role . On one hand , they need to maintain their own population through self-renewal . On the other hand , stem cells also give rise to differentiated cells which carry out most body functions [4] . In order to fulfill the dual role of self-renewal and differentiation , stem cells can undergo two different modes of cell division – asymmetric and symmetric [5] . In the asymmetric division mode , one daughter cell is maintained as the stem cell and the other goes on and evolves into terminally differentiated cells . The stem cells can also divide symmetrically , leading to either two stem cells or two differentiated cells . Asymmetric division is particularly attractive and allows stem cells to accomplish both maintenance and differentiation simultaneously in a single division . However , symmetric divisions are also indispensable in situations such as morphogenesis and tissue injury where stem cells need to proliferate rapidly [6] , [7] . A robust balance between proliferation and differentiation must be maintained to prevent aberrant growth on one hand and tissue loss on the other [5] . Stem cells often form distributed clusters and live in local nurtured structures known as the stem cell niches [8] , [9] . In order to maintain a static hierarchy between different cell types , two different strategies can be employed . In the first strategy ( also called cell asymmetry ) [10] , stem cells engage only in asymmetric divisions where dual roles of self renewal and differentiations can be successfully fulfilled while keeping the stem cell number constant . Population level equilibrium is achieved by maintaining a stasis at the single cell level through asymmetric cell divisions . Studies looking at invertebrate systems , in particular Drosophila melanogaster and Caenorhabditis elegans , have found a predominance of asymmetric divisions where stem daughter cells remain within the niche and differentiated cells exit and evolve into functional cells [11] , [12] . Biological evidence for cell asymmetry is quite strong in many invertebrate systems [10] . In the other extreme ( also called population asymmetry ) , each stem cell division gives rise to one stem cell and one differentiated cell on average [10] . Homeostasis is maintained by having a subset of cells proliferate while other stem cells are lost through differentiation . If the gain and loss are balanced , stasis is achieved at the population level rather than at the level of individual cell divisions [13] . In contrast to invertebrates , it appears that population asymmetry is more prevalent in mammals [14] . Mammalian intestine has become one of the best model systems for studying stem cell dynamics [15]–[18] . Powerful genetic tools together with recently-identified intestinal stem cell markers enable us to directly trace stem cells [19]–[21] . As a result , dynamics of cell populations are becoming accessible to investigation [22] . Interestingly , the relative prevalence of symmetric and asymmetric divisions among studies conducted to date is not yet clear . On one hand , lineage tracing techniques together with models from statistical physics reveal a pattern of neutral drift in a group of equipotent stem cells [23] , [24] . Population asymmetry with a predominance of symmetric divisions is the major mode of stem cell renewal in the adult mouse intestinal crypt . On the other hand , optimal control theory together with experimental data indicates that early development of the mouse intestinal crypt is achieved by a surge of symmetric divisions establishing the stem population followed by a transition to predominantly asymmetric divisions [25] . Moreover , molecular evidence for asymmetric division is starting to accumulate , suggesting that the role of asymmetric divisions might be underappreciated [26] . The dynamics of stem cell renewal , in particular the balance between cell asymmetry and population asymmetry of stem cells , remain enigmatic . Genomic sequencing , in particular single cell sequencing , provides a powerful alternative approach for studying cell lineage relationships . Compared to traditional molecular techniques such as the lineage tracing [27] , spontaneous somatic mutations provide a natural internal cell marker for tracing relationships in a group of cells . In this work , we use a previously published dataset comprising single-cell sequence information collected from mouse intestinal crypts [28] . The authors of this study sequenced multiple microsatellite markers in a repair-deficient mouse strain ( Mlh1−/− ) [28] , [29] . The reduced efficiency of DNA repair machinery results in higher microsatellite mutation rate and thus increased genetic variation , allowing us to discern genealogical relationships in this group of cells . Using traditional phylogenetic methods , the authors of the previous study found that intestinal crypts do not support the immortal strand hypothesis . Instead , they found support for the existence of monoclonal conversion , a process by which multiple crypt cells drift toward monoclonality , where offspring population is only derived from a single ancestor [28] . Population-genetic theory [30] , in particular coalescent theory [31] , [32] , provides a natural framework for studying cells in a population . Population dynamics driven by a combination of symmetric and asymmetric divisions can be explicitly modeled . When we take a sample of cells from a tissue , there will be a genealogical relationship relating individual cells to their common ancestor [33] . The shape of this genealogy is governed by the mode of cell division and thus carries information about underlying population dynamics . In reality , we do not directly observe the genealogy , but rather genotype information ( e . g . microsatellite markers presented here ) collected from individual cells . By considering all possible ancestral relationships compatible with a given pattern of genetic variation ( instead of just a single gene genealogy in the phylogenetic framework [28] ) , we can infer the underlying balance between symmetric and asymmetric divisions using statistical modeling . Here , we demonstrate that this approach , based on classical population genetics , can provide powerful insights into cellular dynamics within an organism and supply fine-scale quantitative description of the processes underlying cellular homeostasis .
We consider a discrete-generation model of tissue homeostasis . In each cell generation , a proportion α of the cells divides symmetrically and gives rise to two descendant stem cells ( type I , Figure 1A ) . A fraction β of the cells divides asymmetrically ( type III ) and 1-α-β cells divide symmetrically and produce two differentiated cells ( type II ) . Because type II divisions do not give rise to any stem cell descendants , the number of stem cells in generation t will be Nt = ( 2α+β ) t×N0 , where N0 is the population size at time 0 . Now , suppose we pick two stem cells at random at time t , the probability that they will have a common ancestor in the previous generation can be computed in two steps . The first stem cell picked must be derived from the type I stem cell division in the previous generation and the probability of picking it is 2α/ ( 2α+β ) . Secondly , the other sampled stem cell must be the pair of the first picked stem cell in the type I division and the probability of picking it is 1/ ( Nt−1 ) . Thus the probability of finding a common ancestor ( i . e . a coalescence ) in a single generation backwards in time for two stem cells is: ( 1 ) The number of stem cells in intestinal crypts is approximately constant . We thus assume that 2α+β = 1 and Nt = N0 = N . Then , the above probability can be rewritten as 2α/ ( N−1 ) . Once we have this single-step probability , other quantities such as the time to the most recent common ancestor for two cells and the coalescent relationship in a sample can be derived following the n-coalescent approach ( Materials and Methods ) [31] , [32] . In this parameterization , both cell asymmetry and population asymmetry are special cases of a general model . For example , strict cell asymmetry will correspond to cases where β = 1 . This general framework will allow us to test and pick the best models based on empirical observations . Each intestinal subunit is composed of two parts: a protrusion compartment called villus , which contains terminally differentiated cells , and an invagination compartment named crypt , which hosts stem cells and highly proliferative transit-amplifying cells . There is a continuous process that replaces functional cells in the villi with cells grown out of the crypts . We used a two-deme population-genetic model to capture continuous renewal of stem cells and transit-amplifying cells ( Figure 1A ) . In each generation , the stem and non-stem cell population follow a dynamic process as described in the previous section . The only exception is that differentiated descendants from the stem cell deme ( population 1 ) are constantly migrating into the other population ( transit amplifying cell deme ) ( Figure 1B ) . One-way migration in the two-deme population model reflects the coupling of the stem and non-stem populations in the intestine . In practice , when we take a random sample of cells from a crypt , we do not know whether they are stem or differentiated cells . Thus , the number of sampled cells from two demes ( denoted as ( m , n ) ) will follow a hypergeometric distribution ( N1 , N2 , m+n ) , where N1 and N2 are population sizes for the two demes . Given ( m , n ) cells from deme 1 and deme 2 , the coalescent process of going backwards in time and finding common ancestors for these lineages can be modeled using a Markov Chain ( Figure 1B and 1C ) . The transition probability between neighboring states can be calculated as a combination of individual coalescence or migration events . For example , a single coalescence and one migration event in deme 2 will change the state from ( m , n ) to ( m+1 , n−3 ) ( Figure 1C ) . For the sake of computational efficiency , we only consider transitions involving a maximum of two events ( either coalescence or migration ) . Probability of three events occurring in one transition is low and is neglected in our calculations ( Text S1 ) . Similar to previous studies [34] , when we compare the Markov Chain results with exact calculations performed through forward simulations , we find that approximate results provide an accurate characterization of the underlying dynamics ( Figure 1D , Materials and Methods ) . Thus , a two-deme population-genetic model and the associated coalescent process can be used to model dynamics underlying cellular homeostasis of the intestinal crypt . Given observed genotype information ( e . g . microsatellite markers , denoted as D ) , obtained by assaying single cells , the likelihood of the data can be calculated as ( 2 ) where θ is the set of model parameters , including symmetric/asymmetric division parameters ( α , β ) and the population size parameters ( N1 , N2 ) . The G ( i . e . gene genealogy ) represents coalescent relationships in a sample of cells . In Eq . 2 , the Pr ( D|G ) can be computed using standard phylogenetic methods and the pruning algorithm can be employed to evaluate this term [35] . The Pr ( G|θ ) can be calculated from the coalescent process using a Markov Chain or forward simulation ( see the following sections ) . Since we do not directly observe the underlying gene genealogy , we need to take into account all possible ancestral relationships that are compatible with the data in order to evaluate the likelihood . By integrating over all these genealogies , the likelihood of observed data can be calculated as a function of the underlying parameters ( Eq . 2 ) . In practice , given the large dimensionality of genealogical spaces , it is not feasible to exhaustively explore all possible ancestry relationships in a sample . Instead , we use a Monte-Carlo approach to compute the likelihood in Eq . 2: ( 3 ) where Gi is sampled from Pr ( G |θ ) . It can be shown that , as k increases , likelihoods from Eq . 2 and Eq . 3 will converge to the same value in the limit [36] . Sampled gene genealogies can be drawn either from the Markov Chain or forward simulations depending on whether a cell population has reached equilibrium ( i . e . stationary distribution ) at the time of data acquisition . Markov Chain calculations assume that a given population has reached equilibrium under a given configuration of symmetric/asymmetric division rates , which may not always be true . On the other hand , forward simulation can be applied to either non-stationary or stationary scenarios , but is computationally much more expensive . Through computational simulations , we found that stationarity has been reached for samples taken on day 340 across most of the parameter space , but not on day 52 ( Text S1 ) . Therefore , we generated genealogies for those non-stationary scenarios by simulating a crypt population history with a phase of crypt morphogenesis , followed by a period of homeostatic renewal ( Materials and Methods ) [25] . The generation number for the associated time point is calculated as the number of cell divisions within a given amount of time . For example , the stem cells are dividing at a rate of about once every 22 hours [25] ( Materials and Methods ) . Since genealogical relationships are directly recorded during the course of the computer simulation , simply picking a sample of cells at the end of a simulation run yields a sample from the distribution of gene genealogies . After averaging over many possible genealogical histories , we can calculate the likelihood of the observed data ( i . e . microsatellite markers ) . Given the likelihood function , maximum likelihood approaches can be employed to infer the most likely parameter values . In particular , we are interested in estimating the proportion of symmetric/asymmetric divisions ( α , β ) in the life history of mice . Single-cell genotype data were taken from a previous study [28] . Two mice from a DNA repair deficiency strain ( Mlh1−/− ) with much elevated mutation rates [28] , [29] were sacrificed at two different ages ( day 52 and 340 ) . From each mouse , two crypts were harvested from the mouse colon . Multiple cells ( 4–6 , Table 1 ) were subsequently isolated from each sample and sequenced at a set of micro-satellite markers . Using a two-step mutation model for micro-satellite markers , we first calculated the genetic distances between all sampled cells ( Materials and Methods ) . As shown in Figure 2 , individual cells within an intestinal crypt are monophyletic and are clonally related . Between-crypt divergence rapidly increases with age ( Figure 2B ) , in agreement with previous observations of fast clonal turnover in intestinal crypts [23] . Using the likelihood approach we outlined above , we calculated the likelihood of the data as a function of asymmetric division rate for each crypt ( Figure 2C–2F ) . For example , using the two crypts sampled from day 52 mice , maximum-likelihood estimates for the proportion of asymmetric divisions are 0 . 76 and 0 . 60 respectively ( Figure 2C–2D ) . Interestingly , when we look at stem cells from the older mice ( day 340 ) , maximum-likelihood estimates are 0 , which means that stem cells are all dividing symmetrically ( Figure 2E–F ) . When we combined the data from both crypts at each life stage , maximum-likelihood estimates for proportions of cells dividing asymmetrically at day 52 and 340 are 0 . 76 and 0 . 0 respectively ( Figure 2G–H ) . Our analyses thus suggest that the stem cell populations have changed from largely asymmetric divisions to solely symmetric ones . Even though the point estimates of the asymmetric division rate is rather different , the confidence in the point estimates is not very strong . This is especially true for the day 52 ( Figure 2G ) . In order to assess the uncertainties in the point estimates , a resampling-based nonparametric bootstrap analysis was conducted [37] ( Materials and Methods ) . As shown in Table 1 , the estimates for the asymmetric division rates at these two time points stay quite disparate , even though the confidence intervals overlap with each other . In order to compare these two point estimates rigorously , we explicitly tested the null hypothesis of H0: β52 = = β340 against the alternative hypothesis Ha: β52≠β340 . Since the null hypothesis is a special case of the alternative hypothesis ( i . e . the models are nested ) , a Likelihood Ratio Test ( LRT ) can be employed to ask whether the null hypothesis can be rejected with confidence . After we calculated the associated test statistic , the LRT gives the p-value of 0 . 024 , which is significant at the nominal cut-off of 5% ( Table 2 ) . Statistical significance is also observed when we compare β52–β340 with zero over the bootstrap samples ( Figure S1 , P = 0 . 04 ) . In other words , there is strong statistical evidence that the proportion of cells undergoing asymmetric divisions is very different between day 52 and 340 . In addition to the two-deme model outlined above , we explored a series of more complicated models that reflect various additional aspects of crypt biology . In general , due to complexity of these models , analytical results are much harder to derive , but can be supplemented with computer simulations ( Materials and Methods ) . For example , when we compute the log-likelihood under a model where the population of the transit-amplifying cell pool is age-structured ( Table 2 , Figure S2 ) , the likelihood ratio test shows even stronger evidence of differences in asymmetric division rates between day 52 and 340 ( P = 6 . 3×10−7 ) . This is also true when we explore spatial structures of the intestinal crypt ( P = 1 . 3×10−5 , Table 2 , Figure S2 ) , as well as continuous-time models where waiting times between events are exponentially distributed and population sizes are allowed to fluctuate within a certain range ( P = 0 . 029 , Table 2 , Figure S3 ) . In addition , we also examined possible variations in mutation rates and found that the results stay qualitatively similar ( Materials and Methods , Text S1 ) . In summary , the conclusion that stem cells transition from asymmetric to symmetric division is insensitive to model details .
Using population genetic theory , in particular the coalescent theory , we have drawn an extraordinary dynamic picture of intestinal crypt homeostasis . Compared to earlier lineage-tracing methods which typically do not allow for individual lineage relationships , this branch of theory provides a more detailed picture of stem-cell crypt dynamics . With single-cell data collected from different life stages , we found strong statistical support for a transition from cell asymmetry to population asymmetry during mouse life history . Intuitively , the reason we can observe this discrepancy is that genealogical trees of crypt cell populations will steadily increase in size as the population evolves to establish equilibrium . This is analogous to the founder population effect in population genetics ( Figure 2I ) . The genealogical trees for day 52 ( before equilibrium ) are expected be much shorter than those from later times if asymmetric division parameters are constant . However , the gene trees at day 52 are observed to be larger than those at day 340 ( Figure 2I ) . The discrepancy between expectation and real observation leads to the inference of higher asymmetric division rate at the early life stage , because asymmetric division will slow down stem cell lineage turnovers and increase genealogical tree size . Previous observations revealed that mouse crypt morphogenesis started with a surge of symmetric divisions establishing the pool of stem cells , followed by a transition to predominantly asymmetric divisions that maintain an equilibrium between stem cell self-renewal and differentiation [25] . Our results support the existence of a second transition from mostly asymmetric stem cell divisions to symmetric divisions during intestinal homeostasis ( Figure 3A ) . It remains to be seen whether this phenomenon also occurs in other systems . The population asymmetry found here for day 340 matches previous observations that adult intestinal stem cells are maintained by replacing randomly-lost cells through predominately symmetric divisions of their neighbors ( Figure 3A ) [23] , [24] . This random neutral drift also allows monoclonal turnover observed previously [28] . The stochastic fate determination of stem cells seems to be quite general across many tissue types and species [13] , [14] , [38] . Our observations raise a number of questions about the dynamics of the transition between cell division modes . Stem cell behaviors are often controlled by both internal signals ( e . g . cellular polarity [39] or telomerase activity [40] ) and external factors ( e . g . BMP pathway in the mesenchyme ) [16] , [18] , [41] . What are the relative roles of these intrinsic and extrinsic factors are still not fully understood and therefore the exact molecular mechanisms behind the transition between cell division modes are still unknown . Furthermore , transition timing is also unclear . If paneth cells are responsible for maintaining much of the intestinal niche [42] , the transition might be quite fast since stem niche and stem cells are derived from the same cell cascade where an upstream triggering signal can easily be propagated downstream and the transition can be hastened through a snow-ball-like effect ( Figure 3A ) . On the other hand , if signals from other mesodermal components ( e . g . mesenchyme ) are also contributing to this transition , we might imagine the stem cell and their niche could be changing asynchronously leading to a variety of histories with drastically different paces [43] , [44] . Interestingly , current biological evidences seem to have a tilt towards a fast transition . For example , lineage tracing study looking at the speed of drift towards monoclonality , has found similar rates for mice of age 1 . 5 , 6 . 5 as well as 8 months [23] . In other words , starting from about 45 days , the crypt dynamics could potentially have shifted to largely symmetric divisions . In our model we were treating asymmetric/symmetric parameter as a fixed unknown constant and the coalescent analysis is a retrospective approach looking at the profile of a recent history before the time of observation ( typically within 4N generations , where N is the population size [30] ) . Our results thus reflect a time-average measurement . The fast transition might have lead to the statistical uncertainties for the asymmetric division rate observed for day 52 . Nevertheless , based on our limited simulations with time-varying asymmetric division rates , changes in transition dynamics should leave very different genealogical signatures , where future studies with dense sampling across time will be able to resolve this landscape more precisely ( Figure 3B ) . After all , the pattern observed here seems to suggest that the switch has started not long before day 52 , which is broadly around mouse maturation ( Figure 3 and following sections ) . Why do the crypt cells need to switch to population level asymmetry , which is an apparently more fragile scheme for long-term tissue maintenance [10] ? There are two possible explanations for this transition . One explanation involves acquisition of the population asymmetry driven by adaptive mechanisms . Asymmetric divisions may be a harder task for cells to perform as cell fate determinants all need to be delivered to the two daughter cells according to the two distinct states [5] . In contrast , the easier mode may be symmetric divisions in which the two daughter cells need not be distinguished . The implication for the transition is that as mice grow old , their cells gradually take the easier mode . In addition , since stem cell function often declines with age , population asymmetry might allow stem cells to effectively repair and restore homeostasis – a key adaptation that can increase capacity for repair and increase life-span [45] . Furthermore , adaptive immune response to environmental insults including gut microbiota infection can also possibly contribute to the homeostatic transition [44] , [46] , [47] . In light of this view , cancer and tissue loss , two flip sides of normal cellular dynamics , might result from disruptions of this cell division equilibrium [5] , [48] . On the other hand , there might also exist a “passive” explanation for this transition . The observed progression can simply be a by-product of natural selection . When a single gene has multiple functions , some of the functions will be beneficial to the organism , while others might be detrimental . Most importantly , when the advantageous gain outweights the deleterious costs , the target gene can still be selected ( i . e . antagonistic pleiotropy ) [49] . During the course of evolution , a gene with multiple effects will be strongly optimized for its function before reproduction , and as a consequence also produce deleterious effects in later life [49] . In this light it is notable that the transition in cell division mode occurs roughly at the time of sexual reproduction ( Figure 3 ) . Many proteins involved in asymmetric cell divisions also function as tumor suppressor genes [5] , [50] , [51] . The loss of asymmetric division might thus simply be driven by the increasing need for active tumor suppressors [52] . The life-history of stem cell division is not yet fully discernable from our results ( e . g . Figure 3 ) . For example , our estimate of the proportion of asymmetric divisions for day 52 is still quite high , even though we have evidence that the proportions at days 52 and 340 are significantly different . Information collected from only two time points also prevents sophisticated models where many of the parameters can be treated as time-varying variables rather than fixed constants . Future studies with denser serial sampling together with larger number of crypts might be able to draw a more concrete picture of the crypt homeostasis . Until now , asymmetric divisions were thought to be rarer in vertebrates than invertebrates [5] . However , new evidence for the existence of this mode is starting to accumulate for a few tissue types such as the central nervous system [53] , skin [54] , [55] , the hematopoietic system [51] , [56] , [57] as well as the intestinal crypts [26] . Since the mechanisms controlling stem cell symmetric and asymmetric divisions are often conserved across the tree of life [16] , the pattern observed here for the mouse intestine is very likely to be quite general . With the advances in genomic technology , in particular whole-genome single-cell sequencing [58] , [59] , we should be able to reveal a more lively cellular orchestra across a wide range of organ types and species , each with its own mechanism and equilibria .
Since population sizes are relatively constant in the intestinal crypt and because type III divisions do not change the number of stem cell descendants , the proportions of two types of symmetric divisions ( type I and II ) have to be balanced and are set to be equal so that the total number of cells is maintained ( Figure 1A ) . In other words , 1-α-β ( proportion of type II divisions ) is be equal to α ( proportion of type I divisions ) . In each generation , fraction α of the stem cells divides symmetrically , each cell giving rise to two stem cells ( type I ) ; fraction β divides asymmetrically ( type III ) and fraction α divides symmetrically with each cell producing two differentiated cells ( type II ) . Differentiated cells from both asymmetric and symmetric divisions migrate to the transit-amplifying cell pool . Since there is a constant extrusion of cells from the crypt into the villi , we capture the dynamics of the transit-amplifying cell pool by allowing only a certain proportion of the cells to participate in reproduction for the next generation . The remaining cells are extruded out of the deme 2 . We set the population size of stem cells ( population 1 ) to N1 and that of transit-amplifying cells ( population 2 ) to N2 . In each generation there are N1 cells migrating to the transit-amplifying cell pool . In the transit-amplifying cell population , fraction γ of N2 cells divide once and give rise to two descendant cells . The remaining ( 1−γ ) ×N2 cells are extruded outside of the population 2 . The value of γ is set to ( N2-N1 ) /2N2 such that the population size in population 2 is constant . Based on previous observations in mouse colon crypts [60] , [61] , we set the population size N1 and N2 to be 15 and 185 . Population sizes such as 250 ( 15 stem and 235 non-stem cells ) are also tried and results stay similar . Given the number of cells we sampled in two demes , the process of going backwards in time and finding common ancestors can be modeled as a Markov Chain . The state transitions are given by combinations of individual coalescence or migration events . For two randomly sampled lineages , the expected time to their most recent common ancestor ( MRCA ) can be computed by directly simulating from a Markov Chain following the appropriate state transitions . The expected value for the time to MRCA for two lineages can also be derived analytically using a first-step analysis of a given Markov Chain ( Figure 1D ) . In Text S1 , we presented the details of the transition probabilities between state spaces for the Markov Chain . The Markov Chain calculation assumes that data are collected from a stationary process . Based on our simulations , we find that , for most of the parameter values , stationary distributions have been reached by day 340 . However , this does not appear to hold for day 52 ( Figure S4 and Text S1 ) . In these cases , forward simulations are used to generate genealogical histories from Pr ( G|θ ) . To achieve this , we simulated population histories with two phases: morphogenesis and homeostasis ( Figure 3 ) . This scenario was chosen to reflect experimental evidence as well as predictions from the Bang-Bang control theory [25] . In this process , the intestinal crypt is founded by first creating N1 stem cells , followed by a series of asymmetric divisions to generate the transit-amplifying cell population . After crypt morphogenesis , populations follow the dynamic process described in the previous section . At the end of our simulations , a random subset of cells is sampled and their genealogical history is recorded ( see Text S1 for details ) . To estimate the number of generations leading to day 52 , we tried a series of approaches . Since we know that crypt morphogenesis starts around post-natal day 7 for mice [62] , [63] and stem cells divide every 22 hours [24] , [25] , postnatal day 52 corresponds to about generation 50 starting from crypt morphogenesis ( roughly 42 generations after crypt formation because crypt morphogenesis takes around 8 generations ) . Because the exact cell generation number will be a random variable around this mean value , we used various forms of random distributions ( e . g . beta distribution or truncated normal distribution ) to model the generation number . Conditioning on the generation number , gene genealogies in the forward simulation at the corresponding generation number is sampled for the Pr ( G|θ ) at day 52 . ( see Text S1 for details ) . Data were taken from a previously-published study [28] where multiple single-cell genotypes were sequenced from a DNA-repair deficiency mouse strain ( Mlh1−/− ) . This strain has elevated mutation rates , allowing for enough informative mutations to enable our analyses [28] , [29] . Two mice at day 52 and 340 were sacrificed . From each mouse , two crypts were harvested from the mouse colon . Multiple cells ( 4–6 , Table 1 ) were then isolated and sequenced at a set of micro-satellite markers . In total , 150 microsatellite markers were genotyped , from which only markers with successful genotyping information from at least two cells were extracted for this work ( 70–90 markers ) . We also tried a series of alternative models to explore other aspects of stem cell dynamics . In the age structure model ( Figure S2A ) , multiple demes exist in the transit amplifying cell population . Each of these demes corresponds to cells with different ages ( number of divisions since leaving the stem cell deme ) . Cells hitting an age limit will be extruded out of the crypt . In the spatial model ( Figure S2B ) , multiple spatial demes corresponding to cells at different localities are constructed . Demes in the transit amplifying cell population have different probability of being extruded from the crypt depending on their physical locations . In the continuous-time models , the time to the next event ( waiting time ) is exponentially distributed with the intensity parameter specified by the cell division rate ( λ ) . The time to the next event for n cells is exponentially distributed with rate nλ . Given the time to the next event , the exact cell that experiences this event is randomly picked among the n cells following statistical properties of the Poisson Process [64] . When a stem cell is picked , the possible events are type I/II/III stem cell divisions , depending on the values of α and β . If a transit-amplifying cell is picked , it either divides or is extruded out of the crypt with the associated probability . In this simulation , we allowed the stem cell population to fluctuate within a size range ( size from 10 to 20 with a mean of 15 ) , a feature we implemented using rejection sampling [64] . The details of these models are presented in Text S1 . In general , analytical results from these models are much harder to derive , but computational simulations can be conducted to sample gene genealogies from the random process . Likelihood calculations follow the same procedures as previous models after sampling gene genealogies from Pr ( G| θ ) . Given a gene tree , the likelihood of the observed data can be computed from the tip of the tree towards the root using the pruning algorithm [35] . For the microsatellite loci , we used a two-step mutational model where repeat number j can mutate to j±1 and j±2 . Previous empirical measurements have found that the average mutation rate is 0 . 01 per site per generation and size-two transitions ( j±2 ) are happening at 1/7 the frequency of size-one mutations ( j±1 ) [28] . We also explored other mutational models and they did not affect our conclusions ( data not shown ) . In order to further explore the possibility of mutation rate variation , we used various forms of the beta distribution to capture uncertainty in the mutation rate . Since the mutation rate measured from previous studies is 0 . 01 per site per generation , we adopted beta distributions with different shape parameters , but transformed to take values between 0 . 0075 and 0 . 0125 . The likelihood of the data can be calculated by partitioning the mutation distribution into discrete bins and taking the weighted sum of individual likelihoods calculated at discrete values of mutation rates [65] ( see Text S1 ) . Unweighted Pair Group Method with Arithmetic mean ( UPGMA ) tree [66] was built using functions implemented in the APE ( Analyses of Phylogenetics and Evolution ) library [67] within the R package ( http://www . R-project . org ) . The pairwise distances between the cells were calculated using the mutation matrix specified in the previous section . The likelihood of the data as a function of the underlying parameters can be computed in a Monte Carlo fashion as in Equation 3 . We sampled 20 , 000 gene genealogies to compute the log-likelihood for each combination of parameter values . A non-parametric bootstrap test was conducted by resampling microsatellite markers from the original dataset with replacement ( 100 replicates ) and re-running the analyses . In the likelihood ratio test , the maximum-likelihood values under the null and alternative model respectively are extracted . Likelihood Ratio Test ( LRT ) is conducted by comparing twice the log-likelihood ratio to the chi-square distribution with one degree of freedom , since the two models we compared were nested with the alternative having one extra parameter .
|
In multi-cellular organisms , there is a static equilibrium maintaining cells of various forms . This homeostasis is achieved by an exquisite balance between stem cell proliferation and differentiation . Understanding how different species and organ types maintain this dynamic equilibrium has been an interesting question for both evolutionary and developmental biologists . Using population genetic theory together with previously published single-cell sequencing data collected from mouse intestinal crypts at two points in development , we have revealed a dynamic picture of stem cell renewal in intestinal crypts . We found that intestinal equilibrium is maintained at the single-cell level through predominantly asymmetric stem cell divisions at early life stages , but progressively switches to a population level homeostasis with only symmetric divisions as the mouse matures to adulthood . This dynamic process , likely to be conserved across species , has important developmental and evolutionary implications in understanding how adult tissues maintain their homeostasis integrating lifelong trade-offs between intrinsic and extrinsic factors .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology",
"stem",
"cells",
"evolutionary",
"biology",
"genetics",
"population",
"genetics",
"biology",
"evolutionary",
"genetics",
"genetics",
"and",
"genomics",
"evolutionary",
"developmental",
"biology"
] |
2013
|
Age-Dependent Transition from Cell-Level to Population-Level Control in Murine Intestinal Homeostasis Revealed by Coalescence Analysis
|
Neurons process information via integration of synaptic inputs from dendrites . Many experimental results demonstrate dendritic integration could be highly nonlinear , yet few theoretical analyses have been performed to obtain a precise quantitative characterization analytically . Based on asymptotic analysis of a two-compartment passive cable model , given a pair of time-dependent synaptic conductance inputs , we derive a bilinear spatiotemporal dendritic integration rule . The summed somatic potential can be well approximated by the linear summation of the two postsynaptic potentials elicited separately , plus a third additional bilinear term proportional to their product with a proportionality coefficient . The rule is valid for a pair of synaptic inputs of all types , including excitation-inhibition , excitation-excitation , and inhibition-inhibition . In addition , the rule is valid during the whole dendritic integration process for a pair of synaptic inputs with arbitrary input time differences and input locations . The coefficient is demonstrated to be nearly independent of the input strengths but is dependent on input times and input locations . This rule is then verified through simulation of a realistic pyramidal neuron model and in electrophysiological experiments of rat hippocampal CA1 neurons . The rule is further generalized to describe the spatiotemporal dendritic integration of multiple excitatory and inhibitory synaptic inputs . The integration of multiple inputs can be decomposed into the sum of all possible pairwise integration , where each paired integration obeys the bilinear rule . This decomposition leads to a graph representation of dendritic integration , which can be viewed as functionally sparse .
For information processing , a neuron receives and integrates thousands of synaptic inputs from its dendrites and then induces the change of its membrane potential at the soma . This process is usually known as dendritic integration [1]–[3] . The dendritic integration of synaptic inputs is crucial for neuronal computation [2]–[4] . For example , the integration of excitatory and inhibitory inputs has been found to enhance motion detection [5] , regularize spiking patterns [6] , and achieve optimal information coding [7] in many sensory systems . They have also been suggested to be able to fine tune information processing within the brain , such as the modulation of frequency [8] and the improvement of the robustness [9] of gamma oscillations . In order to understand how information is processed in neuronal networks in the brain , it is important to understand the computational rules that govern the dendritic integration of synaptic inputs . Dendritic integration has been brought into focus with active experimental investigations ( see reviews [1] , [10] and references therein ) . There have also been many theoretical developments based on physiologically realistic neuron models [11] , [12] . Among those works , only a few investigate quantitative dendritic integration rules for a pair of excitatory and inhibitory inputs [3] , [13] and there has yet to be an extensive investigation of the integration of a pair of excitatory inputs or a pair of inhibitory inputs . In this work , we propose a precise quantitative rule to characterize the dendritic integration for all types of synaptic inputs and validate this rule via realistic neuron modeling and electrophysiological experiments . We first develop a theoretical approach to quantitatively characterize the spatiotemporal dendritic integration . Initially , we introduce an idealized two-compartment passive cable model to understand the mathematical structure of the dendritic integration rule . We then verify the rule by taking into account the complicated dendritic geometry and active ion channels . For time-dependent synaptic conductance inputs , we develop an asymptotic approach to analytically solve the cable model . In this approach , the membrane potential is represented by an asymptotic expansion with respect to the input strengths . Consequently , a hierarchy of cable-type equations with different orders can be derived from the cable model . These equations can be analytically solved order by order using the Green's function method . The asymptotic solution to the second order approximation is shown to be in excellent agreement with the numerical solutions of the original cable model with physiologically realistic parameters . Based on our asymptotic approach , we obtain a new theoretical result , namely , a nonlinear spatiotemporal dendritic integration rule for a pair of synaptic inputs: the summed somatic potential ( SSP ) can be well approximated by the summation of the two postsynaptic potentials and elicited separately , plus an additional third nonlinear term proportional to their product , i . e . , ( 1 ) The proportionality coefficient encodes the spatiotemporal information of the input signals , including the input locations and the input arrival times . In addition , we demonstrate that the coefficient is nearly independent of the input strengths . Because the correction term to the linear summation of and takes a bilinear form , we will refer to the rule ( 1 ) as the bilinear spatiotemporal dendritic integration rule . In the remainder of the article , unless otherwise specified , all the membrane potentials will be referred to those measured at the soma . We note that our bilinear integration rule is consistent with recent experimental observations [3] . In the experiments [3] , the rule was examined at the time when the excitatory postsynaptic potential ( EPSP ) measured at the soma reaches its peak for a pair of excitatory and inhibitory inputs elicited concurrently . We demonstrate that our bilinear integration rule is more general than that in Ref . [3]: ( i ) our rule holds for a pair of excitatory and inhibitory inputs that can arrive at different times; ( ii ) our rule is also valid at any time and is not limited to the peak time of the EPSP; ( iii ) our rule is general for all types of paired synaptic input integration , including excitatory-inhibitory , excitatory-excitatory and inhibitory-inhibitory inputs . Our bilinear integration rule is derived from the two-compartment passive cable model . We then validate the rule in a biologically realistic pyramidal neuron model with active ion channels embedded . The simulation results from the realistic model are consistent with the rule derived from the passive cable model . We further validate the rule in electrophysiological experiments in rat hippocampal CA1 pyramidal neurons . All of our results suggest that the form of the bilinear integration rule is preserved in the presence of active dendrites . As mentioned previously , there are thousands of synaptic inputs received by a neuron in the brain . We therefore further apply our analysis to describe the dendritic integration of multiple synaptic inputs . We demonstrate that the spatiotemporal dendritic integration of all synaptic inputs can be decomposed into the sum of all possible pairwise dendritic integration , and each pair obeys the bilinear integration rule ( 1 ) , i . e . , ( 2 ) where denotes the SSP , denotes the individual EPSP , denotes the individual inhibitory postsynaptic potential ( IPSP ) , , , and are the corresponding proportionality coefficients with superscripts denoting the index of the synaptic inputs . We then confirm the bilinear integration rule ( 2 ) numerically using realistic neuron modeling . The decomposition of multiple inputs integration in rule ( 2 ) leads to a graph representation of the dendritic integration . Each node in the graph corresponds to a synaptic input location , and each edge connecting two nodes represents the bilinear term for a pair of synaptic inputs given at the corresponding locations . This graph evolves with time , and is all-to-all connected when stimuli are given at all synaptic sites simultaneously . However , based on simulation results and experimental observations , we can estimate that there are only a small number of activated synaptic integration , or edges in the graph , within a short time interval . Therefore , the graph representing the dendritic integration can indeed be functionally sparse . Finally , we comment that , in general , it is theoretically challenging to analytically describe the dynamical response of a neuron with dendritic structures under time-dependent synaptic conductance inputs . One simple approach to circumvent this difficulty is to analyze the steady state of neuronal input-output relationships by assuming that both the synaptic conductance and the membrane potential are constant [3] , [12] . Such analyses can be applied to study dendritic integration , but they usually oversimplify the description of the spatial integration , and fail to describe the temporal integration . Another approach to circumvent the difficulty is to study the cable model [14] , [15] analytically or numerically . For the subthreshold regime , in which voltage-gated channels are weakly activated , the dendrites can be considered as a passive cable . Along the cable , the membrane potential is linearly dependent on injected current input . This linearity enables one to use the Green's function method to analytically obtain the membrane potential with externally injected current . In contrast , the membrane potential depends nonlinearly on the synaptic conductance input [12] . This nonlinearity greatly complicates mathematical analyses . Therefore , in order to solve the cable model analytically , one usually makes the approximation of constant synaptic conductance [16] , [17] . The approximation can help investigate some aspects of dendritic integration , however , the approximation in such a case is not sufficiently realistic because the synaptic conductances in vivo are generally time-dependent . On the other hand , one can study the dendritic integration in the cable model numerically . The compartmental modeling approach [14] enables one to solve the cable model with time-dependent synaptic inputs . This approach has been used to investigate many aspects of dendritic integration . For instance , it was discovered computationally that dendritic integration of excitatory inputs obeys a certain qualitative rule , i . e . , EPSPs are first integrated nonlinearly at individual branches before summed linearly at the soma [18] , [19] , which was verified later in experiments [20] , [21] . Clearly , the computational approach can help gain insights into various phenomena of spatiotemporal dynamics observed at the dendrites , however , a deep , comprehensive understanding often requires analytical approaches . Note that this point has also been emphasized in Ref . [22] . Here , our analytical asymptotic method can solve the cable model with time-dependent synaptic inputs analytically and reveal a precise quantitative spatiotemporal dendritic integration rule , as will be further illustrated below .
We begin to study the spatiotemporal dendritic integration of a pair of excitatory and inhibitory inputs . An analytical derivation of the bilinear integration rule is described in the section of Derivation of the Rule . The details of the cable model used in the derivation can be found in the section of Materials and Methods . The validation of the bilinear integration rule using the realistic neuron modeling and electrophysiological experiments is described in the section of Validation of the Rule . The spatial dependence of the coefficient in the rule is described in the section of Spatial Dependence of . So far we have addressed the dendritic integration for a pair of excitatory and inhibitory inputs . A natural question arises: how does a neuron integrate a pair of time-dependent synaptic conductance inputs with identical type ? The dendritic integration of excitatory inputs has been extensively investigated in experiments ( reviewed in Ref . [1] ) , yet a precise quantitative characterization is still lacking . According to our idealized cable model , given a pair of excitatory inputs with input strengths and at locations and and at times and , the dynamics of the membrane potential on the dendrite is governed by the following equation: ( 22 ) with the initial and boundary conditions the same as given in Equations ( 4 ) – ( 6 ) . Similarly , we can represent its solution as an asymptotic series and solve it order by order to obtain the following bilinear integration rule: ( 23 ) where and are EPSPs induced by two individual excitatory inputs , and is the SSP when the two excitatory inputs are present . Similar to the case of a pair of excitatory and inhibitory inputs , the shunting coefficient only depends on the excitatory input locations and the input time difference . It does not depend on the EPSPs' amplitudes . Here will still be referred to as a shunting coefficient because the origin of the nonlinear integration for the paired excitatory inputs is exactly the same as that for the paired excitatory and inhibitory inputs from the passive cable model . The bilinear integration rule ( 23 ) is found to be consistent with the numerical results obtained using the same realistic pyramidal neuron model as the one used in the section of Bilinear Rule for E–I Integration . For a pair of excitatory inputs with their locations fixed on the dendritic trunk , the rule holds when the amplitude of each EPSP is less than . For the case of concurrent inputs , at the time when one of the EPSPs reaches its peak valueis found to be linearly dependent of , as shown in Fig . 6A . This linear relationship indicates is independent of the amplitudes of the two EPSPs . In addition , as shown in Fig . 6B , the bilinear integration rule is numerically verified in the time interval , for , within which the amplitude of EPSPs are relatively large . For the case of nonconcurrent inputs , the bilinear integration rule is also numerically verified in the same way , as shown in Fig . 6C–D . In addition , we find that when the input strengths become sufficiently strong so as to make the depolarized membrane potential too large , i . e . , there is a deviation from the bilinear integration rule ( 23 ) . This deviation can be ascribed to the voltage-gated ionic channel activities in our realistic pyramidal neuron model . After blocking the active channels , the rule becomes valid with a different value of for large EPSPs amplitudes , as shown in Fig . 7 . However , we note that , regardless of input strengths , the amplitude of SC is always two orders of magnitude smaller than the amplitude of SSP . Therefore , the integration of two excitatory inputs can be naturally approximated by the linear summation of two individual EPSPs , i . e . . We then perform electrophysiological experiments with a pair of excitatory synaptic inputs to confirm the linear summation . As expected , this linear summation is also observed in our experiments for both concurrent and nonconcurrent input cases , as shown in Fig . 6E and 6F , respectively . Note that , the linear summation is also consistent with experimental observations as reported in Ref . [24] . Similarly , for a pair of inhibitory inputs , we can arrive at the following bilinear integration rule from the cable model: ( 24 ) where and are IPSPs induced by two individual inhibitory inputs , and is the SSP when the two inhibitory inputs are present . Here , is the shunting coefficient that is independent of the IPSPs amplitudes but is dependent on the input time difference and input locations . The above bilinear integration rule ( 24 ) is consistent with our numerical results using the realistic pyramidal neuron model , as shown in Fig . 8A–D . Our electrophysiological experimental observations further confirm this rule , as shown in Fig . 8E–H . In the previous sections , we have discussed the integration of a pair of synaptic inputs . In vivo , a neuron receives thousands of excitatory and inhibitory inputs from dendrites [2] . Therefore , we now address the question of whether the integration rule derived for a pair of synaptic inputs can be generalized to the case of multiple inputs . Our theoretical analysis shows that , for multiple inputs , the SSP can be approximated by the linear sum of all individual EPSPs and IPSPs , plus the bilinear interactions between all the paired inputs with shunting coefficients , , and respectively ( the superscript labels the synaptic inputs ) , i . e . , ( 25 ) We next validate the rule ( 25 ) using the realistic pyramidal neuron model . It has been reported that , for a CA1 neuron , inhibitory inputs are locally concentrated on the proximal dendrites while excitatory inputs are broadly distributed on the entire dendrites [25] . Based on this observation , we randomly choose 15 excitatory input locations and 5 inhibitory input locations on the model neuron's dendrites ( Fig . 9A ) . In the simulation , all inputs are elicited starting randomly from to . In order to compare Equation ( 25 ) with the SSP simulated in the realistic neuron model , we first measure , , and pair by pair for all possible pairs . We then record all membrane potential traces and induced by the corresponding individual synaptic inputs . Our results show that the SSP measured from our simulation is indeed given by the bilinear integration rule ( 25 ) , as shown in Fig . 9B and 9C . In contrast , the SSP in our numerical simulation deviates significantly from the linear summation of all individual EPSPs and IPSPs . According to our bilinear integration rule ( 25 ) , the dendritic integration of multiple synaptic inputs can be decomposed into the summation of all possible pairwise dendritic integration . Therefore , we can map dendritic computation in a dendritic tree onto a graph . Each dendritic site corresponds to a node in the graph and the corresponding shunting component is mapped to the weight of the edge connecting the two nodes . We refer to such a graph as a dendritic graph . The dendritic graph is an all-to-all connected graph if all stimuli are given concurrently ( Fig . 10A ) . However , the dendritic integration for all possible pairs of synaptic inputs is usually not activated concurrently in realistic situations . For instance , if the arrival time difference between two inputs is sufficiently large , there is no interaction between them . The activated level of the nonlinear dendritic integration for a pair of synaptic inputs can be quantified by the SC amplitude—the weight of the edge in the graph . The simulation result shows that the number of activated edges at any time is relatively small on the dendritic graph ( Fig . 10B–D ) , compared with the total number of edges on the all-to-all connected graph ( Fig . 10A ) . Therefore , for the case of a hippocampal pyramidal neuron , the dendritic graph could be functionally sparse in time . The functional sparsity of a dendritic graph may also exist in neocortical pyramidal neurons . In vivo , a cortical pyramidal neuron receives about synaptic inputs [26] . Most of them are from other cortical neurons [27] , [28] , which typically fire about 10 spikes per second in awake animals [29] , [30] . Thus , the neuron can be expected to receive synaptic inputs per second . The average number of synaptic inputs within ( membrane potential time constants in vivo ) is . The number of activated dendritic integration pairs within the interval is , which is relatively small compared with the total possible synaptic integration pairs . Therefore , the activated integrations or edges in the dendritic graph within a short time window can be indeed functionally sparse ( ) . In general , the neuronal firing rates vary across different cell types , cortical regions , brain states and so on . Therefore , based on the above estimate , in an average sense , the graph of dendritic integration is functionally sparse .
Our bilinear dendritic integration rule ( 21 ) is consistent with the rule previously reported [3] , but is more general in the following aspects: ( i ) Our dendritic integration rule holds at any time and is not limited to the time when the EPSP reaches its peak value . ( ii ) The rule holds when the two inputs are even nonconcurrent . This situation often occurs because the excitatory and inhibitory inputs may not always arrive at precisely the same time . ( iii ) The form of the rule can be extended to describe the integration between a pair of excitatory inputs , a pair of inhibitory inputs , and even multiple inputs of mixed-types . The spatiotemporal information of synaptic inputs interaction is coded in the shunting coefficient , which is a function of the input locations and input arrival time difference . Our bilinear integration rule holds in the subthreshold regime for a large range of membrane potential . When we derive the bilinear rule from the passive cable model , we assume that the input strengths or the amplitudes of membrane potentials require to be small . This assumption forms the basis of the asymptotic analysis , because the second order asymptotic solutions of EPSP , IPSP and SSP converge to their exact solutions as the asymptotic parameters and ( denoting the excitatory and inhibitory input strengths ) approach zero . In general , in the passive cable model , the bilinear rule will be more accurate for small amplitudes of EPSPs and IPSPs than large amplitudes . Importantly , the assumption holds naturally that in the physiological regime when EPSP amplitude is less than 6mV and IPSP amplitude is less than -3mV , and are small . However , even for EPSP amplitude close to the threshold , i . e . , 10mV , which is unusually large physiologically , we can show that the second order asymptotic solution can still well approximate the EPSP with a relative error less than 5% . Thus the bilinear rule is still valid for large depolarizations near the threshold . The validity of the bilinear rule for large membrane potentials is also confirmed in both simulations and experiments . In particular , in the analysis of our experimental data , to validate the bilinear rule , we have already included all the data when the EPSP amplitude is below and close to the threshold because we have only excluded those data corresponding to the case when a neuron fires . Our bilinear dendritic integration rule ( 21 ) is derived from the passive cable model . However , the simulation results and the experimental observations demonstrate that the form of dendritic integration is preserved for active dendrites . Additional simulation results show that for the same input locations , the shunting coefficients are generally larger on the active dendrites than those on the passive dendrites with all active channels blocked . We also note that the value of in simulation is different from the value measured in experiments . This difference may arise from the fact that some parameters of the passive membrane properties , such as the membrane leak conductance , may not be exactly the same as those in the biological neuron , and we have only used a limited set of ion channels in simulation compared with those in the biological neuron . In addition , the input locations in the simulation and the experiments are different , which may also contribute to this derivation . However , the bilinear form is a universal feature in both simulation and experiment . By fixing excitatory input location while varying inhibitory input location , our model exhibits that there exists a region in the distal dendritic trunk within which the shunting inhibition can be more powerful , i . e , a larger , than in proximal dendrites . This result is consistent with what is reported in Ref . [31] . Compared with Ref . [31] , our work provides a different perspective of dendritic computation . In their work , the multiple inhibitory inputs can induce a global shunting effect on the dendrites . However , if we focus on the shunting effect only at the soma instead of the dendrites , our theory shows that all the interactions among multiple inputs can then be decomposed into pairwise interactions , as described by the bilinear integration rule ( 25 ) . In addition , in this work , we focus on the somatic membrane potential that is directly related to the generation of an action potential . However , it is also important to investigate the local integration of membrane potentials measured at a dendritic site instead of that measured at the soma . Asymptotic analysis of the cable model can show that our bilinear integration rule is still valid for the description of the integration on the dendrites . On the dendrites , the broadly distributed dendritic spines with high neck resistances [32] , [33] will filter a postsynaptic potential to a few millivolts on a branch [34] , [35] . Within this regime our bilinear integration rule is valid . Note that our rule may fail to capture the supralinear integration of synaptic inputs measured on the dendrites during the generation of a dendritic spike [36] . However , if the integration is measured at the soma , our rule remains valid even when there is a dendritic spike induced by a strong excitatory input and an inhibitory synaptic input on different branches [3] . The bilinear integration rule ( 25 ) can help improve the computational efficiency in a simulation of neuronal network with dendritic structures . By our results , once the shunting coefficients for all pairs of input locations are measured , we can predict the neuronal response at the soma by the bilinear integration rule ( 25 ) . By taking advantage of this , one can establish library-based algorithms to simulate the membrane potential dynamics of a biologically realistic neuron . An example of a library-based algorithm can be found in Ref . [37] . To be specific , based on the full simulation of a realistic neuron model , we can measure the time-dependent shunting coefficient as a function of the arrival time difference and input locations for all possible pairs of synaptic inputs and record them in a library in advance . For a particular simulation task , given the specific synaptic inputs on the dendrites , we can then search the library for the corresponding shunting coefficients to compute the neuronal response according to the bilinear integration rule ( 25 ) directly . In such a computational framework , one can avoid directly solving partial differential equations that govern the spatiotemporal dynamics of dendrites and greatly reduces the computational cost for large-scale simulations of networks of neurons incorporating dendritic integration .
The animal-use protocol was approved by the Animal Management Committee of the State Key Laboratory of Cognitive Neuroscience & Learning , Beijing Normal University ( Reference NO . : IACUC-NKLCNL2013-10 ) . We consider an idealized passive neuron whose isotropic spherical soma is attached to an unbranched cylindric dendrite with finite length and diameter . Each small segment in the neuron can be viewed as an RC circuit with a constant capacitance and leak conductance density [11] , [38] . The current conservation within a segment on the dendrite leads to ( 26 ) where is the membrane potential with respect to the resting potential on the dendrite , is the membrane capacitance per unit area , and is the leak conductance per unit area . Here , is the synaptic current given by: ( 27 ) where and are excitatory and inhibitory synaptic conductance per unit area and and are their reversal potentials , respectively . When excitatory inputs are elicited at dendritic sites and inhibitory inputs are elicited at dendritic sites , we have ( 28 ) where . For a synaptic input of type , is the input strength of the input at the location , is the arrival time of the input at the location , is the input location . The unitary conductance is often modeled as ( 29 ) with the peak value normalized to unity by the normalization factor , and with and as rise and decay time constants , respectively [38] . Here is a Heaviside function . The axial current can be derived based on the Ohm's law , ( 30 ) where is the axial resistivity . Taking the limit , Equation ( 26 ) becomes our unbranched dendritic cable model , ( 31 ) In particular , for a pair of excitatory and inhibitory inputs with strength and received at and , and at time and , respectively , we have ( 32 ) Similarly , for a pair of excitatory or inhibitory inputs with strengths and received at and , and at time and ( ) , respectively , we have ( 33 ) For the boundary condition of the cable model [Equation ( 31 ) ] , we assume one end of the dendrite is sealed: ( 34 ) For the other end connecting to the soma , which can also be modeled as an RC circuit , by the law of current conservation , we have ( 35 ) where is the somatic membrane area , and is the somatic membrane potential . The dendritic current flowing to the soma , , takes the form of Equation ( 30 ) at . Because the membrane potential is continuous at the connection point ( 36 ) we arrive at the other boundary condition at : ( 37 ) For a resting neuron , the initial condition is simply set as ( 38 ) In the absence of synaptic inputs , Equation ( 31 ) is a linear system . Using a impulse input , its Green's function can be obtained from ( 39 ) with the following boundary conditions and initial condition , For simplicity , letting , , , the solution of Equation ( 39 ) can be obtained from the following system , ( 40 ) with rescaled boundary and initial conditions , where . Taking the Laplace transform of Equation ( 40 ) , we obtain ( 41 ) Combining the two boundary conditions ( is thus eliminated ) , we have ( 42 ) where ( 43 ) whose denominator is denoted as for later discussions . For the inverse Laplace transform , we need to deal with singular points that are given by the roots of . It can be easily verified that these singularities are simple poles and is analytic at infinity . Then can be written as ( 44 ) where is a constant coefficient in the complex domain , and are the singular points . Then taking the inverse Laplace transform of Equation ( 44 ) , we obtain ( 45 ) Now we only need to solve and in Equation ( 45 ) to obtain the Green's function of Equation ( 40 ) . We solve the singular points first . Defining yields ( 46 ) whose roots can be determined numerically . There are solutions for with for and Next , to determine the factors we use the residue theorem for integrals . For a contour that winds in the counter-clockwise direction around the pole and that does not include any other singular points , the integral of on this contour is given by ( 47 ) Using Equations ( 42–44 ) and ( 47 ) , we obtain ( 48 ) where ( 49 ) for . The solution of the original Green's function for Equation ( 39 ) can now be expressed as ( 50 ) We first consider the case when a pair of excitatory and inhibitory inputs are received by a neuron . Similar results can be obtained for a pair of excitatory inputs and a pair of inhibitory inputs . For the physiological regime ( the amplitude of an EPSP being less than and the amplitude of an IPSP being less than ) , the corresponding required input strengths and are relatively small . Therefore , given an excitatory input at location and time , and an inhibitory input at location and time , we represent as an asymptotic series in the powers of and , ( 51 ) Substituting Equation ( 51 ) into the cable equation ( 31 ) , order by order , we obtain a set of differential equations . For the zeroth-order , we have ( 52 ) Using the boundary and initial conditions [Equations ( 34 ) , ( 37 ) , and ( 38 ) ] , the solution is simply ( 53 ) For the first order of excitation , we have ( 54 ) With the help of Green's function , the solution can be expressed as ( 55 ) here ‘’ denotes convolution in time . For the second order of excitation , we have ( 56 ) Because is given by Equation ( 55 ) , the solution of Equation ( 56 ) is ( 57 ) Similarly , we can have the first and second order inhibitory solutions , ( 58 ) ( 59 ) For the order of , we have ( 60 ) whose solution is obtained as follows , ( 61 ) For the numerical simulation of the two-compartment passive cable model [Equation ( 3 ) ] , the Crank-Nicolson method [39] was used with time step and space step . Parameters in our simulation are within the physiological regime [3] , [12] with , , , , , , , . , , , and . The time constants here were chosen to be consistent with the conductance inputs in the experiment [3] . The realistic pyramidal model is the same as that in Ref . [3] . The morphology of the reconstructed pyramidal neuron includes 200 compartments and was obtained from the Duke-Southampton Archive of neuronal morphology [40] . The passive cable properties and the density and distribution of active conductances in the model neruon were based on published experimental data obtained from hippocampal and cortical pyramidal neurons [18] , [19] , [34] , [41]–[50] . We used the NEURON software Version 7 . 3 [51] to simulate the model with time step . The experimental measurements of summation of EPSPs or IPSPs in single hippocampal CA1 pyramidal cells in the acute brain slice followed a method described in Ref . [3] , with some modifications . A brief description of modified experimental procedure is as follows . Acute hippocampal slices ( thick ) were prepared from Sprague Dawley rats ( postnatal day 14–16 ) , using a vibratome ( VT1200 , Leica ) . The slices were incubated at 34°C for 30 min before transferring to a recording chamber perfused with the aCSF solution ( 2ml/min; 30–32°C ) . The aCSF contained ( in mM ) 125 NaCl , 3 KCl , 2 CaCl2 , 2 MgSO4 , 1 . 25 NaH2PO4 , 1 . 3 sodium ascorbate , 0 . 6 sodium pyruvate , 26 NaHCO3 , and 11 D-glucose , and was saturated with gas containing 95% O2 and 5% CO2 ( pH 7 . 4 ) . Whole-cell recording was made from the soma of CA1 pyramidal cells using glass micropipettes under an upright microscope ( BX51WI , Olympus ) equipped with the DIC optics and an infrared camera ( IR-1000E , DAGE-MTI ) . The intra-micropipette solution contained ( in mM ) 145 K-gluconate , 5 KCl , 10 HEPES , 10 disodium phosphocreatine , 4 Mg2ATP , 0 . 3 Na2GTP , and 0 . 2 EGTA ( pH 7 . 3 ) , together with fluorescent dye Alexa Fluor 488 ( , Invitrogen ) to visualize the dendritic trees . Pipette resistance was about 3–4 MΩ , and the access resistance during the whole-cell recording was normally less than 20 MΩ . The same method for micro-iontophoretic application of extracelluar glutamate or GABA at the apical dendrite of CA1 pyramidal cells was used to elicit rapid membrane depolarizations ( EPSPs ) and hyperpolarizations ( IPSPs ) . For all three experimental configurations ( EPSP-IPSP , EPSP-EPSP and IPSP-IPSP summation ) , two micro-iontophoretic pipettes were placed at dendritic locations and from the soma , respectively , in particular for the EPSP-IPSP summation GABA iontophoretic pipette was always placed at the more proximal location than glutamate iontophoretic pipette was placed . For each recorded cell , an electrode was placed at the soma to set the resting membrane potential to about in order to obtain a driving force of for inhibitory GABA inputs . Electrical signals of individual and summed iontophoretic responses were amplified and filtered at 3 kHz ( low pass ) by a patch clamp amplifier ( MultiClamp 700B , Molecular Devices ) , digitalized ( 100 kHz ) by an AD-DA converter ( Digidata 1440A , Molecular Devices ) , and acquired by a pClamp 10 . 3 ( Molecular Devices ) into a computer for further analysis . In order to study the dendritic integration of a pair of excitatory and inhibitory inputs , for fixed input locations and strengths , a moving average technique with time lag was first applied to smooth each individual trace of the EPSP , IPSP , and SSP recorded in our experiments . After smoothing , we measured the amplitudes of EPSP , IPSP , and SSP at different times , including those when EPSP reached its peak value , and denoted them by , , and , respectively . By varying the excitatory and inhibitory input strengths , we measured values of , and . We then constructed a scatter plot of vs . at different times . We divided the range of into approximately 10 bins and averaged all the data points within each bin . The number of bins was chosen to ensure at least 8 data points were used for averaging . However , the qualitative results were not sensitive to the number of bins ( e . g , from 6 bins to 16 bins ) . Using the Curve Fitting Toolbox in Matlab Version 7 . 14 , we finally fitted the averaged data points by a linear function , from which the slope was estimated together with its confidence interval . For the dendritic integration of a pair of identical type , the same data processing procedure was followed .
|
A neuron , as a fundamental unit of brain computation , exhibits extraordinary computational power in processing input signals from neighboring neurons . It usually integrates thousands of synaptic inputs from its dendrites to achieve information processing . This process is known as dendritic integration . To elucidate information coding , it is important to investigate quantitative spatiotemporal dendritic integration rules . However , there has yet to be extensive experimental investigations to quantitatively describe dendritic integration . Meanwhile , most theoretical neuron models considering time-dependent synaptic inputs are difficult to solve analytically , thus impossible to be used to quantify dendritic integration . In this work , we develop a mathematical method to analytically solve a two-compartment neuron model with time-dependent synaptic inputs . Using these solutions , we derive a quantitative rule to capture the dendritic integration of all types , including excitation-inhibition , excitation-excitation , inhibition-inhibition , and multiple excitatory and inhibitory inputs . We then validate our dendritic integration rule through both realistic neuron modeling and electrophysiological experiments . We conclude that the general spatiotemporal dendritic integration structure can be well characterized by our dendritic integration rule . We finally demonstrate that the rule leads to a graph representation of dendritic integration that exhibits functionally sparse properties .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"neuroscience",
"neuroscience",
"biology",
"and",
"life",
"sciences",
"computational",
"biology"
] |
2014
|
Bilinearity in Spatiotemporal Integration of Synaptic Inputs
|
A long-standing goal in biology is to establish the link between function , structure , and dynamics of proteins . Considering that protein function at the molecular level is understood by the ability of proteins to bind to other molecules , the limited structural data of proteins in association with other bio-molecules represents a major hurdle to understanding protein function at the structural level . Recent reports show that protein function can be linked to protein structure and dynamics through network centrality analysis , suggesting that the structures of proteins bound to natural ligands may be inferred computationally . In the present work , a new method is described to discriminate protein conformations relevant to the specific recognition of a ligand . The method relies on a scoring system that matches critical residues with central residues in different structures of a given protein . Central residues are the most traversed residues with the same frequency in networks derived from protein structures . We tested our method in a set of 24 different proteins and more than 260 , 000 structures of these in the absence of a ligand or bound to it . To illustrate the usefulness of our method in the study of the structure/dynamics/function relationship of proteins , we analyzed mutants of the yeast TATA-binding protein with impaired DNA binding . Our results indicate that critical residues for an interaction are preferentially found as central residues of protein structures in complex with a ligand . Thus , our scoring system effectively distinguishes protein conformations relevant to the function of interest .
Proteins are dynamic molecules that adopt multiple structures in vitro and in vivo [1] . To study the role protein dynamics has in protein function , a combination of approaches has been used [1]–[6] . For instance , crystallographic structures of proteins associated with different substrate's analogues have been instrumental in understanding enzymatic function [6] . More recently , the role of protein dynamics in the dihydrofolate reductase function has been analyzed using nuclear magnetic resonance relaxation dispersion [5] . Furthermore , techniques such as NMR , hydrogen-deuterium exchange and mutagenesis experiments have provided insights at specific time-scales of protein dynamics and function [7] , [8]; however , the detailed understanding of protein dynamics usually requires information over a broad range of time-scales . Thus , computational modeling is becoming central in studying the link between protein dynamics and protein function for multiple time-scales [8] . To effectively link protein dynamics to protein structure and function using computational modeling techniques , it is required to know the structure of a protein bound to a natural ligand , considering that protein function at the molecular level is understood by the ability of proteins to bind to other molecules ( e . g . , biological macromolecules and/or small molecules ) . However , public databases of protein structures scarcely show this information: for instance , in September 4 2007 , the PDB release contained 45 , 632 entries including 1 , 856 protein-DNA complexes ( data obtained from the Protein Data Bank [9] ) , and 1 , 700 protein-protein complexes ( PINT database [10] ) . Thus , a computational procedure to identify functional conformations of proteins will facilitate the modeling of protein function in terms of protein structure and dynamics . In this work , we introduce a computational approach aimed at identifying functional conformers of proteins . To explain the basis of our approach , we have established some definitions and axioms . Definitions:D1We refer to a protein function as a process ( group of events over time ) that depends on the intra and inter molecular interactions of proteins . D2A protein conformer is the three-dimensional structure of a protein at a given time , and it corresponds to a local minimum in the free energy surface . D3A functional conformer of a protein is a protein structure that at a given time participates in a particular protein function ( e . g . , catalysis ) . D4Critical residues for a protein function are those residues that upon mutation abolish the activity of the protein . This definition depends on the way the activity was experimentally measured; hence , a ( experimentally determined ) critical residue may be either a residue critical for maintaining the protein structure or a residue critical for the interaction with other molecules , or both . For the proteins analyzed here , residues that did not tolerate more than 2 substitutions without loosing full activity in vivo were considered critical residues . Here we simply refer to these residues as critical residues , unless otherwise specified ( i . e . , critical residues for ligand binding ) . D5Central residues are the most traversed residues with the same frequency in networks derived from a given protein conformation ( see Methods and [11 , 12] ) . The most traversed residues are identified by an automatic procedure [12]and usually involve 20% or less of the residues in a protein conformer . Furthermore , to model protein function in terms of protein dynamics , we will assume as axioms:A1Proteins accomplish their function through a set of conformationsA2Critical residues for protein function play their roles in that set of conformations . Note that experimental evidence supports axiom A1 [1]–[6] , but no evidence exists for axiom A2 . However , if axiom A2 is correct , we should be able to identify functional conformers of proteins by identifying those conformers harboring preferentially the critical residues for ligand binding . In order to relate different conformations with different critical residues we need to estimate a property of the residues that varies with the conformation of proteins; the property used in this study is centrality . One of the reasons to choose centrality comes from the observed alteration in the centrality values of critical residues involved in binding in the dihydrofolate reductase enzyme upon ligand binding [13] . Our method scores for the presence of critical residues as central residues in different protein conformers , thus the conformers with higher scores are postulated to be the conformations associated to the interaction of interest . It is important to note that many possible conformations could be involved in binding a ligand , provided that the ligand as well presents several conformations accessible to the protein . In this regard , our method does not attempt to identify all of them or a specific one . Instead , here we show that our method can determine from a population of protein conformations , which ones are those related to the binding of a ligand . In summary , the goal of our work is to identify the functional conformers of proteins . For that , we describe a method that accounts for the presence of critical residues important for ligand binding in different protein conformations . We tested our method in 24 different proteins and more than 260 , 000 conformations of these proteins both in the absence of a ligand or bound to a ligand . Our results indicate that functional conformers harbor preferentially the critical residues for ligand binding as central residues , thus providing a procedure to effectively identify the functional conformers of proteins .
Our group [11] , [12] and others [14] , [15] have previously reported that network centrality is related to the function of the protein . In most of these previous works , every function of the protein ( e . g . , folding , catalysis ) was limited to the analysis of a single protein structure . Considering axiom A1 ( an ensemble of protein conformations accomplishes protein function ) , the analysis of a single protein structure may not be appropriate to effectively understand protein function . Thus , a procedure that uses multiple protein conformers to identify critical residues may be more reliable . A first step in our approach is to build a network representation of a protein conformer ( two residues were linked if they have at least one pair of atoms at 5 Å or less , see Methods ) . From this network , we determine the central residues as those with the largest transitivity value and the same frequency of occurrence in the network ( see Figure S1 ) . The transitivity values were obtained by counting the number of times a residue was in the shortest paths connecting every pair of residues in the network ( see Methods ) . This may be extended to include as many protein conformers as required . In order to estimate the reliability of our procedure to link critical residues with central ones , we used two parameters: sensitivity and specificity . Sensitivity accounts for the fraction of truly predicted critical residues , and specificity for the fraction of truly predicted non-critical residues ( see Methods ) . To this end , we have reported that using multiple protein conformations derived from the normal modes of vibration improves the sensitivity of predictions based on the transitivity [12] . Here , we extend these results for two well-characterized proteins in terms of structure and function , HIV protease [6] , [16] and T4 lysozyme [17] . We observed that including a large number of experimentally determined protein conformers improved the reliability for predicting critical residues from the residue's transitivity parameter ( see Figure 1 ) . Additionally , we looked at the triosephosphate isomerases ( TIMs ) , a family of enzymes involved in central metabolism . This family includes 16 protein orthologs with known three-dimensional structures in the current PDB release . We observed that central residues shared by most TIM structures , actually correspond to the most conserved residues ( see Figure 2 ) . Thus , including multiple protein conformers does improve the relationship between central residues and critical residues providing support to axiom A1: this improvement could be explained by the presence of different central residues in different protein conformations , which is the basis for the contention that a collection of structures corresponds to the functional conformation of the protein . Our results suggest that different sets of protein conformers harbor different sets of central and critical residues . That is , each protein conformer presents several and different central residues . If this were correct , then it would be possible to find the set of protein conformers harboring the critical residues for ligand binding: the functional conformers . That is the contention of axiom A2 . In Figure 3 , the fraction of identical central residues shared by every pair of protein conformers ( y-axis ) was calculated and normalized to 1; so , Figure 3 shows that even when two conformers are similar ( e . g . , some HIV-1 protease conformers share less than 1 Å RMSD values; see Figure 4 for the RMSD values ) , their central residues are not the same ( no value of 1 was found between any protein conformer compared ) . To determine if there is a relationship between centrality and the structural differences between the conformers ( measured as the Root Mean Square Deviation ) , we plotted the RMSD against the fraction of central residues shared by every conformer; we found that there is no such relationship ( Figure 4 ) . Thus , we have shown that different protein conformers have different central residues despite the small geometrical differences observed between the proteins and , consequently , that there is no relationship between the overall geometrical differences observed between protein conformers and the occurrence of central residues in these conformers . These results provide the basis to assess axiom A2 . We propose that if a protein conformer participates in a given protein function , it must harbor as central residues those that are critical for that function ( axiom A2 ) . For instance , protein conformers of an enzyme solved in the presence of its substrate may show as central residues the critical residues involved in binding the substrate . In order to account for this , the sensitivity values reported in the following sections will use as critical residues those critical for ligand binding only , thus differing from the previous results shown so far . To evaluate axiom A2 , we looked at the HIV protease for which there are multiple protein complexes solved with a substrate or an inhibitor . From crystallographic [6] and mutagenesis studies [16] , it has been shown that the residues Asp25 , Gly27 , Asp29 , Asp30 , Lys46 and Ile50 are critical for substrate binding and/or catalysis . For comparison , we analyzed 42 and 31 HIV protease structures solved in the absence or presence of a substrate analogue , respectively ( see Methods for the list of PDB structures ) . By looking at the fraction of critical residues harbored by these sets of conformers as central residues ( expressed as the sensitivity value ) , we observed that the HIV protease conformers bound to a substrate analogue predominantly show as central residues those that are known to be involved in catalysis ( see Figure 5 ) . We also analyzed multiple computationally generated protein conformers . In these studies , we used the yeast TATA binding protein ( TBP ) , which has been solved both in the presence [18] and in absence [19] of its ligand: the DNA TATA box . It has been previously shown by mutagenesis that at least 53 residues in yeast TBP are involved in DNA binding ( see Table 1 ) . We ran four molecular dynamics simulations , and for each of them 63 , 000 structures were generated . The four simulations included: a ) TBP+WtDNA , TBP in the presence of a high affinity substrate ( the TATA sequence ) , using PDB file 1YTB [19] as the starting structure , b ) TBP-WtDNA , TBP that was solved in the presence of the TATA sequence ( that is 1YTB ) , but the DNA was not included in the simulation , c ) TBP-GCDNA , TBP in the presence of a low affinity substrate ( GC sequence ) generated by in silico substitution of the TATA sequence present in 1YTB by the GCGCGCGCGC DNA duplex and d ) TBP solved without substrate , using PDB file 1TBP [18] as a starting structure . The abundance of critical residues for DNA binding found as central residues in these conformers follows the order: a ) >b ) >c ) >d ) ( see Table 2 and Figure 6 ) . Also , there is no correlation between the RMSD differences of the conformers and the critical residues for DNA binding harbored by these conformers ( see Figure 7 ) . In order to analyze the veracity of axiom A2 and the reliability of our method in a larger data set of proteins , we employed the MolMov set that includes a total of 20 different proteins ( see Methods and Table 3 ) . This set includes a subset of protein structures solved in the absence of a ligand ( subset U ) and a subset of protein structures interacting with a ligand ( subset I ) . A total of 286 alternative conformations were generated for every protein structure in each subset , providing a total of 2 , 860 protein structures in each subset , as derived from the normal modes of vibration ( see Methods ) . The critical residues for ligand binding for each protein were assumed to be those conserved residues on the protein surface ( see Methods ) . This assumption includes some degree of uncertainty ( conserved residues not necessarily are functionally relevant ) and provides an additional way to evaluate our procedure ( see below ) . We observed that on average , the proportion of truly predicted critical residues ( expressed as sensitivity ) in the MolMov subset U is smaller than for the subset I ( see Figure 8A ) but not in all cases ( see Figure 8B ) . We noticed that the MolMov set included 10 proteins for which the predicted critical residues were closer to the ligand ( 3 Å on average per protein , data not shown ) in the crystal structure ( see Figure 8C for an example ) than for the other 10 proteins in the MolMov set ( see Figure 8D for an example ) . Thus , only when the critical residues are truly related to the function of interest , our approach can identify the associated conformations to that function . These results are independent of the nature of either the ligand or the protein analyzed ( see Table 3 ) . The 53 mutants listed in Table 1 were identified with TBP-DNA binding gel-shift assays [20]–[42] . The assay does not distinguish between folding-defective mutants and mutants directly involved in DNA binding . In contrast to the HIV protease , there are not numerous structures of the yeast TBP bound to the TATA DNA , thus limiting our ability to establish the structure/dynamics/function relationship of these mutants . For instance , the assumption that only residues less than 5 Å from DNA are directly involved in binding eliminates residues that are at a longer distance from DNA; yet , these distant residues may be at 5 Å or closer to the DNA in some alternative conformations of TBP bound to DNA . If multiple protein structures are computationally generated to determine which residues always fall within a cut-off distance from DNA , there is no a priori knowledge to determine if all possible conformations were explored . Thus , simply measuring the distance between the ligand and the protein does not provide a comprehensive method to link structure to biological function . Similar reasoning may be applied to energy calculations , since there is no a priori energy value that may be used to specify the relevant residues for binding . In this context , our method does not measure the distance between the ligand and protein , thus is complementary to the criteria based on the distance between the ligand and a protein and could be used to improve our ability to identify critical residues for protein-ligand interactions . All 53 critical residues in TBP involved in DNA binding qualified as central residues in the structures generated during the simulations ( see Table 4 ) . This indicates that the simulations sampled relevant conformations of TBP associated to the function of the 53 DNA-binding null mutants . However , the centrality criteria used to map critical residues onto protein structures does not distinguish between critical residues for structure and binding . Thus , we examined if there are differences in the presence of these critical residues in the simulations . We would expect that critical residues found exclusively in simulations of TBP in the presence of DNA are more likely to be involved in binding , while those residues prevalently found in all the simulations ( frequency> = 0 . 50 ) are more likely to be involved in maintaining TBP structure . From Table 4 , we identified Lys97 , Ser118 , Pro191 , Lys211 , Val213 and Thr215 ( yeast TBP numbering ) as residues critical for binding , whereas critical residues for TBP structure would be Leu67 , Leu76 , Leu80 , Val122 , Leu172 and Leu175 . In agreement with the yeast TBP-DNA structure , all residues that were predicted to be involved in DNA binding are oriented towards it , while those predicted to be involved in TBP structure actually are in the protein's core , with the exception of Val122 , which faces DNA . Moreover , Leu67 , Leu76 , Leu80 , Leu172 and Leu175 were shown to produce misfolded proteins upon mutation to Lysine [18] .
Our results support the notion that protein function is achieved through an ensemble of protein conformations [4] , [43] . The method shown here may be applied to any other protein of interest to identify its potential functional conformers . For that purpose , we have made available the software to identify central residues at http://bis . ifc . unam . mx/jamming/ [12] . The identification of functional conformers of a target protein is indeed useful in many different areas of research , such as drug design , protein function design and protein-protein interaction predictions , among others . Likewise and as shown here , the ability to map critical residues onto protein structures may increase our capacity to link experimental data with structural information . For instance , in many mutagenesis studies of proteins , especially those that test the in vivo function of the mutants , it is not obvious if the defects in function are related to a folding and/or processing problem , or to a more subtle functional effect . Our method may aid in the interpretation of such data .
To study the relationship between conserved residues and central residues in multiple protein structures , two proteins were used: HIV protease and the T4 lysozyme . For the HIV protease , 73 experimentally determined crystal structures were used: 1a30 , 1a8g , 1a9m , 1aaq , 1ajv , 1ajx , 1axa , 1bdr , 1bv7 , 1bv9 , 1bwa , 1bwb , 1cpi , 1dif , 1dmp , 1gnm , 1gnn , 1gno , 1hbv , 1hih , 1hiv , 1hos , 1hps , 1hpv , 1hpx , 1hsg , 1hte , 1htf , 1htg , 1hvc , 1hvi , 1hvj , 1hvk , 1hvl , 1hvr , 1hvs , 1hwr , 1hxb , 1hxw , 1mer , 1mes , 1met , 1meu , 1mtr , 1odw , 1odx , 1ody , 1ohr , 1pro , 1qbr , 1qbs , 1qbt , 1qbu , 1sbg , 1tcx , 1vij , 1vik , 1ytg , 1yth , 2aid , 2bpv , 2bpw , 2bpx , 2bpy , 2bpz , 2upj , 3aid , 4hvp , 4phv , 5hvp , 7hvp , 8hvp , 9hvp . For the T4 lysozyme 23 experimentally determined crystal structures were used: 1ctw , 1cu0 , 1cu2 , 1cu3 , 1cu5 , 1cu6 , 1cup , 1cuq , 1cv0 , 1cv1 , 1cv3 , 1cv4 , 1cv5 , 1cv6 , 1cvk , 1cx7 , 1d2w , 1d2y , 1d3f , 1d3j , 1d3m , 1d3n , 1qsq . To identify functional conformers three sets of protein structures were used: HIV protease , the yeast TATA-Binding Protein ( TBP ) and the MolMov set of proteins . For the HIV protease , the same protein structures described above were used . The PDB code of those structures in complex with a substrate analogue are: 1aaq , 1cpi , 1dmp , 1hbv , 1hih , 1hiv , 1hos , 1hps , 1hpv , 1hte , 1htf , 1htg , 1hvi , 1hvj , 1hvk , 1hvl , 1hvr , 1hvs , 1ohr , 1sbg , 2bpv , 2bpw , 2bpx , 2bpy , 2bpz , 4hvp , 4phv , 5hvp , 7hvp , 8hvp , 9hvp . For TBP , the crystal structures used had the PDB codes: 1tbp for TBP without DNA , and 1ytb for the TBP complex with a TATA box ( TATATAAA ) . In the case of the MolMov set , we used the proteins reported at the database of macromolecular movements [44] . This database includes structures of proteins motions and we have analyzed only those including an interaction with a ligand . Thus , this set includes protein structures in the absence of a ligand ( MolMov subset U ) and the structures of the same protein solved in the presence of a ligand ( MolMov subset I ) . The PDB codes in the MolMov subset U includes: 1bjz , 1beb , 1dqz , 1tre , 1pin , 1dv7 , 4crx , 1ex6 , 1fto , 1omp , 1rkm , 1oib , 1nyl , 1urp , 1akz , 1d6m , 1gp2 , 2pfk and 1pjr . The PDB codes in the MolMov subset I include: 1bjy , 1b0o , 1dqy , 6tim , 1f8a , 1dvj , 1crx , 1ex7 , 1ftm , 3mbp , 1qai , 2rkm , 1quk , 1gtr , 2dri , 1ssp , 1i7d , 1cip , 1pfk and 3pjr . 10 of these proteins showed the predicted critical residues close to the ligand , while the other proteins showed the predicted critical residues not so close to the ligand ( see Table 3 ) . The MolMov set includes very diverse types of ligands and protein architectures ( see Table 3 ) and the number of amino acids per protein ranked from 156 to 647 . Finally , for each structure in these subsets , 26 normal modes of vibration were calculated using ElNèmo [45] and 11 protein conformations derived for each . Thus , the MolMov set includes a total of 5 , 720 protein structures , with 2 , 860 protein structures in each subset . The initial structure for the simulation of free TBP was 1TBP [18] ( PDB code ) . The structure 1YTB [19] ( chains B and D ) , which is the carboxyl terminal domain of TBP from Saccharomyces cerevisiae bound to a TATA box hairpin ( 5′ TATATAAA 3′ , CYC1 ) , was used as the initial structure; the bases in the hairpin were removed , and only 10 basepairs were kept ( the TATA box and one-basepair at the 5′ and 3′ end ) . The complex of TBP bound to sequence 5′ GCGCGCGCGC 3′ ( CG ) was constructed introducing the necessary modifications to the 1YTB structure using the Biopolymer module of InsightII program . The structures were solvated placing the solute molecules on a cubic TIP3 water box and removing all the waters within 2 . 5 Å of the solute . The cubic water box was trimmed to a hexagonal box employing the Simulaid program [46] . Initially , the water molecules and sodium atoms were submitted to an energy minimization using 4 stages of 500 Steepest Descent ( SD ) steps and 2 stages of 1000 Adopted Basis Newton-Raphson ( ABNR ) steps . After solvent minimization , periodic boundary conditions ( PBC ) were turned on employing the CRYSTAL module of the CHARMM [47] program version 28 using CHARMM27 parameters [48] , [49] . The solvent was again minimized with 500 ABNR steps keeping the solute molecule fixed . Two final minimization stages were applied to the whole system with 250 SD steps and 250 ABNR steps . The solvent was equilibrated with 150 ps of molecular dynamics using a 1 . 5 fs step in the NPT ensemble at 300 K with the Leap-Frog integrator . Later , the whole system was equilibrated using the same protocol for the solvent . The Berendsen algorithm was used . A value of 600 . 0 atomic mass units ( amu ) was used for the mass of the pressure piston . The reference pressure was set to 1 atm . The Langevin piston collision frequency was set to 10 . 0 ps−1 . The Langevin piston bath temperature was set to 300 K . The Hoover constant temperature was used . The Hoover reference temperature was set to 300 . 0 K . The mass of the thermal piston was set at 1000 kcal*ps−2 . The target temperature was 300 K . The image and neighbor list update were done when necessary ( heuristic test ) , with a distance cut-off set to 14 Å; electrostatic interactions were shifted , and van der Waals interactions were switched , to ensure smooth forces at the cutoff distance . All calculations were performed using SHAKE algorithm and an integration time step of 1 . 5 fs was used . All the systems were simulated for 10 . 65 ns using PBC with the CRYSTAL module of CHARMM in the NPT ensemble at 300 K with the Leap-Frog integrator saving coordinates every 100 steps . The last 9 ns were used for analysis . Networks were derived from protein structures by a distance criterion . That is , two residues were considered neighbors and consequently to interact if at least 1 atom on each residue is 5 Angstrom ( Å ) apart or closer . The atoms within that distance may be part of the amino acid's main chain and the amino acid's side chain . Therefore , the networks that were built had amino acid residues as nodes and their interactions as links . Links were labeled with identical weights . We previously reported that among 21 different ways to build networks from protein structures ( e . g . , distance between center of masses , charge , different distance cut-off values ) , this way reproduces with better results the prediction of critical residues from central ones [12] . Central residues were defined as those residues with the largest transitivity values having the same frequency in the network ( see Figure S1 for an example ) . The transitivity values were obtained by counting the number of times a residue was in the shortest paths connecting every pair of residues in the network . The frequency of a transitivity value is the number of residues presenting that transitivity value in a network . Thus , each residue will have a transitivity value and a frequency in the network; only those having transitivity values immediately close to the largest transitivity value in the network and with the same frequency as those with the largest transitivity values are considered central . Using this strategy we observe that about 20% or less of the residues were central given a protein structure ( see Figure S1 for an example ) . For these calculations , we used our software available at http://bis . ifc . unam . mx/jamming/ [12] . Transitivity , T , is related to betweeness , B as follows: Bi = Ti/SPi; where Bi is the betweeness value calculated for the i-node , Ti is the Transitivity value of the i-node , and SPi is the number of shortest paths connecting the i-node to the rest of the nodes in the network . Two measurements were used to account for this: sensitivity and specificity . Sensitivity , Se , is defined as Se = ( TP+FN ) /AP , where TP: true positives , FN: false negatives and AP: all positives . In our case , AP are all the critical residues determined experimentally , TP are the critical residues correctly predicted and FN the critical residues not predicted as critical . Specificity , Sp , is defined as Sp = ( AN−FP ) /AN; where AN: all negatives and FP: false positives . In our case , AN are the non-critical residues determined experimentally and FP are the residues predicted as critical , which are not critical . Additionally , in order to compare the sensitivity of the predictions in paired comparisons ( see Figure 4 ) , we defined the Combined Sensitivity parameter as:Where C1 refers to the observed central residues in protein 1 and , C2 refers to the observed central residues in protein 2 . M is the number of central residues that are truly critical residues for either protein 1 or protein 2 . Thus , 2< = CS> = 0 to distinguish it from Sensitivity . The ConSurf server [50] was used for this . The parameters used to run the ConSurf server were: Maximum likelihood method used to calculate the conservation scores , PSI-BLAST E-value = 0 . 001 , maximum number of homologous sequences = 50 and the number of PSI-BLAST iterations = 1 . Conserved residues were those with the most negative score ( color code of 9 ) .
|
Proteins participate in most of the doings of the cells through a variety of interactions . There is an intimate relationship between the function of a protein and its three-dimensional structure , but understanding this relationship remains an unsolved problem , in part due to the limited information on protein structures bound to other biological molecules . On the other hand , thousands of protein structures in the unbound or free form , are made public every year and these differ from those of the bound structures . How to predict the protein structure in the bound form may assist researchers in understanding the structure/function relationship . Here we report that protein structures bound to other molecules tend to present , as central amino acids , those that are critical for binding other molecules . This feature allowed us to identify the protein structures known to be involved in protein interactions from a screening of thousands of structures derived from the free form .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"biology/molecular",
"dynamics",
"computational",
"biology/macromolecular",
"structure",
"analysis",
"biochemistry/bioinformatics"
] |
2008
|
Computer-Based Screening of Functional Conformers of Proteins
|
We measured the amplitude of conformational motion in the ATP-binding cassette ( ABC ) transporter MsbA upon lipopolysaccharide ( LPS ) binding and following ATP turnover by pulse double electron-electron resonance and fluorescence homotransfer . The distance constraints from both methods reveal large-scale movement of opposite signs in the periplasmic and cytoplasmic part of the transporter upon ATP hydrolysis . LPS induces distinct structural changes that are inhibited by trapping of the transporter in an ATP post-hydrolysis intermediate . The formation of this intermediate involves a 33-Å distance change between the two ABCs , which is consistent with a dimerization-dissociation cycle during transport that leads to their substantial separation in the absence of nucleotides . Our results suggest that ATP-powered transport entails LPS sequestering into the open cytoplasmic chamber prior to its translocation by alternating access of the chamber , made possible by 10–20-Å conformational changes .
ATP-binding cassette ( ABC ) transporters harness the free energy of ATP hydrolysis to power the thermodynamically unfavorable trafficking of a wide spectrum of substrates in and out of the cell [1–3] . Cooperative ATP binding and hydrolysis occur in a molecular motor composed of two ATP-binding and hydrolysis cassettes ( ABCs ) , also referred to as nucleotide-binding domains ( NBDs ) [4] . ATP binds at the interface of an NBD dimer sandwiched between the Walker A motif of one subunit and the signature motif of the symmetry-related subunit [5–7] . Two transmembrane domains encode the determinants of substrate binding and provide a passageway across the bilayer . In Gram-negative bacteria , the transport of lipid A from its site of synthesis across the inner membrane is critically dependent on the expression of the ABC transporter MsbA . Loss of MsbA activity inhibits growth and is associated with the accumulation of lipid A in the cytoplasmic leaflet of the inner membrane [8–10] . MsbA has sequence similarity to a subclass of ABC transporters that is linked to the development of multidrug resistance in microorganisms and cancer through the extrusion of structurally dissimilar molecules [11 , 12] . Polyspecifity appears to be a common property of ABC efflux systems , whereas importers are substrate specific , often requiring a dedicated high-affinity binding protein for substrate delivery [1] . The molecular organization of the four domains of ABC transporters was gleaned from crystal structures of a number of importers as well as the bacterial multidrug efflux system Sav1866 [13–15] . The structures confirm the canonical interface observed in isolated NBD dimers [4 , 16] , identify structural elements in the cytoplasmic side that mediate communication between the NBDs and the transmembrane domains , and define the likely pathway of substrate transport and putative gates that control substrate access . Initial structures of MsbA proved incompatible with biochemical and structural data and were subsequently retracted owing to an analysis error [17–20] . A large body of biochemical studies including cryo–electron microscopy ( EM ) analysis [21] , cross-linking studies of P-glycoprotein ( P-gp ) [22–25] , and kinetic and thermodynamic analysis of P-gp and LmrA substrate transport cycles [2 , 26] have delineated many aspects of the transport mechanism . They collectively demonstrate that the energy input of ATP binding and/or hydrolysis is transduced to the mechanical work of an inward-outward–facing cycle of the substrate binding sites [27] . Structural studies of MsbA by site-directed spin labeling [28] and electron spin resonance ( ESR ) spectroscopy suggested that in liposomes , MsbA undergoes substantial conformational changes upon ATP binding and hydrolysis [29] . Reporting on the accessibilities and relative proximities of three transmembrane helices—2 , 5 , and 6—and adjacent regions of the intracellular domain and periplasmic loops , the spin labels revealed the presence of an asymmetric , water-exposed chamber that is open to the cytoplasm in the absence of nucleotides . ATP binding or hydrolysis occludes the chamber to the cytoplasm and increases hydration in the periplasmic side along an alternating access model [29] . Accessibility changes are accompanied by opposite proximity changes on the cytoplasmic and periplasmic sides of the transporter , although the amplitude of these movements was not determined . Conformational changes induced by either ATP binding or by the formation of a ADP/Vanadate ( Vi ) post-hydrolysis intermediate are of similar sign and magnitude , suggesting that the ATP binding provides the power stroke for transport as previously reported for P-gp [21] . The ESR constraints indicate that apo-MsbA samples conformations that depart from the reported crystal structures in helix topology and the extent of opening on the extracellular side [29] . In particular , these experiments indicated that helix 6 is shielded from direct exposure to the bilayer and is likely packed at the dimer interface as concluded from cross-linking studies of P-gp and subsequently confirmed by the Sav1866 structure [13 , 30] . Recent crystal structures have captured importers in inward-facing and outward-facing conformations and presented a model of the amplitude and extent of the underlying structural changes [15] . A central theme of this model is the limited rearrangement of the NBD dimer interface upon ATP binding and hydrolysis . Whether this model applies for efflux ABC transporters is yet to be determined . The Sav1866 structure [13] corresponds to a post-hydrolysis intermediate; thus it does not define the amplitude of the movement associated with chamber reorientation nor does it address the critical question of whether the two NBD domains remain closely packed during the ATPase cycle . In addition , the crystal structures were all determined in detergent micelles and thus may be influenced by the absence of lipids in addition to conformational selectivity imposed by crystal lattice forces [15 , 31] . To further map the conformational changes in the transport cycle of MsbA , we used pulse dipolar ESR spectroscopy [32–36] and fluorescence homotransfer [37–39] to obtain a set of critical distance constraints that monitor the relative separation of the transmembrane domains , the dimer interface , and the packing of the NBDs in detergent micelles and in the native-like environment of lipid bilayers . The distances were determined following the addition of a putative substrate , LPS , and in the high-energy post ATP-hydrolysis intermediate in the presence and absence of LPS . The change in the distance constraints provides evidence of LPS interaction with the transporter and establishes the sign and amplitude of the conformational changes on the cytoplasmic and periplamic sides following ATP hydrolysis .
Spin and fluorescence labels were attached to single cysteines in each MsbA monomer , resulting in the introduction of two symmetry-related probes in the functional dimeric unit . Distances between the probes were determined by ESR and fluorescence spectroscopies in detergent micelles . For a selected set , distances between spin labels were also measured in liposomes to establish the correspondence with conformational changes in detergent micelles . The mutants selected for this study form dimers as demonstrated by their retention times on size-exclusion chromatography . Spin-labeled mutants were shown previously to turn over ATP with rates ranging from 20%–100% of that of the wild type ( WT ) and at least 10-fold higher than the Vi-inhibited WT [29] . Three nucleotide-bound intermediates of MsbA can be stably populated: ATP-bound , ADP-bound , and ADP/Vi-inhibited . The latter is a high-energy post-hydrolysis intermediate often referred to as the transition state of ATP hydrolysis [40] . Previous site-directed spin labeling data indicated that ATP binding and hydrolysis are associated with similar overall conformational changes in MsbA [29] . Therefore , here we focus on the ADP/Vi intermediate . The sites for distance measurements were selected in regions that report changes in the local environment of spin labels [29] following the formation of the ADP/Vi intermediate . These include the intracellular or cytoplasmic sides ( IL in the notation of Dawson and Locher [13] ) of helices 2 , 5 , and 6 , where an overall reduction in accessibility to nickel-ethylenediaminediacetic acid ( NiEDDA ) was observed , and the periplasmic loop 2 ( ECL1 ) , where increased water accessibility accompanies ATP binding and hydrolysis . Double electron-electron resonance ( DEER ) data in liposomes and detergent micelles were obtained in the apo state and following ATP hydrolysis and Vi trapping to form a high-energy ADP/Vi intermediate . Figure 1A shows a representative set of data , and Figure 1B shows distance distributions calculated as previously described [41] . The length of data records displayed in Figure 1A was selected to optimize the trade-off between the signal-to-noise ratio and attainable range and resolution of distances [34 , 42] . We observed distinct dipolar oscillations in the DEER signal for all sites ( except 61 ) in the post-hydrolysis intermediate . They were less distinct in the apo state , and aperiodic decays for the loop site 61 are typical of a wide distribution of distances , pointing to a range of protein conformations . All signals provide accurate distances , as illustrated in Figure 1 and in Figures S1–S5 . The difference in appearance of the signal indicates that there is a distinct protein conformation in the post-hydrolysis intermediate in contrast to a wider distribution in the apo intermediate . Average distances , reported in Table 1 , reveal substantial reconfiguration in all three domains of MsbA . ATP hydrolysis fuels a closing motion in the intracellular domain of approximately 10–20 Å . The location of spin labels in three helices suggests that the distance changes reflect a concerted conformational rearrangement in the transmembrane segment . Site 103 in the intracellular part of helix 2 is particularly instructive , because the attached spin labels do not undergo changes in the motional state or in collision frequency with NiEDDA [29] . This indicates that there is no significant change in the local environment of the spin label , and the distance change at this site thus reflects a rigid body type movement of the backbone . An even larger distance change is reported at the NBD interface where spin labels at sites 539 move closer by almost 30 Å ( we note that the spin label at site 539 is confined such that the uncertainty in distance is 0 . 3 Å ) . In contrast , an opening movement of 10 Å occurs between spin labels at site 61 in ECL1 . Taken together , these distance changes suggest that alternating access of the chamber is induced by substantial movements on both sides of the transporter . We emphasize that the results for detergent micelles and liposomes given in Table 1 are sufficiently close to justify the use of ( detergent ) micelles in the study of MsbA . This was not at all obvious at the outset of the present study and highlights the advantage of pulse dipolar spectroscopy in allowing the measurements of distances in both environments . A similar pattern of distance changes in the three regions of MsbA is reported by fluorescein probes undergoing homotransfer [37 , 39] ( Table 2 ) . The probes were introduced in the same general locations as the spin labels , although the exact residues were adjusted to minimize perturbation by their larger molar volume , which affected reactivity and compromised stoichiometric labeling at sites such as 301 . Unlike DEER , homotransfer is measured at ambient temperatures where the transporter samples all the conformers that are accessible . However , the widths of distance distributions in the apo state as well as for spin labels in the ECL1 loop support a significant range of conformations trapped in frozen samples . Comparison of Tables 1 and 2 shows a general agreement in the sign of the distance changes at both sides of MsbA . The absolute distances between fluorescein labels in liquid solution and spin labels in frozen media are in reasonable agreement to the extent imposed by the difference in the reporter groups , and the average nature of the distances calculated from steady-state fluorescence anisotropy . Zou et al . carried out a systematic comparison of distances calculated by the two methods and suggested that the primary factors accounting for the differences are the extension of the linking arm and the tendency of either probe to undergo specific interactions with neighboring main and side chains [39] . We compared the distances measured in the ADP/Vi intermediate to the crystal structure of Sav1866 . Figure 2 maps the sites of spin labeling onto the crystal structure of Sav1866 [13] . Distances were calculated by modeling the spin label side chain into the protein structure ( e . g . , Figure S6 ) . At solvent-exposed sites , the spin label is not expected to have a preferred orientation relative to the backbone and can be represented as a cone projected along the Cα-Cβ bond with a 7-Å distance between the Cα and the nitroxide oxygen [43] , although exclusions due to the tertiary contacts leading to “unusual” rotamers are possible [44 , 45] . In particular , at site 99 , which is equivalent to 103 in MsbA , the spin labels point in opposite directions , and the predicted distance between the nitroxide oxygens is 50 Å , in close agreement with the experimental distance of 47 Å . A 47–53 Å distance range is obtained by modeling the g+ g+ g+ rotamer of the spin label visualized in a number of crystal structures [44] . At buried sites , repacking due to steric constraints between the main-chain and side-chain atoms biases the dihedral angles along the linking arm and results in a net orientation of the spin label that cannot be easily modeled . However , even in these cases , the deviations of the measured distance from the alpha carbon separation can be rationalized by the expected projection of the spin labels ( Figure 2B ) . To illustrate this effect , we sampled the range of distances between spin labels at site 244 ( 248 in MsbA ) by changing the torsion angles around the Cα-Cβ and Cβ-Sγ bonds . Whereas the sampling grid was selected to represent the extreme distances , it was not exhaustive and we did not attempt to assess the relative energies of spin labeled MsbA for every set of torsion angles . We found the modeled distance between the spin labels to vary from 14 to 24 Å ( Table 1 ) , which is shorter than the distance between the corresponding alpha carbons as expected from projecting the spin label along the Cα-Cβ vector ( Figure 2B ) , but well in line with the experimental data in Figure 1B . At site 297 ( 301 in MsbA ) , the range of distances between the spin labels is 25 to 34 Å which is larger than the alpha carbon separation reflecting a relative outward projection of the spin labels ( Figure 2B ) . Finally , we can model the spin labels at site 536 ( 539 in MsbA ) in a conformation with no steric overlap to yield a separation of 25 Å that agrees well with the measured distance ( Table 1 ) . Similar to multidrug ABC transporters , a spectrum of potential MsbA substrates has been identified based on stimulation of ATP hydrolysis [46 , 47] . Among them is the substrate lipid A and a number of its processed derivatives including Ra LPS . The latter was used in the crystallization of MsbA [48] and was visualized bound to the external surface of the molecule . Ra LPS is relatively more soluble than lipid A , hence its use is more practical for spectroscopic analysis . LPS at a concentration used for stimulation of ATP hydrolysis [46] increases the distance between fluorescein labels at all sites , although the effect on the NBDs appears marginal ( Table 2 ) . The distance increase has the shape of an apparent binding isotherm in the lower range of LPS concentrations ( Figure 3A ) . The effect is detergent-dependent: LPS titration in undecyl-maltoside , which has a higher critical micelle concentration ( CMC ) relative to α-ddm ( dodecyl maltoside ) , shifts the curve to lower concentrations , and exposes an inflection point above which the distance increases monotonically ( Figure 3A Figure S7 ) . Given that LPS forms micelles at submicromolar concentrations , the inflection point may reflect a major structural transition as LPS progressively assumes the role of a solvent . If LPS is acting as a substrate , then it is expected that ATP binding and/or hydrolysis will reduce its affinity [21 , 25] . Indeed , the LPS-induced distance increases are inhibited by prior formation of the ADP/Vi intermediate and the concomitant closure of the chamber ( Figure 3A ) , which is consistent with a model where the LPS molecule or its head group interacts with residues at the cytoplasmic end of the chamber . To further characterize the modes of LPS interaction with MsbA , we probed the structural changes induced by excess LPS concentration ( 1–10 mM ) relative to the initial detergent . LPS induces the appearance of a mobile component in the ESR spectrum at all sites explored and leads to a new population of transporters with longer distances ( Figure 3B ) . At the periplasmic site 57 , the addition of LPS reduces the amplitude of dipolar splittings ( arrow in Figure 3B ) , which reflects spin labels separated by less than 10 Å , implying an increase in distance . An effective distance increase between residues 301 ( Figure 3C ) and 61 ( Figure S8 ) for an LPS concentration of 1 . 5 mM ( which is already in excess , as compared to a protein concentration of ∼100 μM ) is also detected by DEER . The increase in distance at site 61 is more pronounced than for site 301 , with virtually no change for site 248 ( unpublished data ) . This may indicate preferential binding of LPS at the ECL1 loop region that is always accessible due to the spherical shape of micelles . In addition , the same sign of the distance change at sites 301 and 61 is not in line with the opposite signs of comparably large changes produced in opening or closing of MsbA by ATP hydrolysis , but is more indicative of a different structural change . Taken together , the data suggest that at these concentrations , LPS induces an increase in monomer separation in the transmembrane domain . At an LPS concentration of 5 mM , there is no further increase in distances , but the pattern of DEER signals clearly changes and can be interpreted in two ways . The first , and most likely , explanation is that it reflects the presence of a second component with a much longer distance ( >70 Å ) ; this could imply the complete dissociation of the dimer . The second , less likely , explanation is that at excessive concentrations of LPS , lipids bridge two transporters at their periplasmic side ( as was observed in the crystal structure of MsbA-ADP/Vi in presence of LPS in high concentration ) , leading to a more rapid decay of the DEER signal . But this case technically is more difficult to reconcile with the nearly uniform pattern of signal change for sites 61 , 248 , and 301 located at progressively larger distance from spin-labels residing on the suggested second dimer . A more detailed DEER study would be instrumental in establishing the precise nature of the mode of LPS interaction with MsbA Hydrolysis of ATP subsequent to LPS addition resets the distance constraints close to those of the ADP/Vi as long as the LPS concentration is in the range of the binding isotherm ( Figure 3A ) . The ensemble averaging of the homotransfer by steady-state anisotropy detection implies that the partial distance recovery may reflect a population of transporters that did not turn over ATP; i . e . , have a separation similar to that obtained by LPS addition . This interpretation is reinforced by the multi-component nature of the ESR lineshape and its partial recovery at higher LPS concentrations . Notable is the decrease in the population of labels with dipolar coupling at sites 248 and 307 ( unpublished data ) if LPS is added before ATP and Vi are added ( Figure 2C ) . In contrast , the ESR spectra of the preformed ADP/Vi intermediate are unchanged after the addition of up to 10 mM LPS , as illustrated in Figure 3B for site 248 ( compare black and blue traces ) . Thus , high concentrations of LPS relative to the detergent result in substantial structural reorganization , which may reflect a solvent effect rather than the specific interaction of a substrate .
The nucleotide-bound structure of Sav1866 [13] has demonstrated that the post-hydrolysis conformation of ABC efflux transporters has an open chamber to the extracellular or periplasmic side [13] . What is less clear is the orientation of the chamber in the absence of nucleotides and whether the two NBDs undergo cycles of dimerization/dissociation [4 , 6] . Crystallographic analysis of isolated NBDs led to a model wherein ATP binding is required for NBD dimerization whereas its hydrolysis favors dissociation . The ATP-switch model proposes that the NBDs cycle between open and closed dimer conformations upon ATP binding although without dissociation or major reorientation relative to the transmembrane domain [27] . Crystallographic analysis of intact ABC importers has been particularly supportive of a limited separation between the NBD dimer during transport [14 , 15] . The structures suggest that the alternating access can be accommodated with little change in overall transporter architecture [14 , 15] . In this model , the two NBDs , which are independent subunits , are in contact throughout the transport cycle with relative movement confined to the P-loop and the signature motif . In the ADP-bound structure of Sav1866 , the packing of the NBDs was interpreted as challenging dimerization/dissociation models [13] . Our distance constraints obtained in detergent and liposomes provide a scale for the movements in MsbA that mediate the cycling of chamber accessibility reported by spin labels in the intracellular regions , ECL2 , and along helices 2 , 5 , and 6 [29] . In conjunction with previously reported distance changes at site 57 [29] , the data from sites 60 and 61 strongly indicate that the extracellular side of the transporter undergoes opening motion of large amplitude following ATP hydrolysis . This distance is well beyond the uncertainties imposed by possible label conformations . Spin labels at site 57 show distinct dipolar splitting in the continuous wave ESR [29] spectrum , implying closer proximity of ECL1 in the apo conformation relative to the nucleotide-bound structure of Sav1866 . Similarly , the closing of the chamber in the cytoplasmic side that leads to reduction in NiEDDA accessibility upon ATP binding occurs through large movements , although in the opposite direction compared to the periplasmic side . While an outward/inward cycling of chamber accessibility can be accommodated in the constant contact model [15] , the maximum predicted separation between the MsbA monomers is limited . The uniformly large magnitude of the distance changes reported here by two independent probes cannot be easily interpreted by this model . The variations in the amplitude of the distance changes between different sites in the cytoplasmic domain suggest that the relative movement is not a simple relative translation between the two MsbA monomers as noted by Dawson and Locher and schematically illustrated in Figure 4 [13] . More importantly , the distance change between the NBDs cannot be reconciled with constant contact models . The packing of the NBDs in the apo BtuCD structure represents their maximal separation during the cycle [14] . The structure predicts a 13-Å distance at the α carbons of 539 , which is not consistent with our measured distance in apo MsbA . Similarly , comparison of the isolated NBDs of MalK in different nucleotide states reveals a limited association/dissociation cycle and predicts a distance change close to 4Å at the α carbon at the equivalent residue to 539 [6] . The scale of the experimentally measured distance change implies a significantly larger separation of the dissociated NBDs in MsbA as depicted in Figure 4 . It is possible that dissociation of the NBD dimer is a property of efflux ABC transporters reflecting a more substantial reconfiguration of the substrate chamber and the need to accommodate bulky substrates such as lipid A . In the Sav1866 structure , two helices in the ICD ( IL ) of one monomer contact the NBD of the opposite monomer , an interaction not observed in the BtuCD transporter . If these contacts were to be disrupted due to the opening in the ICD region as implied by our data ( sites 103 and 248 ) , this may destabilize the NBD dimer enough to allow its complete dissociation . Our results provide direct structural insight into the interaction of LPS with MsbA . It is clear that the addition of LPS leads to structural rearrangements even at low concentrations . However , as with all lipophilic substrates , the results can be easily confounded by changes in the properties of the micelles and liposomes . In the case of MsbA , high concentrations of LPS induce pronounced structural rearrangements possibly reflecting a solvent-like effect . At all sites explored here , high concentrations of LPS increased mobility of spin labels . We are carrying out a systematic analysis of LPS effects on MsbA structure in detergents and liposomes ( P Zou HS Mchaourab , unpublished observations ) . When analyzed in the context of previous crystallographic and biochemical studies , our results are consistent with the initiation of the transport cycle by substrate binding to an open chamber at the cytoplasmic side of the transporter . After ATP binding , large-amplitude motion is required to form the ABC dimer , consistent with the two domains having significant conformational entropy in the apo intermediate and not being in contact throughout the cycle . In addition to the absolute distances reported in Table 1 , this configuration is also supported by large NiEDDA accessibilities of residues on the cytoplasmic side in the apo intermediate [29] . Furthermore , the reanalyzed structure of apo MsbA shows a large open chamber and the two NBDs are separated by about 50 Å ( G Chang , personal communication ) . Because the cellular concentration of ATP exceeds the Km of MsbA , the open conformation of the apo intermediate is expected to be transiently populated . Current models of ATP hydrolysis propose that ATP turnover resets the conformation of the substrate binding site to high affinity [29] . Indeed , the accessibility profiles of spin labels in the ADP-bound and apo intermediates ( J Dong and HS Mchaourab , unpublished results ) are similar implying that the open state may be populated after ATP turnover and release of inorganic phosphate but before rebinding of ATP . In conjunction with previous accessibility data [29] , our results support a model where a reversal of the chamber polarity gradient through an alternating access mechanism makes the two orientations of the substrate headgroup relative to the transporter energetically equivalent and initiates substrate translocation .
MsbA mutants were expressed and purified as previously described [29] . Briefly , Escherichia coli BL21 ( DE3 ) harboring the mutant plasmids were grown in minimal media and the protein expression was induced at 30°C . MsbA was extracted using α-ddm and purified by a two step nickel-affinity and size-exclusion chromatographies . The mutants were labeled with either the MTSSL ( 1-oxyl-2 , 2 , 5 , 5-tetramethylpyrrolinyl-3-methyl ) -methanethiosulfonate spin label or MTS-fluorescein ( 2-[ ( 5-fluoreceinyl ) aminocarbonyl] ethyl methanethiosulfunate ) ( Toronto Research Chemicals; http://www . trc-canada . com/ ) [39] . Both fluorescein and spin-labeled mutants have retention times similar to the WT on the superdex 200 size-exclusion chromatography column . The spin-labeled mutants reported here were previously shown to turn over ATP at rates 20%–100% of the WT [29] . Reconstitution into liposomes was carried out as previously described [29] except that the lipid to protein molar ratio was adjusted to 2000/1 . For homotransfer , two samples were prepared for each mutant . Stoichiometrically labeled samples were prepared by addition of 10-fold molar excess of fluorescein twice over a period of 4–6 h . The reaction was allowed to proceed overnight at 4 °C . Underlabeled samples were prepared by adding 0 . 2 moles of fluorescein per mole of MsbA followed by addition of a 5-fold molar excess of a diamagnetic analog of the MTSSL ( Toronto Research Chemical ) to block the unreacted cysteines [49] . Labeling efficiencies were determined by comparing the absorbance at 280 nm to that at 492 nm . As shown previously , this ratio can be used to confirm labeling efficiency [39] . All stoichiometrically labeled samples had a 0 . 5 absorbance ratio . Liposomes samples for spectroscopic analysis were in a 50 mM Hepes , 50 mM NaCl , pH 7 . 5 buffer . Ra LPS ( Sigma-Aldrich; http://www . sigmaaldrich . com ) was dissolved into the same buffer containing the appropriate amount of detergent . Typically , MsbA mutants were incubated with LPS at 37 °C for 15 min before or after formation of the ADP/Vi intermediate . The ADP/Vi intermediate was trapped by addition of 1 mM Vi following addition of ATP solutions containing 5 mM MgCl2 and incubated for 20 min at 37 °C . Distance measurements were carried out either on a home-built spectrometer at the National Biomedical Center for Advanced Electron Spin Resonance Technology ( ACERT ) facility at Cornell University , which operates at 17 . 3 GHz , or on a Bruker 580 pulsed ESR spectrometer , which operates at 9 . 36 GHz , using DEER with a standard four-pulse protocol [35] in both cases . For detergent samples , glycerol was added to yield 30% w/w prior to cooling . All experiments were carried out at 50–80 K . DEER signals were analyzed by the Tikhonov regularization and maximum entropy methods ( MEM ) [41 , 50] to determine average distances and distributions in distance , P ( r ) , as illustrated in Figures S1–S5 . The error in the distance was conservatively estimated by taking half of the P ( r ) width at 0 . 7 of the height . Samples were analyzed on a steady state T-format fluorometer ( Photon Technology International; http://www . pti-nj . com/ ) . For each mutant , we collected steady-state anisotropy for a stoichiometrically labeled sample as well as an underlabeled sample . The latter serves as a reference wherein the steady-state anisotropy reflects the intrinsic reorientation of the probe . The extent of labeling was parametrized using the ratio of absorbance at 280 and 492 nm and compared to that expected based on labeling of T4L where the absolute extinction coefficient is available . The fluorescence anisotropy , r , was measured by comparing the polarization of the emitted light to the polarization of the excitation light according to the equation: where Ivv and Ivh refer to the amplitude of fluorescence emission parallel and perpendicular to the plane of excitation light , respectively . The G-factor was determined for each sample to correct for bias in each channel . Distances were calculated using an expression derived by Runnels and Scarlata [38 , 39] . We have calibrated this method using T4 lysozyme as a model protein system and demonstrated the correspondence of distances determined between spin labels and those determined by homotransfer [39] .
|
Clinical multidrug resistance in the treatment of bacterial and fungal infections and cancer chemotherapy can result from the expression of pumps that extrude toxic molecules from the cell . A subclass of these pumps—ATP-binding cassette ( ABC ) transporters—use energy from ATP to remove a wide range of molecules . MsbA is a conserved ABC transporter from Gram-negative bacteria with sequence similarity to human multi-drug ABC transporters . MsbA flips the building block of the outer membrane , lipid A , across the inner membrane . The input of ATP energy occurs in two dedicated nucleotide-binding domains ( NBDs ) , whose configuration in intact transporters is controversial . We determined the amplitude of MsbA conformational motion that couples energy expenditure to substrate movement across the membrane . Using molecular probes introduced into the protein sequence , we found that ATP hydrolysis fuels a relative motion of the NBDs close to 30 Å . The movement of the NBDs is coupled to reorientation of the chamber , which binds the lipid substrate from cytoplasmic-facing to extracellular-facing through large amplitude motion on either side of the transporters . In addition to revealing the structural mechanics of transport , these results challenge current models deduced from studies of substrate-specific ABC importers that envisions the two NBDs in contact throughout the ATP hydrolysis cycle .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biophysics",
"eubacteria"
] |
2007
|
Conformational Motion of the ABC Transporter MsbA Induced by ATP Hydrolysis
|
Horizontal gene transfer ( HGT ) plays a central role in bacterial evolution , yet the molecular and cellular constraints on functional integration of the foreign genes are poorly understood . Here we performed inter-species replacement of the chromosomal folA gene , encoding an essential metabolic enzyme dihydrofolate reductase ( DHFR ) , with orthologs from 35 other mesophilic bacteria . The orthologous inter-species replacements caused a marked drop ( in the range 10–90% ) in bacterial growth rate despite the fact that most orthologous DHFRs are as stable as E . coli DHFR at 37°C and are more catalytically active than E . coli DHFR . Although phylogenetic distance between E . coli and orthologous DHFRs as well as their individual molecular properties correlate poorly with growth rates , the product of the intracellular DHFR abundance and catalytic activity ( kcat/KM ) , correlates strongly with growth rates , indicating that the drop in DHFR abundance constitutes the major fitness barrier to HGT . Serial propagation of the orthologous strains for ~600 generations dramatically improved growth rates by largely alleviating the fitness barriers . Whole genome sequencing and global proteome quantification revealed that the evolved strains with the largest fitness improvements have accumulated mutations that inactivated the ATP-dependent Lon protease , causing an increase in the intracellular DHFR abundance . In one case DHFR abundance increased further due to mutations accumulated in folA promoter , but only after the lon inactivating mutations were fixed in the population . Thus , by apparently distinguishing between self and non-self proteins , protein homeostasis imposes an immediate and global barrier to the functional integration of foreign genes by decreasing the intracellular abundance of their products . Once this barrier is alleviated , more fine-tuned evolution occurs to adjust the function/expression of the transferred proteins to the constraints imposed by the intracellular environment of the host organism .
Horizontal gene transfer ( HGT ) is a major force in bacterial evolution [1–3] . Comparative genomic analyses show that HGT events can be broadly classified into three types: a ) acquisition of a new gene not present in the taxa , b ) acquisition of an orthologous gene in addition to the endogenous chromosomal copy , and c ) direct chromosomal replacement of a gene by its ortholog from other species ( also known as xenologous horizontal gene transfer ) [1] . Koonin et al . also found that all three HGT types are approximately equally common and represent an efficient mechanism for rapid evolution and/or adaptation to new niches [1] . The genetic mechanisms responsible for the horizontal transfer of foreign genes ( i . e . , transformation of naked DNA , conjugation , and viral transduction ) are well characterized [4–6] . However , the material transfer of DNA from other bacterial species is only an initial step . The evolutionary fate of an HGT event ( fixation , elimination by purifying selection , or persistence as a subdominant clone ) depends on the fitness benefit or cost of the newly acquired gene . Previous studies on these fitness effects have arrived at apparently inconsistent conclusions . For example , Sorek et al . [7] expressed multiple proteins from 79 prokaryotic genomes in an expression vector under control of an inducible promoter and measured the ensuing fitness effects in E . coli . They found that expression of many foreign proteins is detrimental to the E . coli host and attributed the fitness cost to a gene dosage-related toxicity . Lind et al . found that inter-species chromosomal replacement of three native genes encoding ribosomal proteins in S . typhimurium was detrimental to fitness , apparently due to low expression of transferred proteins [8] . Although these studies showed that many HGT events incur fitness costs , they did not provide mechanistic or molecular explanations of why this was the case . Meanwhile , other studies have argued that HGT is predominantly neutral rather than deleterious . Insertion of random DNA fragments from other bacteria in the Salmonella chromosome showed no significant fitness effect for ~90% of the inserts [9] . Introduction of foreign and complex subunits in E . coli also showed no loss in fitness [10] . The apparent controversies on the nature of the fitness landscape of HGT events can be attributed to several challenges . A ) Pleiotropy at the molecular level . The starting genetic material has a broad distribution of molecular and sequence properties that are not entirely independent ( e . g . , potential effect of GC-content on RNA stability [11] that could affect transcription/translation , and of protein folding stability and activity [12] that could affect function ) . B ) Pleiotropy at the cellular level . Beyond the foreign gene’s immediate functional context , HGT may affect or be affected by other cellular factors , such as protein-protein , metabolic , or regulatory interaction networks [13–16] . Another example is protein homeostasis ( proteostasis ) machinery , which maintains the integrity of the proteome through assisted folding and degradation and is known to buffer against the deleterious effects of mutations [17–19] . However , the actual effect of proteostasis on horizontal gene transfer is not yet known . C ) Time and length scale in evolution . Similarly to mutations , HGT events can be accompanied by immediate and transient responses of the cell that are particularly hard to detect using comparative genomics , because it analyzes HGT that has survived selection over long evolutionary time scales . Altogether , these challenges need to be addressed to understand the fitness landscape of HGT and the cellular responses that lead to the subsequent accommodation or rejection of a foreign gene . Here we aimed at a molecular- and systems-level mechanistic description of the origin of the fitness cost associated with HGT immediately upon its occurrence , as well as after a period of experimental evolution . In particular , we sought to develop an experimental system that allows full control over the molecular properties of the transferred gene ( Fig 1A ) . Our focus is on the functional barriers to HGT emerging at the protein level rather than genomic barriers affecting transcription and translation . To this end , we used the essential gene folA encoding dihydrofolate reductase ( DHFR ) as a model . DHFR catalyzes an electron-transfer reaction to form tetrahydrofolate , a carrier of single-carbon functional groups utilized in central metabolism , including de novo purine biosynthesis , dTTP formation , and methionine and glycine production [20] . DHFR is also an important target of antifolate therapy by trimethoprim ( TMP ) , a competitive inhibitor that binds with high specificity to the active site of bacterial enzymes [21] . Additionally , comparative genomics studies have demonstrated that HGT plays an important role in the evolution of the folate metabolic pathway , including the spread of antifolate resistance [22 , 23] . Moreover , DHFR is an essential enzyme in E . coli with a relatively low basal expression level ( approximately 40 copies per cell on average [24] ) , and its activity is linked to bacterial fitness in a dosage-dependent manner [17] . As such , DHFR is a convenient model to study the HGT-related fitness effects . We experimentally mimicked multiple HGT events by replacing the folA gene on the E . coli chromosome with its orthologs from 35 phylogenetically diverse mesophiles . This collection of orthologs explores a broad distribution of protein sequences and biophysical properties . We found that immediately after HGT the fitness effects were largely deleterious , whereas after the experimental evolution fitness has improved markedly . Using biochemical and genetic analysis , whole-genome sequencing , and proteomics , we show that the mechanistic origin of the barriers to HGT is the global response of the protein homeostasis machinery . Altogether , this work provides molecular and cellular level insights into the origin of the barriers that shape the fitness landscapes of HGT events .
We initially identified 290 orthologous DHFR sequences from mesophilic bacteria and selected 35 diverse sequences with amino acid identity to E . coli DHFR ranging from 29% to 96% ( Fig 1C and S1 Table and Materials and Methods ) . First , we sought to minimize the contributions from confounding factors that mostly affect transcription and translation of replaced genes , such as GC content [25] , codon-usage pattern [26 , 27] , specific loci at which chromosomal incorporations occur , and the copy number of the transferred genes [6 , 11] . The amino acid sequences of the chosen 35 orthologous DHFRs were converted into DNA sequences using the codon signature of E . coli’s folA gene ( Materials and Methods , S2 Table ) . We used the λ-red recombination system [28] to replace the open reading frame ( ORF ) of folA with the synthetic DNA sequences , while preserving E . coli’s wild-type folA promoter ( see Materials and Methods ) . Thus , the resulting 35 strains carrying the orthologous DHFR gene replacements are identical with respect to the chromosomal location of the folA gene and the mode of regulation of their DHFR expression . In addition , they have similar GC content and codon usage signature ( S1 and S2 Tables ) . We assayed the fitness of the resulting HGT strains by measuring their growth rates at 37°C ( this condition was consistently used throughout the work ) ( see Materials and Methods , Fig 1B and 1C and S3 Table ) . We found that E . coli fitness ( here and below we use the terms fitness and growth rate interchangeably ) is very sensitive to the orthologous replacements of its DHFR . Growth rates are lower than wild-type ( WT ) E . coli in 31 out of 35 strains , with six strains ( DHFR-23 , 35 , 36 , 37 , 38 and 43; highlighted in Fig 1C ) exhibiting a severe fitness loss of 70–85% . DHFR-21 ( from W . paramesenteroides ) did not grow at all under the conditions of the experiments and was omitted from further analysis . Surprisingly , we found no significant correlation between growth rates of the HGT strain and the evolutionary distance between DHFR orthologs , measured as % of amino acid identity relative to E . coli DHFR ( Spearman r = 0 . 16; p-value = 0 . 4 ) ( S1A Fig ) , thus , challenging the notion that sequence similarity between endogenous and transferred genes facilitates horizontal gene transfer . A potential explanation for the behavior of strains that exhibit most detrimental effect of HGT , namely , DHFR-23 , 35 , 36 , 38 , and 43 is that these new proteins are inherently toxic to the cell , as previously found for many horizontally transferred genes [7] . To determine whether expression of orthologous DHFRs is toxic to the cell , we followed the approach of Sorek and co-workers [7] by transforming WT E . coli cells with pBAD plasmids expressing the orthologous DHFRs under the arabinose-controlled promoter ( Materials and Methods ) . We found that none of the orthologous DHFRs caused a significant drop in fitness ( S2A and S2B Fig ) . Alternatively , we transformed the severely affected HGT strains ( DHFR-23 , 35 , 36 , 37 , 38 , and 43 ) with pBAD plasmids expressing WT DHFR ( Materials and Methods ) . As shown in S2C Fig , complementation with WT DHFR completely restored fitness . Thus , the fitness cost associated with the HGT of DHFR-23 , 35 , 36 , 37 , 38 , and 43 is not caused by toxicity , but by the loss of DHFR function in the cells . The high fitness cost of the orthologous replacements of E . coli DHFR demonstrates the existence of a molecular constraint ( “a barrier” ) to HGT . To determine whether the evolutionary process can traverse this barrier , we conducted high-throughput serial passaging of the HGT strains ( see Materials and Methods ) . Overall , we performed 31 passages which amount to ~600 generations for the WT strain . Growth rate measurements after the evolution experiment show that orthologous strains have substantially improved their fitness ( Fig 1C and S2 Table ) . Moreover , ~30% of the strains grew as well as or better than WT after the evolution experiment ( Fig 1B and 1C ) . The improvement in growth rates was especially dramatic amongst strains that experienced the most severe fitness loss upon HGT ( DHFR-23 , 35 , 36 , 37 , 38 and 43; highlighted in Fig 1C ) . Thus , the molecular constraints to horizontal transfer of the DHFR coding genes were largely alleviated during experimental evolution . To determine whether the mechanisms underlying the barrier to HGT are linked to the molecular properties of the horizontally transferred proteins , we looked at the important in vitro biophysical properties of the orthologous DHFRs ( Fig 2A and 2B ) ( see Materials and Methods ) . The thermodynamic stabilities of orthologous DHFRs quantified by Tm ( mid-transition temperature of thermal unfolding ) fall within the range expected for mesophilic proteins ( 42–63°C ) ( Fig 2A and S1 Table ) [29] . Approximately 20% of DHFRs are more thermodynamically stable than E . coli DHFR ( Fig 2A ) . Fig 2B shows that the catalytic activity ( kcat/Km ) of orthologous DHFRs varies widely , with approximately ~70% of enzymes having activities that are comparable to or better than E . coli DHFR . Some enzymes , in fact , exhibit almost 100-fold higher activity than E . coli DHFR ( Fig 2B and S1 Table ) . No significant correlation was found between the in vitro properties and fitness upon the HGT ( S3A and S3C Fig ) . Effective metabolic turnover , however , is not only a function of enzymatic activity , but also of intracellular abundance . Thus , we next measured the total intracellular abundances of DHFR in all strains ( see Materials and Methods ) . We found that immediately upon HGT ( before the evolution experiment ) , soluble and total abundances of orthologous DHFRs in all strains were lower than WT DHFR , and for ~35% of the strains , DHFR abundance was bordering the detection limit ( S3 Table and Figs 2C and S5 ) . After the evolution experiment , however , we observed a significant increase in total DHFR abundances ( S3 Table and Figs 2C and S5 ) ( Kolmogorov-Smirnov ( KS ) -test , p-value = 0 . 049 ) . We found that strains with undetectable DHFR immediately upon HGT ( DHFR-23 , 35 , 36 , 37 , 38 and 43 ) increased their abundances to detectable levels after the evolution experiment ( S3 Table and Figs 2C and S5 ) . This finding suggests that the low intracellular abundance of the foreign DHFR proteins is the direct manifestation of the HGT barrier . Moreover , alleviation of the barrier during experimental evolution leads to an increase in functional copies of DHFR . To understand how HGT shapes the fitness landscape of E . coli cells , we projected the growth rates of the HGT strains onto the catalytic activity of the DHFR proteins ( Fig 3A and 3B ) . No correlation was observed immediately upon HGT , primarily because several of the strains ( DHFR-23 , 35 , 36 , 38 , and 43 ) exhibited severe fitness drops despite carrying orthologous DHFR proteins with kcat/Km values that are comparable to or higher than E . coli DHFR ( Fig 3A ) . We were unable to purify the severely deleterious DHFR-37 , thus there is no kcat/Km value for this protein . After the evolution experiment , however , growth of the severely affected HGT strains improved dramatically ( by ~60–90% ) . This change resulted in a statistically significant correlation between catalytic activity and fitness among all the points ( Spearman r = 0 . 57 , p-value = 0 . 0007; Figs 3B and 4A ) . We note that the theoretical fit in Fig 3 ( green lines ) excludes DHFR-23 , 35 , 36 , 38 , and 43; nonetheless , after the evolution experiment , these initial outliers cluster around the theoretical prediction ( Fig 3B , orange points ) . Why did the activity-fitness relationship emerge only after experimental evolution ? Catalytic activity alone does not fully account for the dependence of growth rates on DHFR properties because the effective functional level of intracellular DHFR is also a function of its intracellular abundance ( i . e . , the product [abundance* ( kcat/Km ) ] ) . Indeed , the mapping between dosage of an essential enzyme and fitness is predicted by flux dynamics theory [30]: fitness~flux=aA ( kcatKm ) [b+A ( kcatKm ) ] ( 1 ) where A is the intracellular abundance , and constants a and b depend on the number of enzymes and topology of the metabolic network . We fitted this function for all strains , except the outliers ( DHFR-23 , 35 , 36 , 38 , and 43 ) . We found dosage-dependence not only for the evolved strains ( Fig 3D ) , but also the naive strains that largely clustered around the theoretical fit ( with the exception of DHFR-43 which remained an outlier ) ( Fig 3C ) . Consistent with the prediction from the flux dynamics theory ( Eq 1 ) , strains with lower DHFR dosage than WT E . coli were less fit than WT E . coli , whereas strains with higher DHFR dosage did not necessarily appear more fit ( plateau-like behavior ) [31] . Additionally , the correlation between gene-dosage ( the product [abundance* ( kcat/Km ) ] ) and growth rate is statistically significant for both naive ( Spearman r = 0 . 56 , p-value = 10−3 ) and evolved ( Spearman r = 0 . 64 , p-value = 10−4 ) strains ( Figs 3C , 3D and 4A ) . Thus , we conclude that the dosage-dependent fitness landscape of horizontally transferred DHFR genes generally accounts for the fitness effects of most HGT events . More generally , this mapping between fitness and the molecular properties of DHFR is an excellent manifestation of the “law of diminishing returns” [32 , 33] . Strains with Abundance* ( kcat/Km ) equal or higher than WT generally lie in the neutral regime of the “plateau” and the rest are deleterious because of low Abundance* ( kcat/Km ) values . Next , we determined the cross-correlation among multiple parameters , including molecular and sequence properties of the orthologous DHFRs , their intracellular abundance , and fitness ( Fig 4A , and Materials and Methods ) . As expected , there is no correlation between GC content and mRNA stability and growth rates ( Fig 4A ) . Apart from the correlation of fitness with Abundance* ( kcat/Km ) and ( kcat/Km ) , which we noted is a consequence of the dosage-dependent fitness landscape ( Fig 3 ) , the only other significant correlations of fitness that we found was with promoter activity and charge ( Fig 4A , 4B and 4C ) . It was previously shown that when a cell loses the catalytic capacity of its DHFR , either through destabilizing mutations or treatment of the inhibitor trimethoprim ( TMP ) , a feedback loop triggers up-regulation of folk promoter activity [34] . So we checked if the low abundance and lack of functional capacity upon HGT ( Figs 2C , 3A and 3B ) could trigger the same behavior . Indeed , we found a highly statistically significant correlation between promoter activity and growth ( orange dots in Fig 4B , r = -0 . 58 , p-value = 10−4 ) . After experimental evolution ameliorates the deficiency in functional abundance ( Fig 3D ) , the correlation between growth rate and promoter activity is lost ( Figs 4A and S4A ) . We note that the anti-correlation between promoter activity and fitness before evolution experiment is driven by the subset DHFR-23 , 35 , 36 , 37 , 38 , and 43 , which experience the most severe fitness loss . These results reaffirm that the direct consequence of the barrier is the drop in intracellular DHFR abundance . Interestingly , we also observed a significant non-linear relationship ( p<10−3 ) between the net charge ( at pH 7 ) of the horizontally transferred DHFR proteins and growth rates of the HGT strains before experimental evolution ( Fig 4A and 4C and S1 Table ) . This relationship follows a quadratic dependence , which reflects the observation that strains with reduced fitness also have DHFRs with high net charge . Again , the subset DHFR-23 , 35 , 36 , 37 , 38 , and 43 drives the trend . The correlation of the DHFR with net charge is largely lost after the evolution experiment ( Figs 4A and S4B ) . We also found a strong correlation between soluble and total intracellular abundances of DHFRs with thermodynamic stability in the strains upon HGT ( p<10−5 ) ( Fig 4D and 4E ) ( see Materials and Methods ) . Although this correlation remains strong for soluble abundance after the evolution experiment , this trend weakens for total abundance ( Fig 4A , 4D and 4E ) . We found earlier , again in the context of mutations in DHFR , that the protein homeostasis degrades less stable proteins at a faster rate , thus giving rise to a relationship between total protein abundance and stability in active cytoplasm . In contrast , in solution stability affects only the distribution between folded and unfolded species leaving the total amount unaffected . Thus , the unchanged dependence of soluble abundance on stability after the evolution experiment is likely to be due to the intrinsic propensity of less stable proteins to aggregate , a property that has not changed throughout the evolution experiment . Altogether , these observations again point out to a potential role of protein homeostasis in HGT events . Evidence suggest that loss of abundance is responsible for the barriers to horizontal transfer of the orthologous DHFR proteins , so next we seek to determine the molecular mechanism for that loss of abundance . To that end we performed whole genome sequencing ( WGS ) of the evolved DHFR-23 , 35 , 37 , 38 , and WT strains ( Fig 5A and S4 Table ) . Out of 8 independent evolutionary trajectories conducted for each of the strains , trajectory 1 was arbitrarily chosen for sequencing . Randomly isolated individual clones as well as whole populations were sequenced for each of the strains ( with the exception of DHFR-23 , for which WGS was done only on the whole population , S4 Table ) ( see Materials and Methods ) . WGS revealed a number of genetic variations in the evolved strains at various allele frequencies , however , the only variation common to all genomes of the evolved HGT strains but not the evolved WT strain , was the IS186 insertion in the clpX-lon intergenic area ( Fig 5A and S4 Table ) . The lon gene encodes ATP-dependent protease Lon that is central in maintaining protein homeostasis [35] . Lon is also known to directly control the turnover rates of several proteins in E . coli [36] . Previously , we showed that for E . coli strains carrying destabilizing mutations in the endogenous DHFR , lon deletion improved fitness by increasing the intracellular abundance of the mutant DHFR proteins . We surmised that a similar mechanism might have taken place in our HGT strains . To prove this conjecture , we randomly isolated 4–5 individual clones from the evolved populations of DHFR-35 , 37 , 38 and WT strains and performed a direct sequencing ( PCR amplification followed by Sanger sequencing ) of the clpX-lon intergenic area and lon coding sequence . We confirmed that the evolved HGT strains , but not the evolved WT , contain the IS186 insertion ( see Materials and Methods and S5 Table ) . DHFR-35 exhibited a variety of outcomes whereby one colony contained the IS186 insertion in the clpX-lon intergenic area , one colony carried a non-synonymous D445E mutation in the lon coding sequence , and the remaining two colonies did not carry any mutations in the lon region ( S5 Table ) . Next , we extended the direct sequencing of clpX-lon intergenic area to whole populations of all 35 naive and evolved HGT strains and WT strain ( S5 Table ) . Again , only evolved strains of DHFR-23 , 35 , 37 , 38 carried the IS186 insertion in the clpX-lon intergenic area , but not the evolved WT strain . This same insertion was detected earlier in bacterial populations that developed antibiotic resistance and was hypothesized to inactivate Lon by blocking its expression [37] , although no experimental evidence was provided to support it . We tested this hypothesis in the current work by measuring Lon intracellular abundance ( see Materials and Methods ) and found that all the HGT strains carrying the IS186 insertion upstream to their lon gene have indeed dramatically reduced the intracellular Lon abundances compared to WT ( Fig 5B ) . These are also the strains that exhibited a dramatic improvement in their fitness upon conclusion of the evolutionary experiment ( Fig 1C ) . Interestingly , we also found that DHFR-23 and 37 carried the IS186 insertion prior to the evolutionary experiment , whereas other naive HGT strains and naive WT did not ( S5 Table ) . The WT E . coli strain ( MG1655 ) used in our experiment to generate the HGT strains also does not carry the IS186 insertion in the clpX-lon intergenic area as shown by direct PCR sequencing . By virtue of the experimental design , each HGT strain in our experiments originated from a single WT cell where a homologous recombination successfully replaced the endogenous folA gene with a DHFR ortholog ( see Materials and Methods ) . However , because of the heavy fitness cost associated with orthologous replacements of the endogenous folA gene with genes encoding DHFR-23 and 37 , we believe that a homologous recombination only resulted in a viable strain if it was accompanied by a compensatory mutation . It is plausible that incorporation of IS186 transposable element in the lon promoter , a known hot spot for IS186 incorporation [38] , was co-selected with the DHFR-23 and 37 replacement because it buffered the otherwise severely deleterious ( possibly lethal ) fitness effects . For DHFR-35 and 38 , the IS186 insertion in the clpX-lon intergenic region was not present in the naive strain , but only arose as a result of the experimental evolution , which was accompanied by a dramatic improvement in fitness . Altogether , the IS186 insertion in the clpX-lon intergenic region was selected independently in four different HGT lineages: Twice as an immediate compensation of the deleterious fitness cost associated with orthologous replacement ( naive DHFR-23 and 37 strains ) , and twice as a result of experimental evolution ( evolved DHFR-35 and 38 strains ) . Since we observed lon inactivation via IS186 insertion only in evolved orthologous transfer strains , but not evolved WT , the direct test of whether Lon is solely responsible for the phenotype would be to knock out lon in the naive orthologous transfer strains , and follow the ensuing effect on fitness and DHFR abundance . To that end we introduced lon knock-out mutations on the background of naive ( pre-experimental evolution ) HGT strains . Previously , we observed that lon knock-out resulted in substantial improvement in growth of most naive HGT strains , including DHFR-23 , 35 , 37 and 38 . In the current work , we measured the effect of lon knock-out on the intracellular DHFR abundance . We indeed found a marked increase in abundances of DHFRs 35 and 38 in a soluble lysate fraction ( an increase from non-detectable levels to levels that comprise ~30% of abundance measured for E . coli DHFR in WT strain ) ( Fig 5C ) . Thus , the underlying mechanism behind the detrimental fitness effects of HGT is indeed the action of Lon on the foreign DHFR . Despite the strong selective advantage of shutting down the Lon protease , especially in the background of very deleterious HGT events ( Fig 4 ) , such a mechanism may not be evolutionary sustainable because Lon has other clients that are central to cellular function [36] . Thus , we hypothesized that mutations specific to DHFR would eventually arise to maintain its required intracellular abundance . This may include mutations in the DHFR promoter and/or coding regions , or gene duplication events . To explore this possibility , we investigated in greater detail 8 independent evolutionary trajectories of DHFR-37 , which suffered a deleterious effect from HGT but dramatically increased fitness after the evolution experiment in all trajectories ( Fig 6A ) . Indeed , whole-genome sequencing of the evolved DHFR-37 ( a single evolutionary trajectory randomly chosen out of 8 independent trajectories after 31 serial passages ) revealed a C to T mutation at position -35 of the folA promoter with a concomitant increase in the intracellular DHFR abundance ( Fig 6B ) . Interestingly , this mutation was previously shown to increase the trimethoprim resistance in E . coli MG1655 strain , presumably by increasing DHFR abundance [39] . As noted in the previous section , DHFR-37 had the IS186 insertion inactivating the activity of its Lon protease already at the naive stage ( Fig 5 ) . We performed direct sequencing of the folA gene in four randomly chosen colonies from each of the 8 independent evolutionary trajectories of DHFR-37 strains . As shown in Fig 6B , we found that two more trajectories acquired mutations in folA promoter: G to A at position -31 , ( trajectory 2 ) and a GTA insertion at position -10 ( trajectory 8 ) . Similarly to trajectory 1 ( C to T substitution at -35 ) , these mutations were accompanied by an increase in DHFR abundance ( Fig 6B ) and improvement in growth rate ( Fig 6A ) . Thus , these results show that once the barrier imposed by the proteostasis is overcome , a more fine-tuned evolutionary response follows . To determine how the effect of HGT percolates throughout the entire E . coli proteome , we next analyzed the systems-level effect of inter-species DHFR replacements before and after the evolution experiment . To that end , we quantified relative ( to WT ) abundances of ~2000 proteins in the cytoplasm using tandem mass tags ( TMT ) with subsequent LC-MS/MS analysis ( Materials and Methods and [34] ) . We picked five strains for proteomic characterization based on their fitness effect upon HGT ( Fig 1B ) : DHFR-23 , 35 and 38 ( severely deleterious ) ; DHFR-22 ( mildly deleterious ) ; and DHFR-39 ( beneficial ) . For reference , we compared the proteomic effects of orthologous replacements with the proteomic effect of treating E . coli with 1 μg/ml of trimethoprim ( TMP ) . For each of the >2000 proteins detected , we quantified their enrichment using the log of relative protein abundances ( LRPA ) that are expressed as z-scores ( see Materials and Methods ) . In Fig 7A , we show the correlation plots of the z-scores between proteomes . The complete gene-by-gene proteomics data are presented in S6 Table . As expected for DHFR-35 and 38 , where HGT is severely deleterious , their proteomes strongly resemble the proteome of TMP-treated wild-type strain ( r = 0 . 42 , p-value = 9 . 5×10−74 and r = 0 . 50 , p-value = 7 . 1×10−105 , respectively ) ( Fig 7A ) . This result suggests that the systems-level response to HGT of the DHFR genes is akin to response to inactivation of the endogenous DHFR protein by TMP . Interestingly , despite significant evolutionary distance between the DHFR alleles from strains DHFR-35 and 38 ( Fig 1C ) , the correlation between their proteomic profiles is significant ( r = 0 . 84 , p-value < 10−300 ) . However , the proteome of DHFR-23 , another orthologous strain with a severely reduced growth , was not similar to the TMP-treated WT proteome ( r = 0 . 07 , p-value = 0 . 0041 ) , suggesting that , at least for some strains , the systems-level response to the partial loss of DHFR function follows a different pattern . The proteome of DHFR-22 , a strain with moderately reduced fitness , was much less similar to the proteome of TMP-treated WT ( r = 0 . 30 , p-value = 8 . 6×10−35 ) . The proteome of DHFR-39 , one of the few strains that grew better than WT upon HGT , bore no resemblance to TMP treatment ( r = -0 . 05 , p-value = 0 . 026 ) . After the evolution experiment , the proteomic profiles of the strains lose their resemblance to TMP-treated WT cells ( Fig 7A ) , which reflects the alleviation of the detrimental effects of HGT ( Fig 3B and 3D ) . Additionally , after the evolution experiment the proteomic profiles of the strains become more similar ( see S6 Fig and note the increased correlation in inter-strain comparison ) . In particular , although the proteome of DHFR-23 is barely similar to DHFR-35 and 38 before the evolution experiment , it becomes similar to DHFR-35 and 38 after it ( S6 Fig ) . This observation is consistent with the clustering of the strains near the growth rate-metabolic turnover curve that is predicted by flux dynamics analysis ( Fig 3B and 3D ) . Next , we carried out a comparative analysis at the level of functional pathways and operons . Using the functional and regulatory classification of genes by Khodursky and co-workers [40] , we screened the gene groups that collectively changed their abundances significantly during the evolution experiment ( Fig 7B ) , using the Kolmogorov-Smirnov test ( see Materials and Methods ) . Among them we found that the genes responsible for cell motility show the largest increase in their abundances , which is notable , because it was previously reported that the loss of DHFR function either by TMP treatment or destabilizing mutations causes a sharp down-regulation of the most of the motility genes [34] . This suggests that through the adaptation process , HGT strains have eliminated the energetic burden which caused shutting down of cell motility in the first place ( fliA operon and motility class in Fig 7B ) . In the same vein , genes responsible for a number of metabolic processes such as synthesis of amino acids and nucleotides as well as turnover of several metals show highly significant changes between the naive and evolved strains ( Fig 7B ) .
In this work we provided molecular-level insight into the origin of barriers that shape the fitness landscape of HGT events . We focused here on xenologous transfer of DHFR coding genes , a common mode of HGT [1] . The chromosomal gene replacements were done while preserving the endogenous promoter . Such an experimental design provided us with a direct control over conditions of expression , enabling us to focus on the link between variation of sequence and biophysical properties of transferred proteins and the ensuing fitness effects of HGT . We were able to purify 33 orthologous DHFRs and measure their activities and stabilities . Over half of them are more catalytically active and stable than E . coli DHFR . Nevertheless , inter-species replacement of E . coli DHFR with non-toxic and much more catalytically active proteins resulted in a considerable loss of fitness . Experimental evolution revealed the main culprit of the severe loss of fitness—Lon protease , which is repeatedly deactivated in strains with most dramatic fitness improvement . While previous work showed that lon knockout could rescue fitness of E . coli containing orthologous DHFRs ( since Lon protease is crucial for DHFR turnover ) , over-expression of the chaperonins GroEL/ES produced a similar fitness improvement [17] . However , we show here that actual evolutionary processes consistently utilize Lon deactivation , rather than GroEL/ES over-expression . Given their similar fitness effects , this suggests that Lon deactivation must be more accessible by the processes that produce the genetic variation on which selection acts . Indeed , we find that Lon deactivation is consistently achieved by insertion of mobile elements in the lon promoter; mobile element insertion is known to occur frequently in bacterial genomes and thus is a major source of genetic variation [27] . Is the partial shutdown of the degradation branch of proteostasis a generic response to HGT events ? Lon inactivation increases the number of functional enzymes , but it also increases the number of misfolded/aggregated species that can be toxic in highly abundant proteins [41] . The toxicity due to misfolding may be less relevant for HGT events involving low abundance proteins . The DHFR abundance in E . coli is about 40 molecules per cell [24] , on the lower end of the range of abundances for the rest of the E . coli proteome , which span at least five orders of magnitude ( 10−1 to 104 protein copies per cell in single-cell measurements [24] or 101 to 106 per cell in bulk measurements [42] ) . Our work unambiguously demonstrates the role of proteostasis in shaping HGT fitness . However , determining the specific response of the proteostasis machinery to different HGT events is an exciting subject of future research . The key finding of this work is that E . coli cells apparently discriminate between own and foreign proteins . Most likely the discrimination is post-translational: the folA promoter gets activated as cells lose DHFR function , similarly to the effect of DHFR point mutations and inhibition by TMP [34] , so that the loss of the protein is due to post-translational degradation and/or aggregation . The mechanism of recognition of foreign proteins by E . coli proteostasis machinery remains somewhat of a mystery . In earlier studies Lon was implicated in degradation of destabilized DHFR and other misfolded proteins [43] . However in this case orthologous DHFRs do not appear unfolded or misfolded , at least in solution in vitro . Global molecular properties of orthologous DHFRs such as amino acid composition or net charge do not correlate with their abundances in E . coli . We also do not observe a correlation between the evolutionary distance of orthologous DHFRs from E . coli and fitness of the HGT strains . It is for the future studies to establish sequence or structural/physical signatures that trigger degradation of these proteins in E . coli cytoplasm . Current phylogenetic methods can now detect HGT events in a wide range of evolutionary ages , even within pathogenic clone outbreaks [44] . However , these bioinformatics approaches are agnostic about molecular and cellular mechanisms that incur fitness costs or benefits of HGT events . As shown in this work , pleiotropy at the level of molecular properties and cellular components could shape the HGT fitness landscape . Swapping out a protein with an orthologous copy ( or adding point mutations to a WT protein [45] ) has a much broader effect than merely changing the protein’s individual molecular properties like stability and activity [34] . As we observe here , HGT events that actually improve DHFR activity may nevertheless be deleterious at first because they interact negatively with proteostasis machinery of the host . We believe that experimental studies to dissect the mechanistic origin of pleiotropy will complement our understanding of HGT from genomics by pointing out to other molecular or sequence signatures of HGT events . Our finding that elements of the proteostasis machinery have the ability to discriminate between “self” and “non-self” proteins points to an important possible constraint on evolution of proteomes to maintain compatibility with their own proteostatic machinery . Future studies might reveal the manifestations of this evolutionary constraint , such as certain proteases and chaperones evolving slower than other components of proteomes , or co-evolution of sequences of components of proteostasis machinery with global features of the whole proteome such as charge distributions or amino acid compositions . Besides being relevant to understanding the evolutionary dynamics of HGT , our approach is broadly applicable to the study of the genotype-phenotype relationship . While the concept of a fitness landscapeis dominant in evolutionary biology , it remains highly metaphoric as its “axes” remain unlabeled . A promising approach to map fitness landscape is by introducing ‘‘bottom up , ” controllable genomic variations that cause known changes of the molecular properties of proteins [17 , 45–51] . However , point mutations and/or random mutagenesis are limited in their ability to generate a broad variation of catalytic activity and other physical properties of proteins . But “borrowing” highly diverged yet catalytically active orthologous proteins from other species allows us to cover a broad range of variation of molecular properties of proteins . By systematically exploring the relation between molecular properties of xenologously replaced proteins and fitness this approach provides an opportunity to quantitatively characterize the global properties of fitness landscapes .
The BLAST analysis against E . coli’s DHFR amino acids sequence of mesophilic bacteria produced 290 unique DHFR sequences [52] . This dataset was used to select 35 sequences with amino acid identity to E . coli’s DHFR ranging from 29 to 96% ( Fig 1 and S1 Table ) . The amino acid sequence of the chosen DHFRs was converted into the DNA sequence using the codon signature of the E . coli’s folA gene . Specifically , the frequency of each codon was calculated , and codons with the highest score were used for protein—DNA sequence conversion ( S2 Table ) . In addition , each DNA sequence , including E . coli’s folA , was fused to a tag encoding 6 histidines . The resulted DNA sequences were synthesized ( GenScript ) and cloned into pET24a+ plasmid ( EMD Millipore ) for recombinant expression and purification , and finally cloned into pKD13 plasmid for homologous recombination ( see below ) . The detailed description of the method can be found in . Briefly , orthologous DHFR sequences placed under the E . coli’s folA endogenous regulatory region ( 191 bp separating the stop codon of the upstream kefC gene and the start codon of folA gene ) were ligated into a pKD13 plasmid flanked by two different antibiotic markers ( genes encoding for kanamycin ( kanR ) and chloramphenicol ( cmR ) resistances ) . The entire cassette was then amplified with two primers tailed with 50 nucleotides homologous to the chromosomal region intended for recombination ( kefC gene upstream and apaH gene downstream of folA ) . The amplified product was transformed into BW25113 strain with induced Red helper plasmids , and the recombinants selected on plates carrying both antibiotics . Strains carrying the putative orthologous replacement in the chromosome were verified by sequencing . Identified chromosomal replacements were then moved to MG1655 strain by P1 transduction and double antibiotic selection ( kan and cm ) and again verified by sequencing . Absence of the endogenous E . coli’s folA gene lingering in the orthologs strains was verified by PCR and by Western Blot with specific anti-E . coli’s DHFR antibodies . Cells were grown from a single colony overnight at 30°C in M9 minimal medium supplemented with 0 . 2% glucose , 1 mM MgSO4 , 0 . 1% casamino acids , and 0 . 5 μg/μL thiamine . Overnight cultures were diluted 1/100 and grown at 37°C . Growth rate measurements were conducted for 10 hours . OD data were collected at 600nm at 20 min intervals . The resulting growth curves were fit to a bacterial growth model to obtain growth rate parameters [53] . Strains were propagated in deep 96-well plates in 8 independent trajectories by serial passaging from a diluted culture to saturation in 12-hour growth cycles at 37°C using the TECAN robotic liquid handling system . We performed 31 of such passages which amounts to ~600 generations ( for the wild-type strain ) , assuming a continuous exponential phase . mRNA folding stability was calculated using the HYBRID-SS-MIN routine of mFold over a window of 42nt starting with -50 upstream of the ATG of the DHFR coding region using the default parameters ( http://mfold . rna . albany . edu/ ? q=unafold-man-pages/hybrid-ss-min ) [54] . Codon adaptation index ( CAI ) was calculated as previously reported [55] using the whole E . coli genome as the reference set . pET24 expression vectors carrying the orthologous DHFR sequences fused to C-terminal ( 6x ) His-tag under isopropyl β-D-1-thiogalactopyranoside ( IPTG ) inducible T7 promoter were transformed into BL21 ( DE3 ) cells . Cultures were grown at 30°C overnight from a single colony in Luria broth ( LB ) supplemented with 2% glucose , diluted 1/100 into Terrific Broth ( TB ) and grown for 4 hours at 28°C . Cultures were then chilled to 18°C , supplemented with IPTG ( 0 . 4 mM final concentration ) and grown overnight . The recombinant proteins were purified from a lysate on Ni-NTA columns ( Qiagen ) followed by gel filtration to over 95% purity . DHFR kinetic parameters were measured by progress-curve kinetics , as described in ( Fierke et al . , 1987 ) . Purified enzymes ( 10 nM ) were pre-incubated with 120 μM NADPH in MTEN buffer ( 50 mM 2- ( N-morpholino ) ethanesulfonic acid , 25 mM tris ( hydroxymethyl ) aminomethane , 25mM ethanolamine , and 100 mM sodium chloride , pH7 ) . Reaction was initiated by addition of dihydropholate ( 20 , 15 , 10 μM final concentration ) and monitored till completion at 25°C following the drop in absorbance signal at 340 nm ( NADPH disappearance ) in Carry 60 spectrophotometer ( Agilent ) . The kinetics parameters ( kcat and Km ) were derived from progress-curves analysis using Global Kinetic explorer ( Johnson et al . , 2009 ) . Thermal stability was characterized by Differential Scanning Calorimetry ( DSC ) , essentially as described [56] . Briefly , DHFR proteins in Buffer A ( 10 mM potassium-phosphate buffer pH8 . 2 supplemented with 0 . 2 mM EDTA and 1 mM beta-mercaptoethanol ) were subjected to temperature increase at a 1°C/min rate between 20 to 80°C ( nano-DSC , TA instruments ) , and the evolution of heat was recorded as a differential power between reference ( buffer A ) and sample ( 40 μM protein in buffer A ) cells . The resulting thermograms ( after buffer subtraction ) were used to derive apparent thermal transition midpoints ( Tmapp ) from the peak in heat consumption . Genes encoding the orthologous DNA sequences were cloned into pBAD expression vector ( EMBL ) under the control of arabinose inducible promoter , and transformed into wild-type E . coli strain . For complementation assay , pBAD plasmid carrying E . coli’s DHFR ( WT ) was transformed into strains carrying orthologous replacements of the folA gene . The transformants were grown from a single colony overnight at 30°C in the supplemented M9 medium ( + 100 μg/ml amplicillin ) , diluted 1/100 and grown at 37°C in presence of 0 . 01% arabinose to achieve 50–100 fold increase in expression , relatively to the baseline chromosomal expression ( for pBAD-phylogeny DHFR ) , or without the inducer to achieve 6–8 fold increase in expression , relatively to the baseline chromosomal expression ( for pBAD-wtDHFR ) ( see S2B Fig ) . Growth rates were determined as described above . Strains were transformed with pUA66 plasmid carrying folA promoter fused to GFP coding gene ( Zaslaver et al . , 2006 ) . Promoter activity is defined as a ratio between fluorescent signal ( excitation 495 nm , emission 510 nm ) and biomass production ( measured as OD at 600nm ) . Cells were grown in supplemented M9 medium for 4 hours at 37°C , chilled on ice for 30 min and lysed with BugBuster ( EMD Millipore ) . DHFR amounts in the soluble fraction were determined by SDS-PAGE followed by Western Blot using mouse anti-His-Tag monoclonal antibodies ( Rockland Immunochemicals ) and goat anti-mouse polyclonal secondary antibodies conjugated with WesternDot 625 ( Life Technologies ) . Total Lon amounts were quantified using anti-Lon antibodies ( a generous gift from the T . Baker’s lab ) following the procedure described in [57] Genomic DNA was extracted using E . Z . N . A Bacterial DNA kit ( Omega Bio-Tek ) following the manufacturer’s instruction . Sequencing was performed on whole-population samples on Illumina MiSeq in 2x150 bp paired-end configuration ( Genewiz , Inc . , South Plainfield , NJ ) . Sequencing was also performed on samples derived from an individual colony from one evolutionary trajectory ( out of eight ) . The raw data were processed with the breseq pipeline [58] on default settings , using the E . coli K-12 MG1655 reference genome ( GenBank accession no . U00096 . 3 ) with a modified folA locus for each strain . Average coverage was ~100–200 bases for each sample . Only mutations with >20% frequency were considered . We rejected variants that only appeared on reads aligning to a single strand as well as variants in loci with significant alignment problems . We confirmed all variants using a separate alignment method ( BWA ( http://arxiv . org/abs/1303 . 3997v2 ) and SAMtools [59] ) and manually inspected each variant using IGV [60] . The results are summarized in S4 Table . To validate WGS data , we designed specific primers to amplify the genomic alleles of interest either from whole populations or independent colonies . Primers GGTGCGTTTGCCGGTCTGGATAAAGTG ( forward primer ) and CGGTGCCGTCAGGCAGTTTCAGCATC ( reverse primer ) were used to amplify the ClpX-Lon intergenic region , while primers GTGAAGCACAGTCGTGTCATCTG ( forward ) and CACTTGAATCCTTCAAGGTACGAACGCG ( reverse ) were used to amplify the lon ORF . The PCR products were subsequently sequenced by Sanger method . For global proteome analysis , whole populations of the evolved strains cells were lysed into 50 mM NaH2PO4 buffer ( pH8 ) supplemented with BugBuster extraction reagent and benzonase ( EMD Millipore ) , and soluble fraction was separated by centrifugation . Soluble cell lysates were trypsinized overnight by Promega ( Madison , WI ) Trypsin/Lys-C enzyme mixture with ratio 1:30 enzyme to protein and labeled with TMT reagent ( TMT , Thermo , San Jose , CA ) followed by nano LC-MS/MS separation and analysis ( for detailed description of the method see [34] ) . Z-scores of the log of relative ( to wild-type ) protein abundance ( LRPA ) were obtained according to Eq ( 1 ) : zistrain=Yistrain−〈Ystrain〉σYstrain ( 2 ) where index i refers to gene , Yistrain=log ( AistrainAiwt ) is LRPA for gene i , 〈Ystrain〉 denotes an average quantity Yi over all genes for a given strain or condition in corresponding experiments , and σYstrain is a standard deviation of Yistrain . After grouping genes according to Sanguardekar et al . [40] , we employed the Kolmogorov-Smirnov test in order to check how much a certain gene group changed upon serial passaging . First , we determined the direction of adaptation by comparing the average z-values of the naive and evolved proteome sets of a given gene group . If the average z-value increases , it is considered up-regulation; otherwise , it is considered down-regulation . Then , we applied the two-sample Kolmogorov-Smirnov test on the two sets , which provides the p-value for the null hypothesis that the two sets were drawn from the same distribution . Hence , we can quantitatively interpret a lower p-value as an indication that the two sets have more different distributions of z-values .
|
Horizontal gene transfer ( HGT ) is central to bacterial evolution . The outcome of an HGT event ( fixation in a population , elimination , or separation as a subdominant clone ) depends not only on the availability of a new gene but crucially on the fitness cost or benefit of the genomic incorporation of the foreign gene and its expression in recipient bacteria . Here we studied the fitness landscape for inter-species chromosomal replacement of an essential protein , dihydrofolate reductase ( DHFR ) encoded by the folA gene , by its orthologs from other mesophilic bacteria . We purified and biochemically characterized 33 out of 35 orthologous DHFRs and found that most of them are stable and more catalytically active than E . coli DHFR . However , the inter-species replacement of DHFR caused significant fitness loss for most transgenic strains due to low abundance of orthologous DHFRs in E . coli cytoplasm . Laboratory evolution resulted in an increase in orthologous DHFR abundance leading to a dramatic fitness improvement . Genomic and proteomic analyses of “naive” and evolved strains suggest a new function of protein homeostasis to discriminate between “self” and “non-self” proteins , thus creating fitness barriers to HGT .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Protein Homeostasis Imposes a Barrier on Functional Integration of Horizontally Transferred Genes in Bacteria
|
Neuronal microcircuits generate oscillatory activity , which has been linked to basic functions such as sleep , learning and sensorimotor gating . Although synaptic release processes are well known for their ability to shape the interaction between neurons in microcircuits , most computational models do not simulate the synaptic transmission process directly and hence cannot explain how changes in synaptic parameters alter neuronal network activity . In this paper , we present a novel neuronal network model that incorporates presynaptic release mechanisms , such as vesicle pool dynamics and calcium-dependent release probability , to model the spontaneous activity of neuronal networks . The model , which is based on modified leaky integrate-and-fire neurons , generates spontaneous network activity patterns , which are similar to experimental data and robust under changes in the model's primary gain parameters such as excitatory postsynaptic potential and connectivity ratio . Furthermore , it reliably recreates experimental findings and provides mechanistic explanations for data obtained from microelectrode array recordings , such as network burst termination and the effects of pharmacological and genetic manipulations . The model demonstrates how elevated asynchronous release , but not spontaneous release , synchronizes neuronal network activity and reveals that asynchronous release enhances utilization of the recycling vesicle pool to induce the network effect . The model further predicts a positive correlation between vesicle priming at the single-neuron level and burst frequency at the network level; this prediction is supported by experimental findings . Thus , the model is utilized to reveal how synaptic release processes at the neuronal level govern activity patterns and synchronization at the network level .
Oscillatory activity patterns in the brain have been linked to sleep , sensorimotor gating , short-term memory storage and selective attention [1 , 2] . Neuronal microcircuits in the brain spontaneously generate oscillatory activity patterns via synaptic interaction between groups of neurons [1 , 2] . Indeed , changes in synaptic transmission cause alterations in neuronal firing and neuronal network activity [3–5] , and synaptic dysfunction can lead to pathological epileptic conditions [6–8] . Even though small alterations in synaptic transmission and in the firing properties of single neurons can alter the spontaneous and evoked activity of entire neuronal circuits [3 , 9] , most computational models of neuronal networks do not explicitly account for the elaborate presynaptic neurotransmission process . Presynaptic transmission is a regulated multistep process that encompasses the loading of neurotransmitters into synaptic vesicles , the translocation to and docking of those vesicles at the plasma membrane ( PM ) , and vesicle preparation for fusion through a calcium-dependent maturation process generally referred to as "vesicle priming" [10–14] . This pool of primed vesicles is the readily releasable pool ( RRP ) , where vesicles undergo immediate fusion with the PM upon acute elevation in intracellular calcium concentration ( [Ca2+]i ) . Another presynaptic pool of vesicles , the recycling pool ( ReP ) , accommodates unprimed vesicles which can undergo maturation and fusion during repetitive synaptic stimulation; all of the remaining vesicles in the presynaptic terminal belong to the reserve pool ( RP ) . Equilibrium of the presynaptic vesicles transition between these pools depends on neuronal activity , synaptic proteins and calcium [15–19] . In the synapses , there are three types of synaptic release modes that rely on the high dynamic range of [Ca2+]i and share the same vesicle pools [20 , 21] ( but see [22 , 23] ) . They are defined by their temporal association with the action potential ( AP ) : a ) synchronous release , driven by a short-lived acute increase in [Ca2+]i , is time-locked to the AP [24–26]; b ) asynchronous release begins several milliseconds after an AP and drives slower vesicle release; this rate is two orders of magnitude slower than that of synchronous release . Asynchronous release is enhanced by slow clearance of residual calcium from the presynaptic terminal , as well as by strontium application [24]; c ) spontaneous release which emerges without any association to previous neuronal activity . Although presynaptic transmission is well understood at the single-neuron level , it is unclear how the aforedescribed manipulation of presynaptic processes modulates the activity patterns and synchronization of the network . Recently , a handful of studies have begun to investigate how manipulations of non-synchronous presynaptic release , such as asynchronous or spontaneous release , modulate neuronal network activity [3 , 6–8 , 24 , 25 , 27–30] . Understanding the determinant properties of spontaneous activity of the neuronal network is highly complex . Therefore , neuronal network computational models are utilized to simulate key features of the network's spontaneous activity . A large group of simulations utilizes computationally light leaky integrate-and-fire ( LIF ) neurons to model the activity of large-scale neuronal networks [31 , 32] . However , these neuronal models are based on very general assumptions regarding neuronal synaptic transmission processes and thus do not simulate critical synaptic mechanisms , such as the transition of vesicles between pools , vesicle maturation steps or calcium-dependent presynaptic release . An important model for neuronal networks , which combines the concept of synaptic resources and neuronal activity , is the tri-state model [33] . The original model , based on three kinetic equations , organized synaptic resources into three states: active , recovered or inactive . Synaptic transmission in this model was determined by the available synaptic resources ( i . e . vesicles ) and a constant utilization factor ( i . e . calcium , according to the calcium-based synaptic release theory ) . This model was later extended to include an increase in the utilization factor as the neuron keeps firing [34–36] , much like the increase in [Ca2+]i occurring in short-term synaptic plasticity . Another extension of the model also included asynchronous synaptic transmission by adding a super-inactive state [29 , 30 , 37 , 38] to generate reverberatory activity in small networks . Nonetheless , this model does not directly simulate the presynaptic vesicle pools , calcium-dependent vesicle priming or calcium-dependent release , which are basic and crucial properties of presynaptic release [25 , 39–42] . Furthermore , in response to evoked stimulations , this model generates very short network oscillations ( each oscillation lasting several milliseconds ) , which are several orders of magnitude shorter and more frequent than the network bursts recorded in vitro ( typically several hundreds of milliseconds of recurrent network activity ) [43 , 44] . In general , none of these models simulate spontaneous release , which is physiologically important [45–47] , and spontaneous activity in these models is generally achieved by artificial injection of current [48–50] . In addition , a model that relates in detail to changes in synaptic processes , and provides a mechanistic explanation and prediction for how changes in synaptic mechanisms at the neuronal level govern the activity patterns and synchronization at the network level is lacking . In this paper , we present a novel computational model that demonstrates how changes in synaptic transmission modulate neuronal network activity patterns . We utilized experimental data from in vitro neuronal networks cultured on microelectrode arrays ( MEA ) that spontaneously generate network-wide synchronized activity patterns , termed network bursts . We used the model to learn about synaptic mechanisms that can explain changes in neuronal network activity following manipulations of the presynaptic release processes [3 , 5 , 43 , 44 , 51] . Our model attempts to strike a balance between detailed cellular models and simplified neuronal network models [15 , 21 , 26 , 52] by extending the LIF neuronal model to simulate both the presynaptic release process and the entire neuronal network . This allowed us to examine how manipulations of specific steps in the presynaptic release mechanism affect neuronal network activity . The model provides putative mechanistic explanations for various network activity patterns in vitro , such as network burst termination , and allows us to predict how changes in the presynaptic release machinery will affect network oscillation frequency .
We previously explored [3] how genetic and pharmacological manipulations of presynaptic release change the spontaneous activity of neuronal networks cultured on MEA plates ( Figs 1A and S1 ) . To do so , we genetically and pharmacologically manipulated different synaptic transmission steps in cultured neuronal networks and examined the effects on neuronal network activity . Pharmacological enhancement of asynchronous release was achieved by strontium application , which has been shown to activate calcium-dependent release mechanisms but is cleared from the presynaptic terminal more slowly than calcium [53 , 54] . Genetic manipulations consisted of overexpressing DOC2B , a presynaptic protein that enhances spontaneous and asynchronous neurotransmitter release [55–57] . Our findings suggested that higher levels of asynchronous release at single synapses , induced by DOC2B overexpression or by strontium application , increase the firing rate within a network burst; on the other hand , facilitation of spontaneous release frequency by overexpression of the DOC2BD218 , 220N mutant [3] led to lower network burst firing rate ( Figs 1B and S1 ) . These findings join other studies that have shown that manipulation of presynaptic proteins has a substantial impact on neuronal network plasticity , information transfer and animal behavior [10 , 58 , 59] . However , it is difficult to infer a mechanistic explanation for these findings . Therefore , we developed a computational model that simulates how changes in different steps of synaptic transmission alter neuronal firing . The model consisted of 800 LIF neurons , spread on a virtual MEA-like 2D surface ( 30% inhibitory neurons; Fig 1C ) [3] . The neurons were connected by the small-world and scale-free topology typically associated with cortical neuronal networks [60–63] ( S2 Fig ) , creating an active neuronal network . A key feature of the model was that neuronal activity and synaptic release were generated from a presynaptic compartment that simulates the multistep process of calcium-dependent synaptic transmission . This presynaptic compartment was simulated for each LIF neuron ( Eq 1 ) and governed the spontaneous , evoked and asynchronous activity of each neuron in the network . All of the chosen parameters were based on up-to-date papers ( Table 1 ) [63] . Our model allowed us to perform in silico experiments , manipulate specific properties of synaptic transmission and study their impact at the network level . It gave us access to multiple cellular parameters , such as vesicular pool capacities , vesicle replenishment rate and [Ca2+]i , and simultaneously follow the macroscale network activity and the interaction between neurons . Each neuron received multiple inputs which accumulated as changes in the PM voltage until they crossed a threshold ( Eq 2 ) and generated an AP or decayed with a predefined time constant ( Table 1 ) . AP generation induced a transient increase in the [Ca2+]i that accumulates when several APs arrive concomitantly ( Eq 4 ) . This increase in calcium was then translated into vesicle release according to a calcium-dependent synaptic release curve ( Fig 1D ) . The release curve ( described in Eq 5 ) linked the free synaptic [Ca2+]i to synaptic release probability ( Pr ) according to well-established release-rate curves [21 , 25 , 26] . According to most calcium-dependent release models , upon AP generation , calcium level increases by almost two to four orders of magnitude in the active zone , inducing an acute shift in the synaptic Pr [25 , 26] . Accordingly , we used the Calyx of Held calcium-dependent release-rate curve as previously described [26] with a small modification to fit the lower Pr of cortical synapses . To recreate the multiscale temporal dynamics of synaptic release , each synapse consisted of three vesicle pools: RP ( 170 vesicles ) , ReP ( 20 vesicles ) and RRP ( 10 vesicles ) ( Fig 1E ) ; the vesicle transportation between pools was bidirectional ( Eqs 7 and 8 ) . Following vesicle release , vesicles underwent refilling according to different rate constants ( Table 1 ) . A variety of neuronal preparations have demonstrated that vesicle recruitment in neurons is enhanced by elevated [Ca2+]i [41 , 64 , 65] , and this enhancement has been recognized as essential for maintaining adequate release during high-frequency bursts of activity [65 , 66] . Therefore , we adapted the rate of vesicle transition from the ReP to the RRP to a similar Michaelis–Menten-type equation ( Fig 1E black frame; Eq 7 ) which has been used to describe the calcium-dependent transition rate from the unprimed pool to the RRP in chromaffin cells [15 , 67] . Each vesicle fusion event contributes a positive or negative voltage upon release ( excitatory or inhibitory postsynaptic potential , respectively ) to the PM of the postsynaptic neuron . Notably , the basal activity in the model was maintained by spontaneous release driven from the Pr of the neuron under resting calcium levels ( Eq 5 ) . This method kept the network active and replaced the common route of keeping computational neuronal networks spontaneously active , i . e . injecting current into the neurons [48–50] . Comparison of network spontaneous activity between these two methods showed that calcium-dependent synaptic release generates network bursts which are more similar to those recorded from neuronal networks cultured on MEA ( S3 Fig ) . Recurrent network-wide bursting activity and abundant inter-burst activity can be observed in the color-coded raster plot of neuronal network spontaneous activity generated by the model ( Fig 1F ) . Hence , the model recreated a pattern of synchronized activity followed by a period of quiescence similar to that in the experimental recordings ( compare Fig 1F to 1A ) . Importantly , the model recreated both network-wide bursting activity ( "full" bursts; green box ) and bursting activity limited to subnetworks ( "aborted" bursts; black box ) . To test the stability and robustness of the network activity under various manipulations , we explored the response of the model to changes in its primary gain parameters: excitatory postsynaptic potential ( EPSP , voltage ) and connectivity ratio ( the percentage of actual connections out of all possible connections in the network; see considerations for choosing these parameters in Methods ) . Quantitative analysis of the basic model activity parameters , such as global and network burst spike rate , network burst frequency and network burst duration , was performed under different levels of the gain parameters ( S4 Fig ) . We found that the model is robust to two- to threefold changes in basic gain parameters while maintaining continuous spontaneous network activity but displaying changes in various network activity properties ( S4 Fig ) . We also showed that even increasing the number of neurons or the number of synapses in the model 10-fold does not change its basic bursting activity; the neuronal network still displayed network-wide bursts followed by periods of relative quiescence ( S5 Fig ) . The stability of the bursting activity of the network following changes in basic gain parameters ( and changes in the number of neurons and number of independent synapses per neuron ) established the robustness of the model and increased its fidelity . Indeed , most of the experimental manipulations did not abolish the basic bursting activity in the network but rather manipulated the inter-burst and intra-burst spiking profiles . This places the model in an excellent position to test the impact of changes in other parameters of synaptic release on the network bursting activity . The established model was utilized to understand two intriguing findings: elevated asynchronous release , but not spontaneous release , at the single-neuron level enhances and synchronizes network burst activity [3]; on the other hand , enhanced spontaneous release reduces synchronization and network burst activity . Experimentally , asynchronous release was elevated by either DOC2B or strontium . Strontium has been suggested to trigger vesicle fusion and neurotransmitter release in the same way as calcium , but is extruded from the synapse more slowly than calcium , causing long-lasting vesicle fusion or asynchronous neurotransmitter release [53 , 68] . Therefore , to mimic the effect of asynchronous release , we reduced the rate of calcium efflux out of the synapse ( Eq 4 , τCafast and τCaslow ) , allowing more time for vesicle fusion [69] . It is important to note that we changed the asynchronous release in both excitatory and inhibitory neurons . We first verified that slower calcium clearance increases the ratio of asynchronous to synchronous release in the model . We followed the change in the probability for vesicle release from single neurons up to 50 ms after an AP , under different calcium-efflux rates ( Fig 2A; see Methods ) . The ratio of asynchronous to synchronous release ( Fig 2A right panel; 'ASync' and 'Sync' , correspondingly ) increased as calcium efflux was reduced [70] . Notably , the increase in asynchronous release did not increase the total neuronal output of a single neuron but only spread the release over a longer time . We then examined how asynchronous release affects the activity profile in the network burst ( Fig 2B ) . Gradually increasing asynchronous release in the model enhanced the network burst firing rate ( Fig 2B; +100% , left panel ) , similar to the experimental results of increasing strontium concentration ( Fig 2B right panel ) . Both manipulations also decreased the time from burst onset to its peak . Interestingly , even when we increased the number of neurons in the network 10-fold ( 8000 instead of 800 ) and also when we increased the number of synapses per neuron 10-fold ( 10 instead of 1 ) , enhanced asynchronous release facilitated network burst firing rate and decreased the network burst's time to peak ( S5 Fig ) . This supports the power of the model in mimicking experimental results and suggests that asynchronous release has a profound effect on neuronal network activity . Next , we focused our analysis on the manipulation of spontaneous release and tested its effects on the network activity . Experimentally , spontaneous release was increased by overexpressing a DOC2B mutant , DOC2BD218 , 220N , that is known to increase spontaneous release [55] . Computationally , spontaneous release was elevated by increasing the Pr at resting calcium ( Fig 2C top panel; Eq 5 ) . This manipulation increases the probability of vesicle release under resting conditions , which is the basic definition of spontaneous release [25] . The increase in spontaneous release in the model led to a significant decrease in the network burst activity , as evidenced by the reduced network burst activity profile and the lower global spike rate in each network burst ( Fig 2C and 2D ) . This manipulation recreated the experimental data of DOC2BD218 , 220N overexpression ( Fig 2C; compare bottom left panel , model , to bottom right panel , experiment ) while reducing the number of spikes and the number of neurons in the network bursts ( Fig 2D ) . Comparison of the changes induced by both manipulations established their opposite effects on network activity ( Fig 2D ) ; while asynchronous release was positively correlated with network burst activity , spontaneous release was anticorrelated . This means that specific activity properties can change in the same direction by an increase in asynchronous release or a decrease in spontaneous release , or vice versa . These opposite effects were more prominent in the global spiking rate and network burst spikes; however , the burst rate , for example , displayed a more prominent difference between spontaneous and asynchronous release upon an increase in the corresponding parameter ( Fig 2D ) ; while higher spontaneous release reduced network burst frequency , lower spontaneous release did not change it ( Fig 2D , bottom panel ) . Therefore , it is important to examine the combination of various network activity parameters to determine the overall effect on the network activity . Next , we examined whether the model recreates the higher-level effects on network activity patterns observed in the experimental results [3] . Evidently , higher asynchronous release in the model significantly increased , while spontaneous release reduced the ratio of neurons participating in the network bursts ( S6 Fig ) . This was measured by classifying network bursts into "full " or "aborted" bursts [3 , 44] . Moreover , analysis of the normalized network burst synchronization in the simulation showed that elevated asynchronous release also increases network burst synchronization , primarily around the peak of the network burst ( S6 Fig ) . These analyses were in agreement with the experimental findings and showed that the model successfully recreates the response to the manipulation of asynchronous and spontaneous release . Thus , using the in silico model , we manipulated specific steps in the release process and linked them to specific experimental changes . Hence , the model reaffirmed a wide range of experimental analyses , from basic firing rate to high-level network synchronization parameters . The high reliability of the in silico model in reconstructing experimental findings allowed us to utilize it to explore the neuronal mechanisms underlying the findings and uncover the model parameters and factors that govern network activity . Specifically , the model allowed us to follow neuronal parameters , such as changes in the various vesicle pools , which are unavailable experimentally . We analyzed the vesicle pool dynamics and [Ca2+]i of the model neurons under baseline release levels ( Baseline ) and under enhanced asynchronous release ( +100% ) . Fig 3A demonstrates changes in the number of RRP vesicles in 4 representative neurons throughout a single network burst . Each neuron displayed different release patterns from the RRP but all displayed a certain degree of vesicle depletion ( Fig 3A , 3B and 3D ) . Analysis of the average RRP occupancy in all neurons in all network bursts ( in 10 simulations ) showed that during the burst , the RRP are depleted by the same percentage under both baseline and enhanced asynchronous release conditions ( Fig 3B ) . Further analysis of the average RRP content showed that most neurons have more than 4 vesicles in the RRP at the onset of the network burst and less than 2 vesicles at its termination ( out of a maximum occupancy of 10 vesicles in the RRP; Fig 3D ) . It can be suggested that under these conditions , where more than 70% of the neurons have less than 2 vesicles left in the RRP ( i . e . less than 20% of the entire synaptic reservoir is available ) , network bursts are terminated . This is not surprising but rather provides a clear connection between vesicle pool depletion and burst termination and a mechanistic explanation for previous experimental results [4 , 71] . This analysis could not explain the increase in network activity under enhanced asynchronous release and therefore we continued to examine the changes in ReP dynamics , which transfers vesicles to the RRP through calcium-dependent vesicle priming . The same analysis applied to the ReP showed that asynchronous release manipulation causes enhanced consumption and larger depletion of vesicles from this pool ( Fig 3C ) ; while only 7% of the neurons had less than 8 vesicles in the ReP at the time of network burst termination under baseline conditions , ~46% of the neurons had less than 8 vesicles under enhanced asynchronous release at the time of network burst termination ( Fig 3E ) . On average , approximately 2–3 additional vesicles were consumed from the ReP during a network burst under enhanced asynchronous release ( an increase of 10–15% in total synaptic release , on average; Fig 3D ) . This suggests that the ReP is the source for the higher output following elevated asynchronous release and that asynchronous release , driven by slower calcium clearance , relies on the replenishment rate of the ReP for support of the increased network activity . To examine this hypothesis , we determined the average cumulative neuronal output throughout the burst ( Fig 3F ) . On average , each neuron with a higher asynchronous release contributed ~2 more vesicles within the first 300 ms of the burst overall . This accumulated increase underlies the higher network activity and synchronization during the bursts; it also supports our hypothesis that the ReP is the source vesicle pool contributing to this network effect . The lower calcium efflux rate from the presynaptic terminal allows faster and larger accumulation of free calcium throughout the network burst ( Fig 3G ) . This , in turn , has two important implications in the neuronal release dynamics throughout the burst: 1 ) higher calcium levels lead to higher Pr; 2 ) higher calcium levels increase the vesicle transition rate from ReP to RRP ( much like the calcium-dependent vesicle replenishment hypothesis ) . Thus , the model revealed that the higher asynchronous release temporally increases the Pr and vesicle availability , causing enhanced neuronal network activity only during bursts . Furthermore , this analysis pointed to the ReP as the vesicle pool that supports this increase in neuronal vesicle release and network synchronization . Our model presented us with an opportunity to predict the effect of manipulations , which can be later examined experimentally . We were therefore interested in testing how changes in priming rate at the single-neuron level affect network activity . To implement this manipulation , we changed the maximum rate of vesicle transition from ReP to RRP ( Fig 4A , circled red marker; parameter 'Rmax' in Fig 1E; Eq 7 τReP→RRP ) . A comparison of raster plots showed that as the priming rate increases , the activity and frequency of the bursts are enhanced , while decreasing the priming rate reduced network activity ( Fig 4B ) . Burst profile and activity parameter analyses supported these findings , suggesting that a 50% increase in priming rate would lead to ~30% increase in the maximum firing rate within the network burst ( Fig 4C and 4D ) and an increase of 4 bursts per minute in network burst frequency , i . e . the network displays a higher rate of oscillations following this manipulation without elevating the inter-burst activity . Next , we overexpressed Munc13-1 , a positive regulator of vesicle priming rate , in neuronal networks plated on MEA , and found that a 2-fold increase in Munc13-1 expression levels increases the frequency of network bursts by 60% ( previously published in Lavi et al . [3]; Fig 4E ) . These results match the model predictions ( Fig 4B , High priming ) and suggest that changes in vesicle priming rate at the neuronal level tune the burst frequency at the network level .
Current computational network models do not simulate synaptic vesicle pools or calcium-dependent processes . Many computational network models based on LIF neurons simulate neuronal activity as the sum of voltage or current input on the PM to recreate neuronal network activity . Although this voltage accumulation causes the generation of an AP in the soma , it is the calcium influx through voltage-dependent calcium channels in the presynaptic terminal that drives the actual vesicle fusion and subsequent synaptic release [72 , 73] . Therefore , free intracellular calcium dynamics gates the transfer of synaptic information from one neuron to the next , and the combination of calcium dynamics and vesicle release probability underlies short-term plasticity in the presynaptic terminal , a key mode of operation in several central synapses [41 , 74–78] . Therefore , it is highly important to integrate calcium-dependent synaptic release , as we did in the current model , into LIF neuronal models . Although the well-established tri-state model [33] did incorporate synaptic transmission into LIF neurons , that model and its succeeding extensions [29 , 30 , 34 , 36–38] did not simulate synaptic vesicle pools , calcium-dependent vesicle priming or vesicle release . The evoked activity simulated in those models generated very short network oscillations ( several milliseconds ) , significantly shorter than the network bursts observed in vitro in dissociated neuronal cultures by single-neuron current-clamp recordings and neuronal network MEA recordings ( typically hundreds of milliseconds ) . Furthermore , to maintain spontaneous network activity , current or voltage are artificially injected from an external source [48–50] . The lack of biological mechanisms in the neuronal model makes it harder to infer physiologically relevant consequences and predictions . The uniqueness of our model lies in its direct simulation of key components of the presynaptic release process , thereby revealing how changes in the release process , such as changes in release probability ( Eq 5 ) , vesicle pool size ( Eqs 7 and 8 ) , and calcium dependency ( Eqs 4–7 ) , affect neuronal network activity . This direct simulation allows us to model synchronous , asynchronous and spontaneous release as derivatives of the same calcium-dependent release mechanism with different ranges of [Ca2+]i [25 , 26 , 41] . The model also incorporates calcium-dependent vesicle priming ( Eq 6 ) , which is usually not modeled in neuronal network models . Since our model directly simulates the presynaptic vesicle pools and calcium-dependent priming based on measured rate constants , the derivation of putative physiologically relevant mechanistic explanations from its predictions is more intuitive . The model generates network bursts that are similar to those observed in MEA recordings ( S3 Fig ) in duration and firing rate , thereby enabling an investigation of mechanisms for burst termination under spontaneous neuronal activity , and linking them to the dynamics of vesicle pool depletion [79] . Note that we are not claiming that it is impossible to create network bursts without incorporation of the presynaptic release mechanism , but rather that through the structure of our model , we were able to relate the network bursts to their underlying realistic and biologically plausible presynaptic mechanisms . Our simulation allowed us to perform long-term in silico experiments ( which we limited to several hours ) , while the model clearly exhibited stability and robustness to changes in the primary gain parameters that control network activity—i . e . , EPSP and connectivity ratio . In agreement with our experimental results , most of the manipulations performed in the model did not abolish the basic network bursting activity but rather manipulated the inter-burst and intra-burst spiking distributions . The fact that the basic bursting activity of the model was not diminished after these manipulations establishes the model's robustness and its provision of a stable platform to uncover the role of asynchronous and spontaneous release in neuronal network oscillatory activity . A recent intriguing experimental finding demonstrated that asynchronous release , but not spontaneous release , enhances network activity and network burst synchronization [3] . The model allowed us to test how changes in asynchronous release and spontaneous release affect network activity at the neuronal level . Supported by experimental results , the model showed that higher spontaneous release leads to lower firing rate , lower neuronal participation in network bursts and lower frequency of bursts . Higher spontaneous release reduced synchronization of the network activity by the superfluous release of vesicles throughout; this excess activity reduced the availability of releasable vesicles from the RRP during network bursts , which resulted in lower intra-network burst activity and synchronization . The model also allowed testing whether the anticipated change following strontium application—enhanced asynchronous release—is translated into enhanced activity during the bursts , and investigating the vesicular source for this effect . The model-simulated increase in asynchronous release elevated the overall activity of the network and various network burst parameters , including synchronization and neuron participation . This was in agreement with the experimental results obtained following gradual application of strontium . These results are supported by previous evidence regarding the link between asynchronous release and reverberatory activity [54] . Previous studies have shown that rapid recovery of the RRP supports asynchronous release at the neuronal level [71] , and have suggested that network bursting activity depends on the vesicle depletion rate from the RRP [4 , 80] . Here we suggest that during ongoing network activity , the neurons in the network are not fully depleted at the termination of the network burst . Rather , examination of the RRP and ReP of all neurons in the network showed that under baseline conditions , it is sufficient that 70% of the neurons have less than 2 vesicles in the RRP ( that is , less than 20% of the overall vesicles available for immediate release in the RRP ) to terminate the network burst . Analysis of the average depletion rate of the ReP and RRP ( over all bursts and all neurons in 10 simulation sets ) following higher asynchronous release in the model suggested that the main resources for the enhanced network activity come from the ReP . This was concluded from the fact that higher asynchronous release did not increase the total number of vesicles released from the RRP but it did increase consumption of the ReP . Interestingly , we found excellent agreement between the degree of consumption of the ReP and enhancement of release at the neuronal level: asynchronous release enhanced depletion from the ReP by ~2 vesicles and respectively , this higher asynchronous release increased the total amount of vesicles released from each neuron by approximately 2 extra vesicles , on average . These 2 extra vesicles , on average , per neuron ( representing 10% of the maximum ReP occupancy ) accumulated and induced large-scale enhancement of neuronal firing during the bursts . This means that as previously suggested [9] , a relatively small change in presynaptic release results in profound changes in the network activity . As synaptic release in the model depends on the intracellular level of free calcium , we followed the change in [Ca2+]i during the network bursts; we found that following the asynchronous release manipulation , the calcium concentration reaches higher levels throughout the bursts . These calcium levels increased the effective Pr and vesicle replenishment rate in each neuron in the network ( due to calcium-dependent vesicle priming ) . These finding are in agreement with previous studies showing that sustained synaptic release requires the contribution of vesicles from the ReP [42] . Together , these analyses explain how additional vesicles are quickly primed from the ReP into the RRP ( due to faster priming rate ) and are readily released ( due to higher Pr ) , leading to faster vesicle replenishment and an overall higher neuronal output only during the burst , when calcium levels are high . What might the effects of higher asynchronous release be on neuronal microcircuits in the brain ? Measurements from rat cortical acute slices and from human cortical slices have shown that cortical fast-spiking inhibitory neurons exhibit asynchronous release as part of their spontaneous activity . Furthermore , the involvement of excess asynchronous release in inhibitory fast-spiking neurons has been linked to epileptic activity in human patients with intractable epilepsy and in the rat pilocarpine model of status epilepsy [7] . This shows that asynchronous release is a fundamental property of synaptic transmission in the brain and not merely induced by drug application [7] . Recent evidence from rats suggests that the excitation-to-inhibition ratio in the adult brain is regulated by reduced GABAergic asynchronous release , which is supported by the more efficient clearance of residual calcium [6] . This might cause alterations in brain network activity in a mechanism that we simulated for in vitro networks . This evidence joins a previous computational model which showed that the higher levels of asynchronous GABAergic release in the cortex of juvenile animals are counterbalanced by postsynaptic shunting inhibition to regulate synaptic transmission in the developing brain [29] . Interestingly , lack of synchronous release but enhanced asynchronous release following synaptotagmin-1 knockout in the hippocampal CA1 region did not impede acquisition of contextual fear memories , but did impair their precision . This suggests that the hippocampal CA1 region can rely on spike bursts to transfer information downstream [58] . These and other recent studies [3 , 5 , 8 , 81–83] demonstrate the importance of understanding in detail how changes in various types of synaptic release at the single-neuron level regulate the activity of the neuronal network in brain function and dysfunction , and further stress the importance of integrating the presynaptic release mechanism into neuronal network computational models . As suggested above , asynchronous release utilizes ReP vesicles to increase network activity . This indicates the important role of vesicle priming rate , as this process regulates the rate of vesicle transfer from the ReP to the RRP . Since the model can be utilized to recreate experimental data and examine parameters that are inaccessible experimentally , we manipulated the maximum vesicle priming rate ( ‘Rmax’ in Eq 7 ) . Manipulation of vesicle priming revealed a positive correlation between priming rate at the neuronal level and burst frequency at the network level . This model prediction was supported by the viral overexpression of Munc13-1 , a presynaptic protein that positively regulates vesicle priming , in neuronal networks cultured on MEA [3] . Munc13-1's higher expression levels—twofold higher than baseline—are physiologically plausible , suggesting that tuning the priming rate might have a great impact on the activity of neuronal networks in general . To infer a direct connection between vesicle priming rate and network burst frequency , an additional experimental manipulation is required that will specifically and acutely reduce the vesicle priming rate; however , the present analysis already demonstrates the model's power in predicting how changes at the neuronal level are transformed to changes at the network level , and suggests manipulation of the frequency of network bursts by changes in the presynaptic release process . The implications of these manipulations for spontaneous neural activity in the neocortex of the behaving animal remain to be tested , together with their implications for learning and memory , as well as pathological disorders . Our model shows how in vitro network oscillations , in the form of network bursts , can be generated and maintained based on calcium-dependent presynaptic release mechanisms , without external stimulation or injection of current . Furthermore , the model recreates some of the complex experimental data obtained from MEA recordings . A growing body of literature is connecting network bursts in vitro to the "up" and "down" states displayed by neocortex brain oscillations in vivo [84–88] . Modulation of the oscillation between "up" and "down" states during spontaneous activity in vivo has been observed during slow-wave sleep , selective attention and short-term memory tasks [89–92] . Therefore , elucidating the principles of spontaneous network activity and its manipulation in culture might contribute to understanding high-order functions in the behaving animal [5 , 27 , 88 , 93] . Interestingly , the oscillatory nature of "up" and "down" states can be explained by a modulation of presynaptic release , and it has been suggested that while non-synchronous synaptic release might maintain the "up" state [92 , 94] , synaptic depression can be used to terminate it and return the activity to the "down" state [95] . Thus , although our model is based on in vitro experimental data , it opens new avenues to examining how presynaptic release mechanisms modulate microcircuit oscillations and subsequently affect higher neural functions such as slow-wave sleep , learning and attention , or are involved in pathologic neurological disorders .
|
The activity of neuronal networks underlies basic neural functions such as sleep , learning and sensorimotor gating . Computational models of neuronal networks have been developed to capture the complexity of the network activity and predict how neuronal networks generate spontaneous activity . However , most computational models do not simulate the intricate synaptic release process that governs the interaction between neurons and has been shown to significantly impact neuronal network activity and animal behavior , learning and memory . Our paper demonstrates the importance of simulating the elaborate synaptic release process to understand how neuronal networks generate spontaneous activity and respond to manipulations of the release process . The model provides mechanistic explanations and predictions for experimental pharmacological and genetic manipulations . Thus , the model presents a novel computational platform to understand how mechanistic changes in the synaptic release process modulate network oscillatory activity that might impact basic neural functions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion"
] |
[] |
2015
|
Shaping Neuronal Network Activity by Presynaptic Mechanisms
|
With a genome size of ∼580 kb and approximately 480 protein coding regions , Mycoplasma genitalium is one of the smallest known self-replicating organisms and , additionally , has extremely fastidious nutrient requirements . The reduced genomic content of M . genitalium has led researchers to suggest that the molecular assembly contained in this organism may be a close approximation to the minimal set of genes required for bacterial growth . Here , we introduce a systematic approach for the construction and curation of a genome-scale in silico metabolic model for M . genitalium . Key challenges included estimation of biomass composition , handling of enzymes with broad specificities , and the lack of a defined medium . Computational tools were subsequently employed to identify and resolve connectivity gaps in the model as well as growth prediction inconsistencies with gene essentiality experimental data . The curated model , M . genitalium iPS189 ( 262 reactions , 274 metabolites ) , is 87% accurate in recapitulating in vivo gene essentiality results for M . genitalium . Approaches and tools described herein provide a roadmap for the automated construction of in silico metabolic models of other organisms .
Genome-scale metabolic reconstructions are already in place or under development for a growing number of organisms including eukaryotic , prokaryotic and archaeal species [1] . Metabolic pathway reconstructions are increasingly being queried by systems engineering approaches to refine the quality of the resulting metabolic models [2] . Curated metabolic models are indispensable for computationally driving engineering interventions in microbial strains for targeted overproductions [3]–[6] , elucidating the organizing principles of metabolism [7]–[10] and even pinpointing drug targets [11] , [12] . Currently , over 700 genomes have been fully sequenced [13] whereas only about 20 organism-specific genome-scale metabolic models have been constructed [1] , [14] , [15] . Figure 1 pictorially demonstrates , in logarithmic space , the widening gap between organism-specific metabolic models and fully sequenced genomes over the past twelve years . It appears that metabolic model generation can only keep pace with about 1% of the fully sequenced genomes . In response to this flood of present and future genomic information , automated tools such as Pathway Tools [16] and SimPheny ( Genomatica ) have been developed that , using homology comparisons , allow for the automated generation of draft organism-specific metabolic reconstructions that can subsequently be upgraded into metabolic models . All of these models remain to some extent incomplete as manifested by the presence of unreachable metabolites [17] and some growth inconsistencies between model predictions and observed in vivo behavior [2] . In particular , optimization-based techniques for automatically identifying metabolites disconnected from the rest of metabolism ( i . e . , GapFind ) and hypotheses generators ( i . e . , GapFill ) for reconnecting them have recently been introduced [17] . In order to resolve substrate utilization prediction inconsistencies , Reed et al . [2] introduced a novel approach for identifying what reactions to add to the genome-scale metabolic models of E . coli to correct some of the in silico growth predictions . In our group , we have taken the next step for gene deletion data by attempting to correct all such growth inconsistencies by allowing not just additions but also eliminations of functionalities in the model ( i . e . , GrowMatch ) ( Satish Kumar and Maranas , submitted ) . As outlined in Figure 2 , in this work , we describe the application of these automated methodologies during the Mycoplasma genitalium model construction process ( as opposed to an a posteriori mode of deployment ) . M . genitalium has received considerable attention as it is the smallest organism that can be grown in pure culture , having a genome size of ∼580 kb and approximately 480 protein coding regions [18] , [19] . An examination of its genome content revealed limited metabolic capabilities [20] , leading researchers to suggest it may be a close approximation to the minimal set of genes required for bacterial growth [19] , [21] . Several researchers have carried out genomic and proteomic analysis of M . genitalium to quantify this minimal set . For example , Mushegian and Koonin have carried out a detailed comparison of M . genitalium and H . influenzae proteins to derive a set of 256 genes that they suggested are necessary for viability [22] . Further , genomic analyses of these species revealed that Mycoplasma genes encode for several catabolic and metabolite transport proteins but for only a limited number of anabolic proteins suggesting that Mycoplasma species need to scavenge for the required nutrients from the surrounding environment [20] . More recently , Glass and co-workers performed global transposon mutagenesis and established that 382 of the 482 protein coding sequences are essential genes for this minimal bacterium [19] . These gene sets and essential gene analyses , however , have not been put into context of a complete functional metabolic model . Mycoplasma genitalium is not only the closest known approximation of a minimal cell but also an important sexually transmitted human pathogen . It is a cause of nongonococcal urethritis in men and is associated with genital tract inflammatory diseases in women , including endometritis , cervicitis , pelvic inflammatory disease , and tubal factor infertility ( for a recent review see [23] ) . Additionally , evidence suggests that M . genitalium infection increases the risk of contracting HIV-1 [24]–[26] . Mycoplasmas , the generic name for the bacteria that comprise the Mollicutes taxon , evolved from the low G+C Gram positive bacteria through a process of massive genome reduction [27] . Their salient characteristics in addition to small genomes are a lack of a cell wall , and an almost complete inability to synthesize the building blocks of DNA , RNA , proteins , and cell membranes . The above underlines the importance of investigating the molecular biology of mycoplasma and M . genitalium in particular . However , a major hindrance to M . genitalium research and laboratory diagnosis of infection has been their cultivation in vitro . While defined media are present for some mycoplasmas [28]–[30] , researchers have often had to resort to complex media to cultivate most mycoplasmas , including M . genitalium . M . genitalium , and many other mycoplasmas are cultured in vitro in SP-4 medium . This extremely rich medium contains several undefined additives including peptones , yeast hydrolysate , yeast extract and 17% fetal bovine serum [31] . The use of complex undefined growth media has interfered with the molecular definition of mycoplasma metabolic pathways , genetic analyses , estimation of growth requirements , characterization of auxotrophic mutants and examining the nutritional control of bacterial pathogenecity . In this paper , we highlight the development of an in silico model of metabolism of M . genitalium . It was subjected to network connectivity gap detection and reconnection as well as restoration of consistency with in vivo gene essentiality experiments [19] . We subsequently used the model to pinpoint components in the growth medium that are needed for the production of all components of biomass in an effort to eventually eliminate the need for non-defined components such as serum in the growth medium .
The metabolic reconstruction of M . genitalium was carried out in a series of successive refinements ( see Figure 2 and Materials and Methods ) . Of the 482 predicted open reading frames ( ORFs ) , 113 ( 23% ) only have annotations of ( conserved ) putative or hypothetical proteins . Of the remainder , 369 ORFs have well-defined annotations , with functions either shown biochemically or predicted for 272 of them ( 42% ) [18] , [19] . From these well-annotated genes , 82 ( 17% ) are not involved in specific metabolic transformations , but rather encode proteins whose roles include DNA/RNA polymerization , DNA repair , protein folding and adhesion . The model construction process started with the application of an automated procedure for creating a draft metabolic reconstruction from the genome sequence of M . genitalium ( see Materials and Methods ) [32] , [33] . This auto-generated model contained 150 genes and 249 unique metabolites associated with 167 reactions ( see Table 1 ) . The 150 genes comprise 31% of the ORFs present in the genome and provided a solid starting point with very little manual effort . Additional homology searches of genes not included in the auto-model against the NCBI database increased the number of model components to 187 genes and 263 distinct metabolites associated with 179 reactions ( Table 1 ) . These genes comprise 39% of the ORFs present in the genome . These reactions enable , for example , the uptake of glycerol into the cell , thymidine kinase , and ribonucleotide diphosphate reductase ( Table 2 ) , as well as the remaining annotated ribosomal proteins that were not previously incorporated . As indicated in Table 2 , the bidirectional protein-protein BLAST ( i . e . , BLASTp ) expectation values exhibited by these genes when compared to the biochemically-characterized counterpart in other organisms provided strong support for their inclusion in the model . In addition , we also included nine nucleoside di- and tri-phosphate kinase associated reactions based on the observation that the kinase pool of M . genitalium has relaxed substrate specificity [34] . As part of the initial model generation , we also checked the BLASTp scores , gene annotations , and the cluster of orthologous groups ( COGs ) ontology [35] of all genes in the automodel , to guard against the erroneous inclusion of functions in the model ( see Materials and Methods ) . A metabolic reconstruction has been described as a 2-D annotation of a genome [32] . The generation of a computations-ready model requires the complete assignment of metabolites to reactions , inclusion of exchange reactions , resolution of gene-enzyme associations , and derivation of biomass equations . Here , we largely follow the steps put forth by [33] in the latest E . coli metabolic reconstruction . The computations-ready model along with the biomass description allows for the use of optimization-based techniques for testing and correcting for the presence of connectivity gaps ( step 3 ) and growth prediction inconsistencies ( step 4 ) . The initial model was constructed almost exclusively based on homology searches within model libraries . This procedure led to the presence of many network gaps [17] preventing 177 reactions ( 99% of total ) from carrying flux under all uptake conditions ( i . e . , they were blocked ) . As a consequence , these blocked reactions precluded the formation of some of the biomass components . Using GapFind [17] we found that a total of 175 ( 70% ) cytoplasmic metabolites could not be produced inside or transported into the intracellular space . These metabolites included a number of biomass precursor metabolites ( e . g . , some amino acids , cofactors and metal ions ) that had not been assigned uptake reactions . Of all the blocked metabolites , thirteen were involved in nucleotide metabolism and eight were metal ions without an identified transporter . We also note that 40 of these metabolites are charged/uncharged tRNA molecules , which are active in closed reaction cycles used in forming the protein component of biomass . Through the use of GapFill [17] we subsequently sought to bridge these network gaps through the addition of reactions , transport pathways and relaxation of irreversibilities of reactions already in the model . Reactions known not to be present in M . genitalium ( e . g . , an incomplete TCA cycle ) were excluded as gap filling candidates . We first applied GapFill to unblock constituents of biomass guided by the known components in the growth medium . We unblocked biomass production by adding 65 reactions , for which most ( i . e . , 43 ) were involved in metabolite transport , such as for the uptake of amino acids ( 14 ) , folate , riboflavin , metal ions ( 8 ) , and cofactors such as CoA . Among the remaining reactions were those responsible for the hydrolysis of dipeptides ( 15 ) and eight reactions involving other biotransformations . We performed an additional round of BLASTp comparisons of genes annotated with these reactions against the M . genitalium genome to determine if we could associate any of these reactions with specific genes in M . genitalium . We found five proteins catalyzing these reactions that had BLASTp scores smaller than 10−5 ( see Table 2 ) . For example , GapFill suggested the addition of reaction glutamyl-tRNA ( Gln ) amidotransferase in the model to allow the formation of the gln-tRNA molecule . BLASTp searches allowed us to link this activity with the genes encoding for the three subunits ( MG098 , MG099 and MG100 ) . Note that these three genes ( and others added during this step ) were not added earlier ( steps 1 and 2 ) on account of their ambiguous functional characterization . By bringing to bear both homology ( though BLASTp ) and connectivity restoration ( through GapFill ) , here we rely on multiple pieces of evidence when appending a new functionality and corresponding genes to the model . Even after unblocking biomass formation , 43 metabolites remained blocked and were subsequently analyzed by GapFill . The results from GapFill are summarized in Figure 4 . We were able to reconnect three metabolites by treating three reactions as reversible . We also found that the originally assigned ( based on the auto-model ) directionality of 1-acyl-sn-glycerol-3-phosphate acyltransferase was incorrect . It was subsequently reversed and found to be in accordance with both KEGG and MetaCyc entries . An additional 21 metabolites were reconnected by adding 18 reactions from the KEGG and MetaCyc databases ( see Materials and Methods ) . The addition of these 18 reactions also introduced an additional nine metabolites ( three of which were involved in glycerolipid metabolism ) to the model . Finally the incorporation of uptake/transport reactions reconnected an additional four metabolites . We performed an additional round of BLASTp comparisons and we were able to associate three out of 22 reactions with specific genes ( see Table 2 ) . We found that the associated gene ( i . e . MG066 ) for 1-deoxy-D-xylulose 5-phosphate synthase was already included in the model but with a different functionality ( i . e . , transketolase ) . The secondary synthase functionality , revealed by GapFill/BLASTp , was subsequently associated with gene MG066 in the model . A similar situation occurred with MG053 , which was already associated with phosphomannomutase in the model . In addition , gene MG259 ( annotated as “modification methylase , HemK family” in the Comprehensive Microbial Resource , http://cmr . jcvi . org ) was added to the model to carry out the glutamine-N5 methyltransferase activity elucidated by GapFill/BLASTp . The model statistics after correcting for network gaps are summarized in Table 1 . Based on in vivo gene essentiality data [19] we deduced that there are 174 essential genes and 19 non-essential genes among the 193 genes provisionally present in the model ( after steps 1 , 2 , and 3 ) . We note that the in vivo gene essentiality experiments were performed using non-defined medium containing serum and yeast hydrolysate among other rich components . During the in silico model predictions/comparisons , we allowed the uptake of all extracellular metabolites with transport reactions , except for sugars other than glucose , in order to computationally approximate this medium . Using a recently proposed diagnostic of the percentage of correctly-identified essential genes [38] , [39] , we found that the model correctly identified 137 out of a total of 174 essential genes ( i . e . , specificity of 79% ) and 16 out of a total of 19 non-essential genes ( i . e . , sensitivity of 84% ) . This implies that the model ( after steps 1 , 2 , and 3 ) was 79% correct in its overall accuracy in growth predictions ( i . e . , 153 of 193 ) . Most of the mismatches ( 92% ) were over-predictions of the metabolic capabilities ( i . e . , predicting growth when none is observed in vivo ) instead of under-predictions ( i . e . , predicting no growth when growth is observed in vivo ) . We subsequently deployed the GrowMatch method ( Satish Kumar and Maranas , submitted ) to rectify as many as possible of the erroneous essentiality predictions by the model . GrowMatch functions by identifying the minimal number of model modifications required to restore consistency between growth predictions and gene essentiality experiments ( see Materials and Methods ) . Model under-predictions include mutants ( MG410 and MG411 ) , which encode the subunits for the phosphate transporter , preventing in both cases the uptake of phosphate . This implies that M . genitalium must have an additional uptake route of phosphates . Even though GrowMatch suggested a number of phosphate uptake alternatives to resolve this conflict and Glass and coworkers [19] had posited the activity of a putative phosphonate transporter ( MG289 , MG290 , and MG291 ) , we decided not to add them to the model as no direct evidence exists to ascertain their presence . For instance , the putative phosphonate transporter might be nonspecific thus also enabling uptake of phosphate . Alternatively , the unidentified phosphonate substrate might be catabolized to yield phosphate through a number of reactions . The other incorrect under-prediction involved MG138 ( homologous to elongation factor 4 in E . coli ) , which had been associated with macromolecule formation during the automodel construction . We observed that deletion mutants of the homolog in E . coli ( lepA ) are viable [40] . Based on this information , we removed this gene and its erroneous association as an essential component of the biomass equation from the model . Interestingly , three of the 37 erroneous over-predictions were corrected by adding three membrane components to the biomass equation ( see Table 3 ) . These components were not added during the initial model construction because it was not clear which ( if any ) of this class of metabolites were essential . An additional three erroneous predictions of non-essentiality were corrected by suppressing two reactions . One of these reactions , inosine kinase , was added during GapFill but not linked to an associated gene . Suppression of this reaction corrected two over-predictions but did not invalidate any correct model predictions , suggesting that the reaction activity is unlikely to be present in vivo , at least under the experimental conditions , and perhaps is not an activity encoded by M . genitalium . An additional six over-predictions involved two metal ion ABC transporters . GrowMatch identified each transporter to be essential when the other one was suppressed . We rejected co-regulation of the two transporters as a model restoration mechanism . Instead , we restored consistency for three of the six genes by assigning the cobalt uptake to the complex with a better homology to characterized cobalt transporters ( MG179 , MG180 , MG181 ) . The remaining three genes were removed from the model . An alternative interpretation of the GrowMatch results is that some other ion uptake reaction ( s ) are uniquely associated with these transporters and are thus responsible for the in vivo phenotype . Overall , the application of GrowMatch to the metabolic model led to the generation of a number of testable hypotheses regarding the presence or absence of specific functionalities and emphasized the importance of determining the substrate specificity of the transporters . We also identified reactions that had to be inactivated only for certain knock-outs suggesting their dependence on the genetic background in addition to the specific environmental conditions . For example , MG112 ( ribulose-phosphate 3-epimerase ) had to be suppressed in conjunction with two single gene deletions ( i . e . , conditional suppressions ) to restore consistency with the in vivo data , suggesting possible regulation events . Figure 5 summarizes the complete GrowMatch results . Considering only those changes that could be repaired with global model adjustments conservatively raised the overall percent accuracy of the model ( versioned as iPS189 ) from 79 to 87% . By recalculating the diagnostics of the percentage of correctly- identified essential genes [38] , [39] we found that the model is now 87% ( i . e . , 149 of 171 ) correct in its essentiality predictions ( specificity ) and 89% ( i . e . , 16 of 18 ) correct in its non-essentiality predictions ( sensitivity ) . The iPS189 model predicts that M . genitalium uptakes fructose via a PTS system . The fructose is converted to fructose 1 , 6-bisphosphate ( fdp ) and finally enters the glycolytic pathways to produce lactate via lactate dehydrogenase . Neither fructose nor glycerol uptake was found to be essential as glucose could be efficiently taken up and converted . Specifically , glucose is transformed to glucose 6-phosphate and finally to fdp via phosphofructokinase . As expected , the model also indicates that co-enzyme A ( CoA ) is taken up , since M . genitalium has no coA biosynthesis genes . Additionally , accetal-CoA ( accCoA ) is not formed via pyruvate formate lyase but rather by pyruvate dehydrogenase . Interestingly , we find that should acetate be taken up , it is converted to acetyl phosphate ( actp ) and finally to accoa by phosphotransacetylase . Sources for acyl-CoA ( aCoA ) and CDP-glucose are also required for lipid production . The metabolites riboflavin and nicotinic acid ( niacin ) are taken up for synthesis of the cofactors FAD and NAD , respectively . In addition , both spermidine and putrescene are directly imported as biomass components . Similarly , we also found that D-ribose ( rib-D ) is needed to fuel the truncated pentose phosphate pathway . Examination of fluxes indicated that the uptake of rib-D results in production of 5-phospho-α-d-ribose 1-diphosphate , which enables the conversion of adenine to amp . We also deduced that only adenine and cytidine are precursors to nucleotides and nucleosides ( CTP , dCTP , UTP , dUTP , dTTP ) . Interestingly , the model required the direct uptake of GTP and could not be produced through the uptake of guanine . Model modifications that eliminate this requirement using GrowMatch resulted in a number of incorrect gene essentiality predictions . The need for the direct uptake of GTP is consistent with the fact that in M . mycoides the guanine nucleotide pathways depend on transport of preformed guanine derivatives [34] , and that a number of other Mycoplasmas are not able to grow on medium that only contains guanine as a nucleobase [41] . In addition , all amino acids are imported directly from the environment as either monomers or dipeptides . Unlike many other mycoplasmas , M . genitalium is an arginine nonfermenting species , and not surprisingly arginine deiminase activity was not present in the model . Furthermore , in iPS189 , the only participation of the amino acid arginine is its direct incorporation into biomass . Finally , flux predictions revealed that lactate is the main product of M . genitalium fermentation . A key targeted application of the iPS189 metabolic model is to drive the development of a defined growth medium . As noted above , gene essentiality experiments were performed using a non-defined medium , SP-4 , which contains beef heart infusion , peptone supplemented with yeast extract and fetal bovine serum . The use of an undefined medium can confound the characterization of gene essentiality , as the exact environmental conditions are not fully specified . Furthermore , the lack of a defined growth medium complicates the understanding of nutritional control and regulation of pathologies , evaluation of drug susceptibility , characterization of auxotrophic mutants and performing genetic analysis . Using trial-and-error researchers have already attempted to formulate defined media by systematically deleting components from an undefined or complex media [28] . For example , defined media have been constructed for the growth of Mycoplasma capricolum [42] , Acholeplasma laidlawii [43] , Spiroplasmas [44] , and a semi-defined medium was recently formulated for two Mycoplasma mycoides subspecies [45] . However such approaches do not take into account the balance and availability of chemical species in cellular metabolic pathways to systematically guide medium design . Genome-scale models of metabolism provide maps for tracing missing components needed for biomass formation , redox potential and ATP maintenance [46]–[48] . These models have already been successfully employed to establish minimal reaction sets needed for growth under several uptake environments [49] , elucidate substrate uptake requirements for several microbial organisms such as Helicobacter pylori [50] and Haemophilus influenzae [50] and more recently design complete growth media [51]–[53] . Motivated by these medium-associated shortcomings , we used the iPS189 metabolic model as a roadmap of the available transporters , metabolites and internal interconversions to seek out the minimum number of growth medium components necessary for biomass production . We used as a starting point the components of the C5 medium ( Rodwell , 1983 ) , which is used as a component of SP-4 , with the addition of folate and biotin and the replacement of thiamine by the four individual deoxynucleosides ( see Table S4 ) . By minimizing the total number of additional components that are needed for growth ( see Materials and Methods ) , a number of additional components were identified as required ( see Table 4 ) . The purported reduced bio-availability of the amino acids in their monomeric form might be , in part , the reason for M . genitalium's requirement of yeast extract . The listed dipeptides were found to be needed components consistent with the presence of dipeptide transporters that were found to be essential genes both in vivo and in iPS189 . In addition to components identified based on the metabolic model , we identified that some of the enzymes present in the model required cofactors not included present in the model . In addition , a number of additional carbon source supplements could be investigated . We note that these medium predictions generate a necessary but not a sufficient list of components needed in the medium . For instance , a number of components of biomass such the lipids are not fully specified in the model . Furthermore , additional signaling molecules might be necessary signals for allowing growth of M . genitalium in a defined medium . We anticipate that the iterative process of testing and refining a defined medium followed by updates to the model will successively help pinpoint the precise metabolic capabilities and requirements of M . genitalium .
In this paper , our focus was two-fold: ( a ) to construct the metabolic model for the minimal organism and pathogen M . genitalium , and ( b ) to introduce and bring to bear automated procedures that streamline the construction of metabolic models . The procedure does not require a fully annotated genome and can serve to complement existing annotations [2] by generating testable hypotheses of functionality . We made use of BLASTp to associate functionality to ORFs in the multiple stages , but alternate methods of determining enzymatic function , such as profile-based approaches [54] , [55] could provide additional or alternative assignments . Hypothesizing novel pathways [56] will also likely become increasingly important as metabolic reconstructions for more diverse organisms are carried out . Many genome annotation errors are caused [57] by the use of non-specific reaction compound associations or partially qualified Enzyme Commission numbers [58] . For instance , a comparative study identified an 8% difference in ORFs across three different functional annotations of M . genitalium [59] . Therefore , the direct use of genome-annotation derived reconstructions ( e . g . , KEGG and Pathway Tools generated models ) can lead to inaccurate descriptions of metabolic behavior [60] . Specifically , a recent study has shown that a permissive inclusion of pathways from these reconstructions can lead to models that overpredict the metabolic capabilities of Lactococcus lactis [61] . To safeguard against this issue , we have used manually curated metabolic models as libraries of biotransformations . We note that this procedure allows for the straightforward incorporation of reactions that are charge and elementally balanced , which is not the case with many reactions in KEGG [62] . Earlier efforts have examined the general metabolism of Mycoplasmas [63] or targeted some specific subsections such as purine and pyrimidine metabolism [41] , [64] . Although there is an overlap between the reaction set in our reconstruction and those available in previous studies , developing a M . genitalium specific model with growth requirement as a constraint revealed novel uptake and non-gene associated reactions that previous studies were unable to identify . The identification of which metabolites are produced internally or transported directly from the extracellular environment was complicated by the lack of a defined medium for M . genitalium and its fastidious growth-requirements . Notably , no other metabolic model reconstruction efforts to date were complicated by the lack of both a well-defined biomass composition and a defined growth medium . Here , we allowed the uptake of all metabolites known to be present in the current medium . We also included metabolites either with identified transporters or those necessary for reconnecting blocked metabolites/reactions . Even though missing metabolites and pathways still exist in iPS189 , we were able to achieve a high degree of agreement between the model predictions and in vivo gene essentiality data ( 87% ) . We note that the most recent iteration of the metabolic model for E . coli , an organism which has both a well-defined biomass composition as well as chemically defined growth media , has an overall agreement with in vivo gene essentiality data of 91% under aerobic glucose conditions [33] . Becker and Palsson [38] have recently reported that most in silico models correctly predict less than 45% of essential genes ( called specificity in [39] ) ; for the most-recent E . coli model this diagnostic was 66% . iPS189 has similar performance on both the overall accuracy in growth predictions and specificity ( both 87% ) because of , in part , the much higher percentage of essential genes in M . genitalium but also its careful construction . Additional in vivo gene essentiality studies using a fully defined medium could usher a more accurate elucidation of the true metabolic capabilities of M . genitalium , as well as suggest improvements to the reconstruction . The metabolic model iPS189 is smaller than the 256 genes suggested as a minimal gene set based on comparison of M . genitalium and H . influenzae proteins [22] . In large part , this seeming discrepancy results from the intentional exclusion from the model of many genes that are essential ( e . g . , those encoding DNA and RNA polymerases ) though not directly related to metabolic processes . If such genes were included , the model size would increase by ( at least ) 59 genes to 248 . In addition , 68 genes that are essential in vivo have unknown function , and thus cannot ( yet ) be incorporated into the model . It is possible that they could , in part , carry out some of the non-gene associated reactions that were proposed during the GapFill procedure . Determining their metabolic function through biochemical and molecular biology techniques , as well as determining the substrate specificity for non-characterized transporters , would improve subsequent metabolic models . Looking to the future , we note that it was recently shown that it is feasible to transplant the genome from one mycoplasma species to another [65] , thus opening the door to the transplantation of a perhaps completely synthetic genome . Furthermore , the recent announcement of the de novo synthesis and assembly of the complete M . genitalium genome [66] brings closer to reality the ab initio design of microbes from scratch that are exquisitely tuned for specific biotechnological applications . The constructed metabolic model iPS189 could serve as a core of metabolic functions to add upon so as to bring about the desired biological functionalities and/or production capabilities .
The first step in our reconstruction of the genome-scale model of metabolism of M . genitalium involved analyzing the annotated M . genitalium G-37 genome sequence with the SimPheny automated model generation platform developed by Genomatica ( San Diego , CA ) . This automated procedure leveraged the content contained in Genomatica's manually curated models to expedite the initial reconstruction of M . genitalium metabolism . For the M . genitalium model , reconstructions for E . coli , Haemophilus influenzae , Geobacter sulfurreducens , B . subtilis , and Saccharomyces cerevisiae were used in the comparisons . Subsequently genes-protein-reaction ( GPR ) associations were established based on the evidence provided by homology identity provided by sequence analysis and annotation information . Specifically , forward and reverse protein-protein BLAST ( BLASTp ) comparisons were carried out to identify genes in the library of manually curated models that most closely resemble those in M . genitalium . In the forward BLASTp step , each M . genitalium ORF was compared against each gene in the manually curated models . Then , in the reverse BLASTp step , the best hit from manually curated models was compared against the M . genitalium genome ORFs . The top hits for each of the searches were stored , and a third list was compiled consisting of the cases in which the top hits are identical in both the forward and reverse BLASTp searches . For only these pairs , the GPR association ( s ) from the manually curated model was used as a template to assign those in the M . genitalium model . This two-way BLASTp search procedure improves the likelihood that M . genitalium orthologs , not paralogs , to genes in the manually curated models are identified and assigned by the automodel process . Although the automated genome-model comparisons enabled a fast generation of a draft model , the auto-generated reconstruction was extended by searches for relevant reactions not included in the library of curated models . First , all high-quality annotated “non-metabolic” genes , such as those encoding transcription factors and proteins involved in replication and repair , were excluded from subsequent searches . We then extracted the remaining sequences of all the ORFs not yet included in the M . genitalium model and performed BLASTp analysis to identify additional proteins with biochemically characterized functionalities in other organisms . The target database was the nr ( non-redundant ) database at NCBI ( http://www . ncbi . nlm . nih . gov/blast ) . Next , all target proteins that the BLASTp search reported Expect ( E ) values less than 10−5 were further investigated for biochemical characterization . Next , we carried out a reverse BLASTp analysis of each of these characterized proteins against sequences in the M . genitalium genome . If the initial M . genitalium gene was ranked with the best score and had an acceptable E value ( less than 10−5 ) at the end of this step , that gene and its associated reactions were included in the model as with the library of curated models . During the construction of the initial model , all of the reactions added during the auto-generated reconstruction were also subsequently examined for internal consistency . For example , reactions involving compartments ( which are not present in the bacterium M . genitalium ) were updated to occur in the cytosol , three reactions for ATP synthase were merged , and two macromolecule synthesis reactions were combined and their GPR associations were manually updated . We also added GPR associations for missed characterized members of complexes ( e . g . , a component of an ABC transporter system ) when the automodel associated the other members , as well as corrected those GPR associations for proteins that were misidentified as isozymes instead of complexes . Furthermore , we investigated the cluster of orthologous groups ( COGs ) ontology [35] of all genes in the automodel , as well as examined all ORFs using the NCBI nr database as an added guard against the erroneous transfer of a functionality into the M . genitalium model . Those having classification within non-metabolic classifications , such as transcription ( K ) ; replication , recombination and repair ( L ) ; general function prediction only ( R ) ; function unknown ( S ) ; or none assigned ( N/A ) were inspected further . Although most of the ORFs within ( R ) and ( N/A ) were retained , we removed all those within ( K ) , ( L ) , and ( S ) from the model ( see Table S3 ) . All reactions , including those added from the BLASTp searches and subsequent additions , were checked and corrected whenever necessary for charge balance ( protonation state ) and elemental balance . The biomass equation was generated by accounting for as many as possible of the constituents that form the cellular biomass of M . genitalium . Unfortunately , no complete compilation is available in the open literature for M . genitalium , however some information on the soluble protein composition of it [67] and data on closely related mycoplasmas is available . Therefore , we started with the biomass equation from the Gram positive B . subtilis model [68] and the core biomass equation from the E . coli iAF1260 metabolic model [33] and subsequently modified them accordingly . First , all components related to the cell wall were removed as M . genitalium lacks one . Next , the amount of precursor molecules involved in DNA production were adjusted by altering their ratios in accordance with the lower G+C content in M . genitalium ( 31 . 7% ) as compared to E . coli ( 49 . 8% ) and B . subtilis ( 40 . 9% ) and its smaller genome size . The amino acid relative percentages were likewise adjusted based on utilization patterns and codon biases . Amino acid utilization was incorporated as charged and uncharged tRNA molecules in the biomass equation as reactants and products , respectively . Note that the charged gln-tRNA becomes an uncharged glu-tRNA and the charged fmet-tRNA becomes an uncharged met-tRNA through the biomass equation . Metal ions known to be present in the active sites of catalytic enzymes were also included in the biomass equation . Various membrane and lipid components were subsequently added in accordance with the GrowMatch predictions . Testing the metabolic model using optimization-based approaches requires the definition of a number of sets and parameters . Upper and lower bounds , UBj and LBj , were chosen as not to exclude any physiologically relevant metabolic flux values . The upper bound for all reactions was set to 1 , 000 . The lower bound was set equal to zero for irreversible reactions and to −1 , 000 for reversible reactions . The non-growth associated ATP maintenance limit was set to LBj = 8 . 4 gDW−1 h−1 . The maximum transport rate into the cell was 5 mmol gDW−1 h−1 for any external carbon containing metabolite ( i . e . , LBj = −5 ) . The lower bound for the remaining source exchange reactions was −20 mmol gDW−1 h−1 [33] . Using the principle of stoichiometric analysis along with the application of a pseudo-steady-state hypothesis to the intracellular metabolites [69] , an overall flux balance can be written as follows: ( 1 ) When constructing the model , we also generated the gene-protein-reaction ( GPR ) associations that link the ORFs to the reactions that are catalyzed by their gene products using standard conventions [70] . Once a mathematical representation of the metabolic model was generated , we first determined using GapFind [17] which metabolites could not be produced ( i . e . , cannot carry any net influx ) , given the availability of all substrates supported by the model . We next applied GapFill to modify the existing genome-scale model in order to reconnect these metabolites to the model . The reaction source databases used by GapFill in this work were the KEGG [71] and MetaCyc [72] databases . These reactions were provisionally added to the model , after evaluating charge and elemental balancing . We also performed additional homology searches to try to identify any additional GPR associations . Care was taken when applying GapFill so as not to introduce functionalities known to be absent in M . genitalium ( e . g . , incomplete TCA cycle and disconnected pentose-phosphate pathway ) . The introduction of the GPR associations introduces additional complexity in that an additional layer of detail is needed to fully characterize the network when a single gene is deleted . We made the following definitions to this end: Set contains all ORFs ( genes ) that are included in the metabolic reconstruction . We used a fictitious gene s0001 , as in [33] , to map and track spontaneous reactions that are known to be non-enzymatic ( e . g . , diffusion of CO2 across the cell boundaries ) , but this fictitious gene was not included when enumerating the total genes included in the model , nor could it be knocked out . Note that a gene k may have more than one GPR association when it is involved with more than one reaction . Likewise , a reaction j may be involved in more than one GPR association , as is the case with isozymes or multi-protein complexes . We restricted reaction fluxes , vj , using the binary variable wj as follows: ( 2 ) Equation ( 2 ) ensures that the flux in reaction j can take a non-zero value only if the reaction is active ( i . e . , wj = 1 ) . Parameter describes the impact of the deletion of gene k on reaction j . For instance , if only a single gene k1 is associated with reaction j , then we assign . On the other hand , if two genes k1 and k2 encode isozymes that catalyze reaction j , then we assign . However , if two genes k1 and k2 are required for the formation of a multi-protein complex that catalyzes reaction j , we set both and equal to one . More complex associations were handled in a similar manner . We tested the in silico growth predictions of the M . genitalium metabolism network by examining the flux of the biomass equation . Given the deletion of a single gene k , we solved the following formulation: ( 3 ) subject to ( 1 ) ( 2 ) ( 4 ) Equation ( 4 ) ensures that the flux in reaction j is zero when the gene k that is necessary for its activity is deleted . Note that this equation takes advantage of the fact that in this work we only examined single gene deletions and thus needed not write more complicated GPR-related constraints . This formulation was solved for each gene j in the model using CPLEX version 11 accessed within the GAMS modeling environment . The predictions were compared against in vivo gene essentiality data [19] . Because the experiments were performed in a non-defined rich medium [19] , all metabolites that had identified transporters or were known to be able to cross the membrane into the cell in a non-mediated way were allowed to have exchange reactions to enter the system . Exceptions were the metabolites known not to be present in the growth medium: sugars other than glucose , as well as acetate and lactate . In this study , we used the growth cutoff was that proposed in the recent study [73] which is defined as one–third of the average growth exhibited by all the single gene deletions under consideration . However , we found that the in silico growth predictions for iPS189 were insensitive to this value . We next applied the GrowMatch method to reconcile inconsistencies between in silico and in vivo growth predictions across single gene deletion mutants ( Satish Kumar and Maranas , in preparation ) . To this end , we first classified growth prediction inconsistencies into two categories: ( a ) a mutant is termed a “Grow/NoGrow” ( GNG ) mutant if the in silico model predicts growth whereas there is no observed growth in vivo and ( b ) a mutant is termed a “No Grow/Grow” ( NGG ) mutant if the in silico model predicts no growth in contrast with observed in vivo growth . In GNG mutants , the model overpredicts the metabolic capabilities of the organism . GrowMatch automatically restores consistency in these mutants by suppressing reaction activities to prevent in silico growth ( i . e . , by identifying erroneously added reactions or missing regulation ) . Conversely , in NGG mutants , the model underpredicts the metabolic capabilities of the organism . GrowMatch restores consistency in these mutants by adding functionalities that ensure in silico growth consistent with in vivo data . As when GapFill was applied , the additional reactions were carefully monitored . In all cases , GrowMatch operates so as not to perturb any correct growth predictions . We pose the problem of identifying the minimum number of added components to the growth medium so as to allow for the formation of all biomass constituents as an optimization problem . To this end , we introduce the following additional sets and variables: We minimized the total number of activated substrate exchange reactions that enable the uptake of growth medium components through the use of the following optimization formulation: ( 5 ) subject to ( 1 ) ( 6 ) ( 7 ) ( 8 ) Equation ( 6 ) ensures that the fluxes for all reactions that are not associated with the uptake of a substrate are within the bounds defined earlier . When an exchange reaction j is active ( uj = 1 ) , Equation ( 7 ) permits its flux to assume non-zero negative values , thus allowing the uptake of the corresponding substrate . Conversely , when an exchange reaction j is inactive ( uj = 0 ) , Equation ( 7 ) ensures that the exchange reaction can only remove the corresponding product from the extracellular environment ( i . e . , its flux can only assume positive values ) . Constraint ( 8 ) ensures a minimum amount of biomass formation , with the cutoff set to be the same as that used for the gene essentiality predictions above ( i . e . , one–third of the average maximum biomass flux exhibited by all the single gene deletions ) . This MILP problem was also solved using CPLEX version 11 accessed within the GAMS modeling environment .
|
There is growing interest in elucidating the minimal number of genes needed for life . This challenge is important not just for fundamental but also practical considerations arising from the need to design microorganisms exquisitely tuned for particular applications . The genome of the pathogen Mycoplasma genitalium is believed to be a close approximation to the minimal set of genes required for bacterial growth . In this paper , we constructed a genome-scale metabolic model of M . genitalium that mathematically describes a unified characterization of its biochemical capabilities . The model accounts for 189 of the 482 genes listed in the latest genome annotation . We used computational tools during the process to bridge network gaps in the model and restore consistency with experimental data that determined which gene deletions led to cell death ( i . e . , are essential ) . We achieved 87% correct model predictions for essential genes and 89% for non-essential genes . We subsequently used the metabolic model to determine components that must be part of the growth medium . The approaches and tools described here provide a roadmap for the automated metabolic reconstruction of other organisms . This task is becoming increasingly critical as genome sequencing for new organisms is proceeding at an ever-accelerating pace .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/bioinformatics",
"computational",
"biology/metabolic",
"networks",
"computational",
"biology/systems",
"biology"
] |
2009
|
A Genome-Scale Metabolic Reconstruction of Mycoplasma genitalium, iPS189
|
Chronic disease caused by infections , cancer or autoimmunity can result in profound immune suppression . Immunoregulatory networks are established to prevent tissue damage caused by inflammation . Although these immune checkpoints preserve tissue function , they allow pathogens and tumors to persist , and even expand . Immune checkpoint blockade has recently been successfully employed to treat cancer . This strategy modulates immunoregulatory mechanisms to allow host immune cells to kill or control tumors . However , the utility of this approach for controlling established infections has not been extensively investigated . Here , we examined the potential of modulating glucocorticoid-induced TNF receptor-related protein ( GITR ) on T cells to improve anti-parasitic immunity in blood and spleen tissue from visceral leishmaniasis ( VL ) patients infected with Leishmania donovani . We found little effect on parasite growth or parasite-specific IFNγ production . However , this treatment reversed the improved anti-parasitic immunity achieved by IL-10 signaling blockade . Further investigations using an experimental VL model caused by infection of C57BL/6 mice with L . donovani revealed that this negative effect was prominent in the liver , dependent on parasite burden and associated with an accumulation of Th1 cells expressing high levels of KLRG-1 . Nevertheless , combined anti-IL-10 and anti-GITR mAb treatment could improve anti-parasitic immunity when used with sub-optimal doses of anti-parasitic drug . However , additional studies with VL patient samples indicated that targeting GITR had no overall benefit over IL-10 signaling blockade alone at improving anti-parasitic immune responses , even with drug treatment cover . These findings identify several important factors that influence the effectiveness of immune modulation , including parasite burden , target tissue and the use of anti-parasitic drug . Critically , these results also highlight potential negative effects of combining different immune modulation strategies .
Immune check point inhibitors show great potential for the treatment of various cancers[1 , 2] . These drugs modulate regulatory pathways which suppress anti-tumour immune responses . Major targets for blockade include CTLA4 , PD1 , PDL1 , LAG3 , TIM3 and IL-10 , but a number of molecules are also being tested for immune activation , including CD27 , CD40 , OX40 , CD137 and glucocorticoid-induced TNF receptor-related protein ( GITR ) [3–5] . Many of these molecules play important immunoregulatory roles during infectious diseases , including malaria and leishmaniasis[6 , 7] , thereby providing opportunities to use these treatment strategies in different clinical settings . However , testing immune therapies for treating parasitic diseases lags way behind their development for cancer treatment . Visceral leishmaniasis ( VL ) is a potentially fatal disease caused by Leishmania donovani in the Indian subcontinent and East Africa , and L . infantum ( chagasi ) around the Mediterranean and in Central and South America[8] . These protozoan parasites infect macrophages in visceral organs , with the spleen , bone marrow , lymph nodes and liver being the main tissues affected[9 , 10] . Despite the majority of infections being asymptomatic , individuals that develop VL have a high likelihood of death if they do not receive appropriate chemotherapy[11–13] . Treatment in the Indian subcontinent has improved dramatically with the recent implementation of a single dose administration of liposomal amphotericin B ( Ambisome ) [14–16] . However , problems remain with this regime , including cost and toxicity[8] . In addition , treatment in other endemic regions still relies on more extensive regimes , such as an up to 30 day course of pentavalent antimonials for VL patients in East Africa[8] . Again , toxicity is an important problem and there is a high chance of relapse and/or development of complications such as post kala-azar dermal leishmaniasis ( PKDL ) [8] . Therefore , there is an urgent need to improve current treatments . Interleukin 10 ( IL-10 ) has emerged as a major regulatory cytokine in experimental VL models and in VL patients[7 , 10 , 17 , 18] . IL-10 can act directly on antigen presenting cells to suppress their functions , as well as on T cells to inhibit activity[6 , 19] . Blockade of IL-10 in whole blood samples from VL patients caused significantly greater IFNγ and TNF production by CD4+ T cells in response to stimulation with parasite antigen , thus identifying IL-10 as a potent suppressor of cell-mediated immunity in VL patients[17 , 20] . However , responses to IL-10 signaling blockade were variable and not always positive , suggesting a level of heterogeneity amongst individuals in the importance of IL-10 for suppressing anti-parasitic immune responses[17] . Other immunoregulatory molecules that influence T cell responses during experimental VL have been identified , including PD1[21] , CTLA4[22] , OX40[23] , CD40[24–26] , IL-27[27] and TGFβ[28] . We previously showed that activation of T cells via GITR could enhance anti-parasitic CD4+ T cell responses in experimental VL[29] . GITR activation with an agonist mAb improved anti-parasitic CD4+ T cell responses , but only when administered after establishment of infection when GITR expression increased on all T cell populations . Furthermore , despite only having a minor effect on established hepatic infection when administered alone , anti-GITR mAb acted synergistically with a sub-optimal dose of anti-parasitic drug to improve parasite clearance in the liver and spleen[29] . Here we extend these findings into a clinical setting by testing whether GITR activation improved antigen-specific immune responses in VL patient samples . We also used patient samples and an experimental model of VL to investigate how GITR activation influenced cellular immune responses during infection when administered alone or in combination with IL-10 signaling blockade . Our results uncover a complex outcome following immune activation in both VL patient samples and experimental VL that is influenced by parasite load and with the use of anti-parasitic drug . These findings have implications for the development of immune therapy to treat infectious diseases , as well as broader implications for understanding potential consequences of this approach in any disease setting .
All patients presented with symptoms of VL at the Kala-azar Medical Research Center ( KAMRC ) , Muzaffarpur , Bihar , India . Their diagnosis was confirmed either by the detection of amastigotes in splenic aspirate smears or by rk39 dipstick test . Patients were treated either with Amphotericin B or Ambisome . In total , 58 patients and 10 healthy controls were enrolled in this study . The use of human subjects followed recommendations outlined in the Helsinki declaration . Written informed consent was obtained from all participants and/or their legal guardian when under 18 years of age . Ethical approval ( Dean/2011-12/289 ) was obtained from the ethical review board of Banaras Hindu University ( BHU ) , Varanasi , India . The aggregate clinical data of enrolled subjects are given in Table 1 . All animal procedures were approved by the QIMR Berghofer Medical Research Animal Ethics Committee . This work was conducted under QIMR Berghofer animal ethics approval number A02-634M , in accordance with the “Australian Code of Practice for the Care and Use of Animals for Scientific Purposes” ( Australian National Health and Medical Research Council ) . Heparinised blood was collected from VL patients ( n = 7 ) before and after treatment , as well as from endemic controls ( EC; n = 5 ) . PBMCs were isolated by Ficoll-Hypaque ( GE Healthcare , NJ ) gradient centrifugation and collected directly into RNAlater and stored at -70°C until mRNA isolation and analysis . Total RNA was isolated using RNeasy mini kit and Qiashredder homogenizers ( Qiagen , Venlo , Netherlands ) according to the manufacturer’s protocol . The quality of RNA was assessed by denaturing agarose gel electrophoresis . cDNA synthesis was performed in 20 μL reactions on 0 . 5–1 . 0 μg RNA using High-Capacity cDNA Archive kit ( Applied Biosystems , Foster City , CA ) . Real-time PCR was performed on an ABI Prism 7500 sequence detection system ( Applied Biosystems ) using cDNA-specific FAM–MGB labelled primer/probe for GITR . The relative quantification of products was determined by the number of cycles over 18S mRNA endogenous control required to detect the gene expression of interest . In a separate experiment , PBMC were isolated from VL patients ( n = 7 ) before drug treatment , as well as from EC’s ( n = 5 ) , for cell surface staining of GITR and FACS analysis . Prior to treatment , splenic needle aspirates were collected from VL patients ( n = 15 ) at KAMRC . The left over splenic aspirate after formation of a smear on a glass slide ( to demonstrate the presence of parasite ) , was collected in 0 . 8ml RMPI-1640 medium supplemented with 10% fetal bovine serum ( FBS ) , 10mM L-glutamine , 100U/ml penicillin and 100μg/ml streptomycin ( Invitrogen , Carlsbad , CA ) . Spleen samples were transported to Banaras Hindu University for further analysis at a temperature of 14–18°C . Spleen cell culture was performed as reported previously[17] . In brief , a small fraction of spleen cell suspension ( 150μl ) was inoculated in a 96 well plate containing blood agar for base line quantification of parasites by limiting dilution assay . The remaining sample was plated in duplicate ( 300μl ) in 96 well tissue culture plates ( Thermo Fisher , Waltham , MA ) and incubated for 3 days at 37°C and 5% CO2 in presence of either anti-GITR mAb ( clone TRX518; 20μg/ml , Tolerx , Cambridge , MA ) or isotype control ( human IgG1;Sigma , St Louis , MO ) . TRX518 is a non-depleting , humanized IgG1 anti-human GITR mAb with a heavy chain asparagine 297 substitution to alanine to eliminate N-linked glycolsylation and abrogate Fc region functionality[30] , produced under good manufacturing process ( GMP ) conditions . After 3 days , culture supernatants were removed and replaced by promastigote growth medium ( M199 ) , supplemented with 20% FBS , 10mM L-glutamine , 100U/ml penicillin and 100μg/ml streptomycin , 40mM HEPES , 0 . 1mM adenine ( in 50 mM HEPES ) , 5mg/ml hemin ( in 50% triethanolamine ) . The splenic aspirates were transferred to blood agar plate for limiting dilution assay using 3-fold serial dilutions and kept at 25°C for 10–12 days to allow parasite growth . Whole blood assays were performed as previously described[20 , 31] . In brief , heparinised blood was collected ( n = 19 ) from active VL patients . To remove background plasma cytokines , the plasma was replaced with an equal volume of FBS . Whole blood cells were cultured in absence or in presence of soluble leishmania antigen ( SLA ) . To detect whether GITR activation could enhance antigen specific immune response; either alone or in combination with IL-10 signaling blockade , specific mAbs and their isotype controls ( IgG2b and IgG1 for anti-IL-10 and anti-GITR mAbs , respectively ) were added ( each 20μg/ml ) along with SLA . In all experiments , a non-stimulated group was included , and although minimal cytokine production was detected in these samples , the levels detected were subtracted from corresponding antigen-stimulated samples . Anti-IL-10 mAb ( clone25209 ) and mouse IgG2b isotype ( clone 20116 ) were purchased from R&D Systems ( Abingdon , United Kingdom ) . Whole blood cultures were kept at 37°C and 5% CO2 for 24 hours . Supernatants were collected and IFN-γ levels were measured using an ELISA kit ( Biolegend , San Diego , CA ) , as per manufacturer instructions . To test the effect of drug treatment on responses to antibodies in whole blood assays , a similar experiment was conducted on a separate set of whole blood samples ( n = 10 ) , where heparinized blood was collected from active VL patients before the start of treatment and one day after single-dose Ambisome treatment[15] . Female C57BL/6J were purchased from the Australian Resource Centre ( Canning Vale , Western Australia ) and Walter and Eliza Hall Institute for Medical Research ( Victoria , Australia ) , and maintained under conventional conditions at the QIMR Berghofer Medical Research Institute . B6 . RAG2-/- mice were bred at the Queensland Institute of Medical Research . Mice used were sex- and age-matched ( 6–10 weeks ) and groups of 5 to 6 mice were used in experiments . L . donovani ( LV9 strain ) were maintained by passage in B6 . RAG2-/- mice and amastigotes were isolated from the spleens of chronically infected animals . Mice were infected by injecting 2 x 107 amastigotes i . v . via the lateral tail vein , killed at the times indicated in the text by CO2 asphyxiation and bled via cardiac puncture . In experiments examining low dose infections , mice were infected with 5 x 106 amastigotes i . v . . Spleens and perfused livers were removed at times indicated and parasite burdens were determined from Diff-Quik-stained impression smears ( Lab Aids , Narrabeen , Australia ) and expressed as Leishman-Donovan units ( LDU ) ( the number of amastigotes per 1 , 000 host nuclei multiplied by the organ weight in grams ) [32] . Spleen parasite burden was also determined by limiting-dilution analysis[17] . Hepatic mononuclear cells and splenocytes were isolated as previously described[33] . All hybridomas ( DTA-1 ( anti-GITR[34] ) and 1B1 . 3a ( anti-IL-10R[35] ) ) were grown in 5% FCS , RPMI 1640 containing 10 mmol/l L-glutamine , 200 U/ml penicillin , and 200 mg/ml streptomycin . Purified mAbs were prepared from culture supernatants by protein G column purification ( Amersham Biosciences , Uppsala , Sweden ) , followed by endotoxin removal ( Mustang membranes , Pall , East Hills , NY ) . For in vivo stimulation of GITR , mice were injected intraperitoneal ( i . p . ) with 0 . 5 mg DTA-1 mAb in 200μl 0 . 9% sodium chloride ( Baxter ) per mouse on day 14 post-infection . Anti-IL-10R blocking mAb ( 1B1 . 3a ) was administered in 0 . 5mg doses , via i . p . injection in 200μl 0 . 9% sodium chloride ( Baxter , Old Toongabbie , NSW , Australia ) per mouse on days 14 , 19 and 24 post-infection . Control mice were administered the same quantities of control rat IgG ( BioXcell , West Lebanon , NH ) at the same time points as the stimulatory or blocking mAbs . The pentavalent antimonial , sodium stibogluconate ( Sbv; Albert David Ltd , Kolkata , India] , was administered at 500mg/kg/day doses in 200μl 0 . 9% sodium chloride ( Baxter ) per mouse on days 14 and 21 post-infection and administered i . p . . In multiple-dosing experiments , mice were treated on days 14 , 16 , 18 , 20 , 22 , 24 and 26 post-infection with Sbv at 50mg/kg/day doses in 200μl 0 . 9% sodium chloride ( Baxter ) per mouse and administered i . p . , based on drug dosing used by others[36] . Mouse studies: Brilliant Violet 421-conjugated anti–TCRβ chain ( H57-597 ) , PerCP-Cy5 . 5–conjugated anti–TCRβ chain ( H57-597 ) , Brilliant Violet 605–conjugated anti-NK1 . 1 ( PK136 ) , Allophycocyanin-Cy7–conjugated anti-NK1 . 1 ( PK136 ) , Brilliant Violet 605–conjugated anti-CD4 ( GK1 . 5 ) , Allophycocyanin-Cy7–conjugated anti-CD4 ( GK1 . 5 ) , Alexa Fluor 700–conjugated anti-CD8α ( 53–6 . 7 ) , FITC–conjugated anti-CD11a ( M1714 ) , PE-Cy7–conjugated anti-CD49d ( R1-2 ) , Biotinylated anti-CD49d ( R1-2 ) , PerCP-Cy5 . 5–conjugated anti–KLRG-1 ( 2F1 ) , PE-Cy7–conjugated anti-CD25 ( PC61 ) , Allophycocyanin-conjugated anti-Foxp3 ( MF14 ) , Allophycocyanin-conjugated anti-Tbet ( eBio4B10 ) , PE-Cy7-conjugated anti-Tbet ( eBio4B10 ) , PE-conjugated anti-IFNγ ( XMG1 . 2 ) , Allophycocyanin-conjugated anti-IFNγ ( XMG1 . 2 ) , Brilliant Violet 421-conjugated anti-IFNγ ( XMG1 . 2 ) , PE-conjugated anti-IL-10 ( JES5-16E3 ) , Allophycocyanin-conjugated anti-IL-10 ( JES5-16E3 ) , PE-conjugated anti-TNFα ( MP6-XT22 ) , PE-conjugated anti-CD279 ( PD-1 ) ( J43 ) and Allophycocyanin-conjugated anti-CD223 ( LAG-3 ) ( C9B7W ) were purchased from BioLegend or BD Biosciences ( Franklin Lakes , NJ ) . Biotinylated antibodies were detected with streptavidin conjugated PE-Cy7 . Dead cells were excluded from the analysis using LIVE/DEAD Fixable Aqua Stain or LIVE/DEAD Fixable Near I-R Stain ( Invitrogen-Molecular Probes , Carlsbad , CA ) , according to the manufacturer’s instructions . The staining of cell surface antigens and intracellular cytokine staining were carried out as described previously[37] . FACS was performed on aLSRFortessa ( BD Biosciences ) , and data were analysed using FlowJo software ( TreeStar , Ashland , OR ) . Serum and/or tissue culture supernatants were assessed for the presence of soluble cytokines using flexset bead array kits ( BD Biosciences ) according to the manufacturers’ instructions . Human studies: Allophycocyanin-conjugated anti-CD3ε ( HIT3a ) and FITC–conjugated anti-CD4 ( A161A1 ) were purchased from Biolegend . PE-conjugated GITR ( 110416 ) was purchased from R&D Systems . FACS was performed on a BD FACS Calibur and data were analysed using FlowJo software . RPMI 1640 was supplemented with 20% heat in activated FCS , penicillin ( 2Mm ) , streptomycin ( 1000 U/ml ) , 2-mercaptoethanol ( 0 . 05 mM ) , sodium pyruvate ( 2 mM ) and 1M HEPES ( 200 μm Final ) . Mouse spleens were processed to obtain a single-cell suspension , diluted at 2 x 106 cells/ml in RPMI 1640 complete medium and aliquoted in 96 well plates at 1 x 105 cells/well . Cells were stimulated with fixed L . donovani amastigotes at 2 x 106 parasites/well for 24 hours , as previously described[38] . In all experiments , 100μL/well supernatant was removed 4 hours prior to cell harvest for measurement of cytokine levels and replaced with 100μL/well fresh RPMI 1640 complete medium containing Brefeldin A ( Sigma , St Louis , MO ) , at a final concentration of 10 μg/ml . Intracellular cytokine staining was performed as described above . Comparisons between two groups were performed using non-parametric Mann-Whitney tests in mouse studies and Wilcoxon matched-pairs signed rank test or paired t-test , as appropriate , in human studies . Comparisons between multiple groups were made using a Kruskal-Wallis test and corrected using Dunn’s multiple comparisons test . Differences of p< 0 . 05 were considered significant . Graphs depict mean values ± SEM . All statistical analyses were performed using GraphPad Prism 6 software .
We first examined the therapeutic potential of GITR in VL patients by measuring GITR mRNA accumulation in PBMC’s from VL patients before and after drug treatment , and comparing with GITR mRNA accumulation in PBMC’s from healthy , endemic controls ( Fig 1A ) . GITR mRNA levels increased in cells from VL patients , compared with endemic control samples , and then declined after the completion of drug treatment . FACS analysis on blood samples revealed that there was a significantly higher frequency of GITR-positive CD4+ T cells in VL patients , compared to endemic controls ( Fig 1B ) . Therefore , similar to our findings in experimental VL[29] , there was a greater frequency of GITR-positive CD4+ T cells during active VL , thus making this molecule a potential therapeutic target . We next tested whether GITR can be targeted to improve parasite killing by adding anti-GITR mAb to cell cultures comprising splenic aspirates taken from VL patients as part of routine diagnostic procedures . We used a non-depleting , humanized IgG1 , anti-human GITR mAb with reported ability to enhance effector T cell responses and block regulatory T cell-mediated suppression[30] . We found no change in the number of viable parasites in culture following anti-GITR mAb addition , compared with control samples ( Fig 2A ) , suggesting that targeting GITR had minimal impact on anti-parasitic immune responses under these conditions . To examine whether targeting GITR had any effects on cell-mediated immune responses in VL patients , we next employed a whole blood assay in which cells were cultured overnight in the presence of parasite antigen and IFNγ levels were measured the following day , as previously reported[17 , 20] . Initial experiments showed no improvement in antigen-driven IFNγ production following the addition of anti-GITR mAb ( Fig 2B ) . Past work demonstrated that IL-10 was a major suppressor of anti-parasitic immunity in this assay[17] , and hence , we speculated IL-10 might prevent any positive effects of targeting anti-parasitic IFNγ production . Therefore , we tested whether combining anti-GITR mAb treatment with IL-10 signaling blockade could improve this response . As previously reported[17 , 20] , IL-10 signaling blockade resulted in increased IFNγ production in this assay following stimulation with parasite antigen ( Fig 2C ) . Surprisingly , the addition of an anti-GITR mAb reversed this effect ( Fig 2C ) . Thus , targeting GITR on antigen-activated PBMC’s from VL patients had no effect alone , and a detrimental effect on their ability to respond to IL-10 signaling blockade . To try and better understand the above results , we returned to the experimental VL model caused by intravenous infection of C57BL/6 mice with L . donovani amastigotes . Infections were allowed to establish for 14 days before treating mice with an agonistic anti-GITR mAb , an anti-IL-10R mAb or a combination of both , and then measuring parasite burdens 14 days later ( Fig 3A ) . Although there was a consistent reduction in parasite burdens following anti-GITR mAb administration , this did not reach statistical significance , relative to control treated mice , in line with our previous findings[29] . As expected[39 , 40] , IL-10 signaling blockade resulted in significantly reduced parasite burdens in both liver and spleen , compared with controls , but combining this treatment with GITR activation had little impact on parasite burdens . Serendipitously , in an experiment where mice were mistakenly infected with a 4-fold reduction in parasite number , we observed that GITR activation limited the anti-parasitic activity caused by IL-10 signaling blockade in the liver , as we had observed in our human studies . Therefore , we established infections using a lower parasite inoculum and examined the impact of immune modulation ( Fig 3B ) . Again , despite a consistent reduction in parasite burdens following anti-GITR mAb administration , this did not reach statistical significance , relative to control treated mice . IL-10 signaling blockade resulted in significantly reduced parasite burdens in both liver and spleen , compared with controls , but combining this treatment with GITR activation limited these effects in the liver . Therefore , the liver response in this lower burden experimental VL model appeared to better reflect the immune environment in human VL patients , so we therefore included the lower parasite burden model in our studies and focused on responses in this tissue site . We next examined various immune parameters that might explain the negative impact of GITR activation . Th1 ( Tbet+ , IFNγ-producing CD4+ T cells ) cells[41] and IL-10-producing CD4+ T cells[18] cells have been previously shown to influence disease outcome in experimental and clinical VL , but we found only relatively minor differences in these T cell populations , as well as IL-10-producing Th1 ( Tr1 ) cells , in the liver between mice treated with different combinations of agonistic anti-GITR and anti-IL-10R mAbs , regardless of whether mice were infected with a low ( 5 x 106; Fig 4B ) or high ( 2 x 107; Fig 4C ) parasite inoculum . Similar observations were also made in spleen tissue ( S1 Fig ) . Interestingly , in the spleen we observed a greater number of Th1 cells in control mice with low dose infection , compared to the same group in with high dose infection , but a similar number of Tr1 cells in both groups ( S1 Fig ) . Thus , the ratio of Th1 to Tr1 cells was much higher in the control mice receiving the lower dose of infection . Differences in both frequency and number of KLRG-1-expressing Th1 cells , possibly representing functionally exhausted cells[42 , 43] , were apparent ( Fig 4B and 4C ) , with significant increases noted in the livers of mice treated with combined anti-GITR and anti-IL-10R mAbs , compared with control mice . However , no significant difference in the frequency or number of KLRG-1-expressing Th1 cells were found in the livers of low dose-infected mice treated with anti-IL-10R mAb alone and mice treated with combined anti-GITR and anti-IL-10R mAbs ( Fig 4B ) , suggesting changes in KLRG-1 expression on Th1 cells may not explain the antagonistic effects of GITR activation on the anti-parasitic effects of IL-10 signaling blockade . Instead , we found that the KLRG-1+ CD4+ T cells produced significantly more IFNγ on a per cell basis in all antibody-treated groups ( Fig 4D ) . Remarkably , mice infected with lower parasite numbers ( Fig 4B ) had significant increases in the frequency and number of KLRG-1-expressing Th1 cells , compared with corresponding treatment groups of mice infected with higher parasite numbers ( Fig 4C ) . For example , control mice infected with the low-dose inoculum had 28 . 2 ± 1 . 5% and 6 . 0 x 106 ± 6 . 1 x 105 KLRG-1+ hepatic Th1 cells , compared with 18 . 0 ± 1 . 5% and 3 . 7 x 106 ± 3 . 6 x 105 KLRG-1+ hepatic Th1 cells in the same group infected with the higher dose ( P < 0 . 001 for both frequency and number; n = 16–19 mice/treatment group ) . Similarly , combined anti-GITR and anti-IL-10R mAb-treated mice infected with the low-dose inoculum had 41 . 1 ± 1 . 9% and 9 . 8 x 106 ± 1 . 0 x 106 KLRG-1+ hepatic Th1 cells , compared with 30 . 4 ± 2 . 3% and 7 . 0 x 106 ± 4 . 9 x 105 KLRG-1+ hepatic Th1 cells in the same group infected with the higher dose ( P < 0 . 01 for frequency and P < 0 . 001 for number; n = 16–19 mice/treatment group ) . Again , similar observations were made in the spleen ( S1 Fig ) . Hence , a lower infection dose resulted in more KLRG-1-expressing Th1 cells , but although this was exacerbated following all antibody treatment , it did not appear to explain why GITR activation reversed the anti-parasitic effects of IL-10 signaling blockade . Additional analysis of the frequency or number of hepatic effector ( Fig 5B ) , central memory ( TCM; Fig 5C ) and effector memory ( TEM; Fig 5D ) CD4+ T cells , based on previously described markers[44 , 45] , revealed few differences in any of the treated groups . Furthermore , the expression of IFNγ and PD-1 on these CD4+ T cell subsets was not different between infected groups ( Fig 5B–5D ) . Thus , it is unlikely that the negative impact of anti-GITR mAb treatment on IL-10 signaling blockade can be explained by either the promotion of T cell exhaustion or impairment of T cell activation or memory T cell differentiation . Given that any immune therapy for a parasitic disease is likely to be combined with anti-parasitic drug treatment , we next examined the impact of combined anti-GITR and anti-IL-10R mAb treatment with a sub-optimal dose of the anti-parasitic drug sodium stibogluconate ( Sbv ) . A low-dose infection was allowed to establish for 14 days before beginning drug treatment and/or antibody administration ( Fig 6 ) . Although drug reduced hepatic parasite burdens in all antibody treated groups , a significant improvement compared to mice treated with sub-optimal drug dose alone was only achieved when either anti-GITR mAb alone or combined anti-GITR and anti-IL-10R mAb were used ( Fig 6A ) . We next assessed anti-parasitic cellular responses ex vivo . Spleen cells were used for these experiments because in our experience , cultured hepatic CD4+ T cells grow poorly , possibly reflecting a more advanced differentiated state that makes them more susceptible to death . When cellular responses from mice treated with antibodies combined with sub-optimal drug were compared to cell samples from mice treated with sub-optimal drug dose alone , significantly increased IFNγ and TNF production was only observed when combined with anti-GITR and anti-IL-10R mAb ( Fig 6B and 6C ) . However , drug treatment with both combined anti-GITR and anti-IL-10R mAb , as well as anti-IL-10R mAb alone , resulted in significantly increased IFNγ and TNF production , compared to groups treated with the antibodies alone ( Fig 6B and 6C ) . There was little difference in IL-10 levels , except in groups that received IL-10 signaling blockade alone , where IL-10 production was higher by cells from drug-treated mice . Together , these data indicate that combined antibody treatment together with anti-parasitic drug was effective at controlling parasite growth in the liver and promoting anti-parasitic immune responses . Finally , we examined whether drug treatment affected the ability of combined mAb’s to influence cell-mediated immune responses in VL patients by again using IFNγ production in response to parasite antigen in a whole blood assay as a readout . Blood samples were taken from VL patients upon admission to clinic and 24 hours after drug treatment with a single-dose of liposomal amphotericin B[15] . Again , prior to treatment , anti-IL-10 mAb improved parasite-specific IFNγ production ( Fig 7A ) , but when combined with anti-GITR mAb suppressed the enhanced IFNγ production following IL-10 blockade ( Fig 7A ) . After drug treatment , IL-10 signaling blockade again improved parasite-specific IFNγ production ( Fig 7B ) . However , at this time , despite no improvement in IFNγ production , there was no significant decrease in IFNγ levels by the addition of anti-GITR mAb to cells receiving IL-10 signaling blockade . Together , these results indicate that drug treatment reduced the negative impact of anti-GITR mAb on IL-10 signaling blockade , but anti-GITR mAb treatment is unlikely to offer significant therapeutic benefit to VL patients over IL-10 signaling blockade alone .
Our previous work identified GITR as a potential therapeutic target for treating VL[29] . Here , we show that not only did targeting GITR fail to improve anti-parasitic immune responses in VL patient samples , but had a negative impact on IL-10 signaling blockade , which alone can significantly improve parasite-specific T cell responses . Studies in an experimental model demonstrated that this negative effect was prominent in the liver , dependent on parasite burden and was associated with more Th1 cells expressing high levels of KLRG-1 . Interestingly , this latter finding did not appear to indicate the promotion of T cell exhaustion because these KLRG-1+ Th1 cells produced high levels of IFNγ and displayed no increase in PD-1 expression . Instead , we speculate that the negative impact of anti-GITR mAb treatment on IL-10 signaling blockade was caused by either temporal changes in cytokine production by CD4+ T cells or the promotion of an unidentified immunoregulatory pathway . Nevertheless , combined IL-10 signaling blockade and anti-GITR mAb treatment could be improved when used with anti-parasitic drug . However , additional studies with VL patient samples indicated that targeting GITR had no advantage over IL-10 signaling blockade alone at improving anti-parasitic immune responses , even with drug treatment cover . Hence , our findings identify several important factors that influence the effectiveness of immune modulation , including parasite burden , target tissue and the use of anti-parasitic drug . Importantly , our results also highlight potential negative effects of combining different immune modulation strategies . Results obtained from VL patient samples showed that despite increased expression of GITR by CD4+ T cells , anti-GITR mAb treatment failed to improve anti-parasitic immunity in a whole blood assay , and critically , reversed improvements in antigen-specific IFNγ production caused by IL-10 signaling blockade . Initial attempts to investigate this phenomenon in an experimental mouse infection failed to recapitulate these effects . However , when mice were infected with a lower parasite number , we found a similar negative effect of GITR activation on liver anti-parasitic immunity . It was surprising that a lower parasite inoculum resulted in more activated Th1 cells , compared with the frequency and numbers found in mice infected with a higher number of parasites . However , earlier studies by Parish and colleagues[46–48] described distinct patterns of cellular and humoral immune responses depending on antigen dose , with high dose of pathogen suppressing cell-mediated immunity . In fact , IL-10 levels are elevated in VL patients , and this cytokine suppressed parasite-specific CD4+ T cell IFNγ production , but after parasite burdens were reduced by drug treatment , IL-10 levels fell and parasite-specific CD4+ T cell IFNγ production increased[17 , 20 , 31] . Similar pathogen burden-dependent immunosuppressive mechanisms have been reported in other infections , including malaria[49 , 50] and hepatitis B[51] . Data presented in this study also showed a greater ratio of Th1 to Tr1 cells in control mice with low-dose infection , compared to the same group with a high dose infection , and this may contribute to a more activated status that results in a greater proportion of KLRG1+ cells . These findings raise questions about which experimental settings provide the best reflection of clinical VL . The mouse model we have employed exhibits an acute , resolving infection in the liver , but not sterilizing immunity . In contrast , the spleen becomes chronically infected and maintains a relatively high parasite burden , associated with tissue pathology[52] . Hence , these two sites of infection in the mouse model appear to represent different ends of the human clinical VL spectrum . In recent years , VL patients generally present at clinics earlier during disease progression ( Shyam Sundar , BHU , personal observations ) , possibly due to increasing awareness of disease and improved access to treatment via various programs aimed at VL elimination in the Indian subcontinent . Thus , it is possible that a low parasite burden setting in the liver of mice best reflects anti-parasitic immune responses in early clinical VL . However , this may not always be the case , and is likely to differ in situations where disease progression is either more rapid or advanced , such as might occur in different geographical locations where disease control programs are more limited[53] . Combining immune checkpoint inhibitors such as anti-CTLA4 and anti-PD1 mAbs can have dramatic effects on anti-tumor immune responses , and result in significant improvements in clinical outcomes , compared with single antibody treatment[54] . However , combining immune therapies does not always work . In a recent study , administration of agonistic anti-OX40 mAb to mice infected with Plasmodium yoelii resulted in significantly enhanced anti-parasitic T cell responses and improved control of parasite growth . However , when combined with PD-1 blockade , a strategy previously shown to improve anti-parasitic immunity during malaria[55] , the beneficial effects of OX40 activation were reversed[56] . In this study , the combined therapy caused excessive T cell IFNγ production . Although we did not find enhanced Th1 responses with combined antibody treatment in our VL model , we did observe increased Th1 cell exhaustion , which was especially prominent in the liver when mice were infected with a lower parasite inoculum . Thus , it appeared that combined GITR activation and IL-10 blockade promoted excessive Th1 cell expansion , and subsequent exhaustion , without associated improvement in anti-parasitic immunity . The molecular mechanisms mediating this effect are unknown , but warrant further investigation if we wish to better understand and predict detrimental outcomes from immune modulation . Importantly , when a sub-optimal drug treatment regime was incorporated into the combined anti-IL-10R and anti-GITR mAb treatment , significant improvements in anti-parasitic immunity were observed . Sub-optimal drug treatment worked most effectively with anti-GITR mAb alone or the combined anti-IL-10R and anti-GITR mAb treatment , relative to the control antibody treatment group . However , significantly improved IFNγ and TNF production in response to parasite antigen were as good in mice that received anti-IL-10R mAb alone , compared with mice receiving combined antibody treatment . We also examined the effect of drug administration on combined antibody treatment with VL patient samples by comparing responses to parasite antigen in whole blood assays 24 before and 24 hours after drug treatment with liposomal amphotericin B[14] . Although responses following combined anti-GITR and anti-IL-10 mAb administration after drug treatment were not significantly reduced , as was the case before drug treatment , there was no improvement in response , compared with IL-10 signaling blockade alone . One important difference between our studies in mice and humans was that pentavalent antimonials were used in mice , while liposomal amphotericin B was used in VL patients , and this may explain some differences we observed . We were unable to use amphotericin B in our studies because the efficacy of this drug against the LV9 strain ( MHOM/ET/67/HU3 ) of L . donovani we use was poor , possibly reflecting the East African origins of this strain and the relatively poor efficacy of amphotericin B in VL patients in East Africa [57] . Nevertheless , based on the available data , GITR activation provided no additional advantage to anti-IL-10 mAb treatment alone at improving anti-parasitic immune responses within 24 hours of drug treatment . The effectiveness of the recently introduced , single-dose drug treatment protocol in VL patients in our study site means that patients are released after 24 hours after drug treatment , making sample collections at later times difficult , and changes in responsiveness to combined antibody treatment later during recovery harder to assess . However , given the substantial welfare advantages of the minimized hospital admission time with the new drug treatment protocol , any immune modulation aimed at improving anti-parasitic immunity would have to be administered at the time of drug treatment , thus reinforcing the superiority of anti-IL-10 signaling blockade alone . Efforts to eliminate VL are likely to require coordinated efforts involving improved patient treatment , sustained vector control programs and strategies to reduce parasite loads in individuals to below levels that allow parasite transmission . A major risk factor for developing VL is that a household member has previously had this disease[58] , thereby indicating that VL patients can act as parasite reservoirs for transmission . If transmission occurs after drug treatment , then the time of treatment might provide a window of opportunity to enhance immunity to reduce persisting parasite numbers . As described earlier , there have been many potential immune check point molecules identified that could be targeted and combined with drug treatment for beneficial outcomes . Furthermore , there is ample evidence for the establishment of regulatory mechanisms that suppress anti-parasitic immunity in VL patients[7] . Therefore , the development of immunomodulatory strategies , whether as single or combined treatments , has the potential to improve VL elimination programs . The challenge is to identify the right targets that achieve the above goals without causing harm . In summary , our study shows that immune modulation can be an effective adjunct to parasite drug treatment . However , the effectiveness of such approaches will vary depending on clinical parameters such as parasite burden , and in the case of combined therapeutic approaches , whether the host response to individual components is complimentary or antagonistic . Therefore , targeting immune checkpoints has the potential to improve anti-parasitic immunity , but efforts to improve these responses by combining reagents to target different immune molecules should be evaluated carefully , taking into consideration the inclusion of conventional drug treatment approaches .
|
Despite a decline in the numbers of visceral leishmaniasis ( VL ) cases and the availability of drugs to treat patients , there is still a need for more efficient treatment options . This is especially important as many VL endemic areas enter phases of disease elimination where parasite burdens need to be reduced as far as possible to minimize the risk of future parasite transmission . Immunotherapy involving the activation of anti-parasitic immunity in infected individuals is a promising approach to help achieve these goals . We tested whether immune modulation could improve anti-parasitic immunity using VL patient samples and an experimental mouse model of infection . Our results provide guidance on the utility of the experimental model , as well as highlighting potential problems associated with immune modulation to improve disease outcomes . These findings have implications for understanding how immunotherapy should be tested and implemented .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"t",
"helper",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"spleen",
"immunology",
"tropical",
"diseases",
"parasitic",
"diseases",
"preventive",
"medicine",
"pharmaceutics",
"neglected",
"tropical",
"diseases",
"vaccination",
"and",
"immunization",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"white",
"blood",
"cells",
"zoonoses",
"animal",
"cells",
"t",
"cells",
"drug",
"therapy",
"protozoan",
"infections",
"immune",
"response",
"cell",
"biology",
"physiology",
"leishmaniasis",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"antiparasitic",
"therapy"
] |
2016
|
Combined Immune Therapy for the Treatment of Visceral Leishmaniasis
|
In Bangladesh , pharmacy-purchased oral rehydration solution ( ORS ) is often used to treat diarrhea , including cholera . Over-the-counter sales have been used for epidemiologic surveillance in the past , but rarely , if ever , in low-income countries . With few early indicators for cholera outbreaks in endemic areas , diarrhea-related product sales may serve as a useful surveillance tool . We tracked daily ORS sales at 50 pharmacies and drug-sellers in an urban Bangladesh community of 129 , 000 for 6-months while simultaneously conducting surveillance for diarrhea hospitalizations among residents . We developed a mobile phone based system to track the sales of ORS and deployed it in parallel with a paper-based system . Our objectives were to determine if the mobile phone system was practical and acceptable to pharmacists and drug sellers , whether data were reported accurately compared to a paper-based system , and whether ORS sales were associated with future incidence of cholera hospitalizations within the community . We recorded 47 , 215 customers purchasing ORS , and 315 hospitalized diarrhea cases , 22% of which had culture-confirmed cholera . ORS sales and diarrhea incidence were independently associated with the mean daily temperature; therefore both unadjusted and adjusted models were explored . Through unadjusted cross-correlation statistics and generalized linear models , we found increases in ORS sales were significantly associated with increases in hospitalized diarrhea cases up to 9-days later and hospitalized cholera cases up to one day later . After adjusting for mean daily temperature , ORS was significantly associated with hospitalized diarrhea two days later and hospitalized cholera one day later . Pharmacy sales data may serve as a feasible and useful surveillance tool . Given the relatively short lagged correlation between ORS sales and diarrhea , rapid and accurate sales data are key . More work is needed in creating actionable algorithms that make use of this data and in understanding the generalizability of our findings to other settings .
Despite well over a hundred years of research on the disease , cholera remains a threat to public health in many parts of the world , causing more than 2–3 million cases each year and over 100 , 000 deaths [1] . While many of the global cholera deaths come from large unexpected outbreaks , thousands die in highly endemic settings , like Bangladesh , each year [2] . Any improvements to cholera surveillance systems that can provide an early warning of an increase in transmission may provide an opportunity to intervene with vaccine or water and sanitation improvements , averting cases and saving lives . Most cholera surveillance systems in endemic countries are passive and clinic-based; ultimately capturing only the severe cases that are able to access healthcare services . Active case-identification systems at the community level are rare due to the costs involved; however , opportunities may exist in some areas to tailor novel custom surveillance systems to the ways in which people seek care for diarrhea . Pharmacies often serve as the first point of contact with the healthcare sector when individuals become ill . In Bangladesh , people commonly use oral rehydration solution ( ORS ) purchased at pharmacies to treat diarrhea , including Vibrio cholerae , in part due to years of mass educational campaigns and social marketing [3–6] . Trends in the sales of different diarrhea-related products may serve as a useful tool for detecting increases in cholera within a community , perhaps before clinic-based surveillance systems detect the signal of an outbreak . Pharmacy sales have been used in surveillance within high-income countries to gain insight into disease trends including diarrhea [7–10] and influenza [11] . To our knowledge , this tool has not been employed in low resource settings , where disease surveillance is often much poorer and pharmacy sales are usually not routinely tracked electronically . Surveillance systems to track pharmacy sales may provide even greater utility in countries like Bangladesh than in high-income countries , given the lower baseline quality of disease surveillance systems within the country . We implemented real time mobile phone- and paper-based surveillance systems for ORS sales at pharmacies in an endemic community in Dhaka , Bangladesh and compared it with active , case-based diarrhea surveillance at local hospitals . Our objectives were to determine if the mobile phone system was practical and acceptable to pharmacists and drug sellers , whether data were reported accurately compared to a traditional paper-based system , and whether ORS sales were associated with future incidence of cholera hospitalizations within the community .
Arichpur , is a 1 . 2 km2 urban community located 15 km north of Dhaka , Bangladesh in the Tongi sub-district . Arichpur has approximately 129 , 000 residents living in 29 , 000 households where many nuclear families share one room and up to 10–15 families may share a stove , toilet , and water source . Although the cholera burden in Arichpur has not been estimated , patients from this area frequently visit the icddr , b ( International Centre for Diarrheal Disease Research ) cholera treatment center in Dhaka and have suffered from outbreaks of other water-borne diseases including hepatitis E [12] . All pharmacies and drug sellers in Arichpur were enumerated and mapped . We then enrolled 50 pharmacies , out of 124 due to cost and human resource constraints , with equal probability , replacing those who declined to participate with the next on the list ( Fig 1 ) . We used uniform random selection , as opposed to a more advanced surveillance site selection [13 , 14] technique , as no prior sales data , nor high-resolution population data , were available . At each pharmacy we identified the employees primarily responsible for diarrhea-related sales and trained each on the use of paper forms and phone reporting systems . All pharmacies were enrolled between 25 March and 30 April 30 , 2013 , and continued reporting to the surveillance system through October 2013 . Pharmacy employees reported ORS sales through two parallel systems; a paper based system , and a phone-based interactive voice response system [15] . With the paper-based system , employees recorded the number of customers each day and the number of ORS packets purchased by each . The paper forms were collected by study staff at the end of each month and were entered into an electronic database at icddr , b . To initiate the phone-based reporting system , we first collected mobile phone numbers of the key sales staff at each pharmacy and entered them into the phone system database . We instructed pharmacy staff to call the system number and hang-up after the first ring ( this allowed the pharmacy personnel to avoid incurring the cost of the call ) . This ‘missed call’ triggered the phone system to ( 1 ) associate the caller with a specific pharmacy , and ( 2 ) call the individual back with a prerecorded message ( in Bengali ) asking ‘how many ORS customers have you had today ? ’ In response , the sales person entered the digits corresponding with the number of customers . The system then read back the number they entered and provided and an opportunity to correct any mistakes . Following the call , the system sent a text message back to the sales person confirming that their report had been received . If no calls were received from a pharmacy within a 36-hour period , the system automatically called each number associated with the pharmacy a maximum of two times to attempt to collect data on sales . If none of these attempts led to a successful sales report , an SMS alert was automatically sent to the field team and a phone call or in-person visit was made by study staff to troubleshoot and to collect data from the missing period . Field teams visited each pharmacy weekly to collect the paper-based forms and solve any difficulties faced by the pharmacy staff in completing the forms or using the phone-based system . Pharmacies were paid approximately $5 USD ( a loaf of bread costs approximately $0 . 50 USD at the time of writing this manuscript ) per month as compensation for their participation . Our formative research indicated that the majority of individuals from Arichpur who seek care at hospitals for diarrhea attend one of two main hospitals , the Tongi Sub-district Hospital and the icddr , b Dhaka Hospital . From April through October 2013 , we screened patients at both hospitals on admission to identify Arichpur residents aged 2-years or older hospitalized with diarrhea . Consenting patients provided a stool sample for cholera culture . Study staff visited households of confirmed cholera cases within 3 days of the hospital visit and all household members with diarrhea older than 2 years were asked to provide a stool sample for cholera testing . Stool samples from households and the Tongi hospital were transported in Cary-Blair media to the microbiology lab at the icddr , b Dhaka hospital . All samples were cultured using standard methods onto taurocholate-tellurite-gelatin agar media [16] . In addition , a portion of each sample was enriched in Bile Peptone broth overnight and then cultured for enrichment to improve specificity [17] . Trained study staff asked each eligible patient or their guardian for their written informed consent to participate in the study . This study was approved by the Johns Hopkins University Institutional Review Board and the icddr , b Ethical Review Committee . We first explored the time-lagged relationship between community-wide excess ORS sales and hospitalized diarrhea cases ( separately for cholera-confirmed and all diarrhea ) by estimating the cross correlation function at different time lags ( +/- 14 days ) along with asymptotic 95% confidence intervals . We then modeled a series of lagged relationships with generalized linear models assuming a Poisson or quasi-Poisson error distribution adjusting for one potential confounder , the mean daily temperature . We used the Akaike Information Criteria ( AIC ) [22] from Poisson models at different lags to choose the best lag for each outcome ( S1 and S2 Tables ) . Due to the over-dispersed nature of the diarrhea outcomes , and based on previous publications [23 , 24] , we used quasi-Poisson models for the main analyses . All analyses were conducted using the R statistical programming language ( version 3 . 0 . 3 ) and many plots were produced using the ggplot2 package [25 , 26] . Source code and data for these analyses are available at https://github . com/HopkinsIDD/cholera-ORS/ .
We identified 315 residents of Arichpur hospitalized for diarrhea from 17 April to 30 October 2013 ( Fig 2A ) ; 155 from the icddr , b Dhaka Hospital and 160 from Tongi District Hospital . Among these , 69 ( 22% ) had a positive stool culture for cholera with 28 ( 41% ) of these admitted to the Tongi Hospital ( Fig 2B ) . While we did not collect age data for all diarrhea cases , the median age of the 53 confirmed cholera cases for which we have age-data on was 22 ( IQR 14–35 ) . These confirmed cholera cases were well dispersed throughout the study area though we detected two significant space-time clusters of cholera , one from 3-July-2013 to 8-July-2013 ( Cluster 1 ) and the other from 20-August-2013 to 29-September-2013 ( Cluster 2 , S1A Fig ) . On 31% of days ( 62/197 ) no eligible patients were hospitalized for diarrhea . We detected confirmed cholera on 23% ( 45/197 ) of the study days . On days we identified diarrhea patients , we enrolled between 1 and 9 cases per day , up to 5 of which had cholera cultured from their stool ( Fig 2 ) . We initially enrolled 50 out of the 124 enumerated pharmacies into our surveillance network and randomly substituted two new pharmacies for two that dropped out within the first month of the study . Few pharmacies had days with no sales report after their initial enrollment , with an average of 0 . 28 days of missing data per pharmacy ( range 0–2 ) . In total we recorded 47 , 215 customers purchasing 140 , 614 packets of ORS over the study period . Pharmacies had an average of 5 ORS customers per day ( range 0–68 , interquartile range [IQR] 2–6 ) each purchasing on average 2 . 8 ORS packets ( range 1–60 , IQR 1 . 8–3 . 3 ) . On average 37% percent of the ORS sales were intended for individuals less than 15 years old . We found highly variable patterns in normalized ORS sales between pharmacies throughout the study period with no apparent synchrony in periods of higher or lower than expected sales ( Fig 3B ) . For example , early on in the surveillance period a number of pharmacies had a period of higher than expected sales ( e . g . , pharmacies on the top rows of Fig 3B ) , while others had lower than expected sales ( e . g . , pharmacies on the bottom rows of Fig 3B ) . Community-wide excess sales , a statistic capturing deviations in ORS customers throughout the community , ranged from -1 . 0 in October to 0 . 5 in June ( Fig 3C ) . While we identified groups of pharmacies with similar ORS sale trajectories , we found no clear evidence that these clustered together geographically within the community . However , we found that pharmacies tended to have increasing similarity up to distances of about 100 meters from one another ( S2 Fig ) , though this trend was not statistically significant ( by t-test on linear spline regression model ) . Given that we found some evidence of space-time clustering of cholera cases , we explored whether pharmacy sales within the cluster were elevated compared to those outside of the cluster . We found that ORS sales during the clustering period from pharmacies inside the primary space-time cholera cluster tended to be higher than those outside the cluster ( S1B Fig ) though the differences were small . In analyses unadjusted for mean daily temperature , ORS sales were significantly associated with incident hospitalizations for diarrhea up to 9 days later , with the peak association occurring 7 days after the sales . One additional community-wide excess ORS sale was associated with a 3 . 5 fold rise in diarrhea hospitalization risk seven days later ( 95% Confidence Interval ( CI ) 1 . 74–7 . 18 ) . Similarly , using the cross correlation function , we found that diarrhea cases were significantly associated with community-wide ORS sales up to a week before ( Fig 4A ) . While unadjusted analyses may be helpful for surveillance , daily temperature may confound our estimates of the relationship between ORS sales and diarrhea . In analyses adjusted for the mean daily temperature , the association between ORS sales and incident diarrhea was attenuated , with no significant association ( Relative Risk [RR] 2 . 03 , 95% CI 0 . 89–4 . 76 ) between ORS and diarrhea found in the best fitting ( 7-day lag ) model . Mean daily temperature explained more of the variation in diarrhea incidence in this model , with a one degree Celsius increase in temperature ( above the period mean ) associated with a 14% ( RR 1 . 14 , 95% CI 1 . 02–1 . 28 ) increase in the risk diarrhea hospitalization risk seven days later . However , ORS sales were significantly associated with incident diarrhea hospitalizations up to 2-days later in adjusted models , with one additional community-wide excess ORS sale was associated with a 2 . 7 fold rise in diarrhea hospitalization risk two days later ( 95% CI 1 . 15–6 . 47 ) . At short time lags ( 1–2 days ) , mean daily temperature was not independently associated with diarrhea incidence after adjusting for ORS sales . ORS sales were associated with cholera-confirmed diarrhea hospitalizations only up to one day later in both unadjusted and adjusted models . The best fitting model and the cross correlation analysis suggested that the peak association occurred one day after the sale ( Fig 4B ) . We find that one additional community-wide excess ORS sales unit was associated with an 11 . 1 fold higher cholera risk one day later ( 95% CI 2 . 46–53 . 44 ) in an unadjusted model and a 6 . 5 ( 95% CI 1 . 15–40 . 49 ) fold increase in a model adjusted for mean daily temperature . Mean daily temperature was not significantly associated with cholera hospitalizations in this best fitting model .
Through tracking the sales of ORS at pharmacies and conducting surveillance for hospitalized diarrhea cases in an urban Bangladeshi community , we find that ORS sales were significantly associated with hospitalized diarrhea cases up to nine days later in unadjusted models and two days later in adjusted models , and hospitalized cholera cases up to one day later . These results suggest that first-line healthcare providers could serve as the basis for a new avenue for disease surveillance in low-resource settings where traditional surveillance systems may not adequately capture disease trends in the community . To our knowledge , this is the first study in a low resource setting to explore the association of over the counter sales with diarrhea incidence . Previous studies in North America and Europe , have found conflicting results on the utility of this approach for detecting outbreaks of GI illnesses [8 , 9] . One study in Canada , found over-the-counter sales of anti-diarrheals and anti-nauseants to increase contemporaneously with diarrheal cases ( hospitalized and identified through case investigations ) in outbreaks of cryptosporidium , campylobacter , and E . coli O157:H7 [9] . Another study , in California ( USA ) , found no significant association of over-the-counter pharmacy sales and county-level reports of GI illness [8] . Local diarrheal disease epidemiology , care-seeking behaviors and statistical methods used to assess the relationships between sales and diarrhea incidence may be responsible for these apparent differences . The estimated short time lag between ORS sales and cholera compared to that of non-cholera diarrhea could be due to differences in the natural histories of the putative pathogens . While the incubation period of cholera is unlikely to be significantly faster than many other common gastro-intestinal illnesses , our findings could be due to the relative rapidity of severe diarrhea caused by cholera [27 , 28] . Alternatively , differences in care-seeking behavior in response to cholera and non-cholera diarrhea could contribute to differences in their lagged association with ORS sales . While the phone-derived sales data reflected the data collected through the paper system , more work is needed to ensure that pharmacies regularly report and that gaps in reporting can be accurately and efficiently backfilled . With paper-based forms , this was often done through a review of the pharmacy records , but the phone-based system was only designed to ask whether the first report after a reporting gap was cumulative ( i . e . , did the current report include all the sales since the last report ) or whether it pertained only to new sales that day . Future implementation of this system should improve on the ability to fill in non-reporting gaps . More research is needed on the ways in which people most easily interact with a phone-based reporting system ( e . g . , possibly the inclusion of SMS/text message-based reporting ) to improve data quality . These modifications may ultimately lead to a viable electronic system producing timely data in low resource settings with similar data quality to paper forms . Though promising , our findings come with a number of limitations . Our analyses are based on data from one small community over the course of a single , six-month , cholera season , limiting the generalizability of our findings . For these results to be more practically useful in outbreak detection , longer-term surveillance of both sales and disease is needed . With more data across seasons , a robust predictive model capable of continuously integrating new data points could be constructed , which could allow for diarrhea and cholera incidence forecasts or development of simple outbreak alert algorithms similar to those proposed for other diseases like influenza [29] . With an enhanced phone-based system and improved algorithms using data from a larger network of pharmacies , public health practitioners could receive actionable data on a daily basis , if not in real-time . This delay would likely be small enough to take advantage of the estimated lagged relationships between ORS sales and diarrhea ( and cholera , 1–9 days ) and improve the public health utility of this method . Further exploration of the spatio-temporal patterns of ORS sales and their specificity for diarrhea and/or cholera cases is warranted . If sales in pharmacies around clusters of cases are elevated before cases appear at clinics , a modified version of our approach may not only provide a case alert about the influx of cases , but also the potential locations . Additionally , we illustrate the potential , yet not statistically significant , correlation in sales ORS trajectories for pharmacies within 100 meters of one another . With more data and alternative analyses , understanding this relationship between sales at different pharmacies could help in more efficiently choosing surveillance sites . We reduced each pharmacy’s sales trajectory to a single dimension with a clustering algorithm and this does fully capture the essence of the complex sales patterns . Future analyses could use additional dimension reduction techniques to characterize pharmacies’ sales patterns . Moderate amounts of missing data , like what we observed with the phone-based system , make summary statistics like community-wide excess sales less useful due to the high variation in baseline sales between pharmacies . This measure is robust to randomly missing pharmacy reports but could lead to spurious results when missing data ( i . e . , non-reports ) are associated with individual pharmacy sales and diarrhea in the community . If the relationship between ORS and diarrhea in the community is to be explored with more missing reports or missing reports that are thought to be non-random , community-wide sales should be modeled in a more structured manner . Pharmacies and drug sellers in Dhaka sell an array of diarrhea-related products , including antibiotics; and many administer intravenous rehydration therapies . More research is needed to understand if different diarrhea-related products may better predict diarrhea and cholera incidence in the community , and whether they may do so with longer lead times . When conducting the formative research for this study we learned that ORS is often used for rehydration even in the absence of diarrhea on hot days . Unsurprisingly , we found that ORS sales within this study were significantly associated with mean daily temperature ( S4 Fig ) . We also found that mean daily temperature was associated with incident diarrhea though not with cholera . Previous analyses of non-cholera diarrhea from Bangladesh have shown a similar trend of increasing diarrhea with increasing temperature [30 , 31] , though a more nuanced relationship with cholera and temperature has been noted in other analyses [32] . Our estimates of the relationship between lagged ORS sales and diarrhea and cholera suggest that future predictive models should consider the influence of temperature on the relationship between sales and incidence . Combinations of different products , with changes based upon season or other exogenous factors , may allow for enhanced disease-specific surveillance . In other areas with high diarrhea burden and poor health surveillance systems , like much of Sub-Saharan Africa , ORS use is low , particularly when care is provided by the private sector like the pharmacies and drug-sellers within this study [33 , 34] . Expanding this concept to populations beyond Bangladesh will require careful study of their care seeking behaviors and may require tracking multiple products simultaneously . This study illustrates the potential to use first-line health providers in both epidemiologic surveillance and research within low-resource settings . While , significant hurdles remain to translate these results into practical and scalable surveillance tools , results from this study point toward many possible solutions .
|
In Bangladesh , people often purchase oral rehydration solution ( ORS ) at their neighborhood pharmacy to treat diarrhea , including cholera . Over-the-counter sales have been used for epidemiologic surveillance , but rarely in low-income countries . With few early indicators for cholera outbreaks in endemic areas , diarrhea-related product sales may be a useful surveillance tool . We tracked daily ORS sales at pharmacies and drug-sellers in an urban Bangladesh community with both a mobile phone and paper-based system while conducting surveillance for diarrhea hospitalizations among residents . We found that increases in ORS sales were significantly associated increases in hospitalized diarrhea cases up to two days later and hospitalized cholera cases up to one day later . Our findings suggest that surveillance systems based on over-the-counter product sales may be a feasible and useful way to detect outbreaks in low-income settings and that mobile technology may make it even easier to collect implement .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Tracking Cholera through Surveillance of Oral Rehydration Solution Sales at Pharmacies: Insights from Urban Bangladesh
|
The place theory proposed by Jeffress ( 1948 ) is still the dominant model of how the brain represents the movement of sensory stimuli between sensory receptors . According to the place theory , delays in signalling between neurons , dependent on the distances between them , compensate for time differences in the stimulation of sensory receptors . Hence the location of neurons , activated by the coincident arrival of multiple signals , reports the stimulus movement velocity . Despite its generality , most evidence for the place theory has been provided by studies of the auditory system of auditory specialists like the barn owl , but in the study of mammalian auditory systems the evidence is inconclusive . We ask to what extent the somatosensory systems of tactile specialists like rats and mice use distance dependent delays between neurons to compute the motion of tactile stimuli between the facial whiskers ( or ‘vibrissae’ ) . We present a model in which synaptic inputs evoked by whisker deflections arrive at neurons in layer 2/3 ( L2/3 ) somatosensory ‘barrel’ cortex at different times . The timing of synaptic inputs to each neuron depends on its location relative to sources of input in layer 4 ( L4 ) that represent stimulation of each whisker . Constrained by the geometry and timing of projections from L4 to L2/3 , the model can account for a range of experimentally measured responses to two-whisker stimuli . Consistent with that data , responses of model neurons located between the barrels to paired stimulation of two whiskers are greater than the sum of the responses to either whisker input alone . The model predicts that for neurons located closer to either barrel these supralinear responses are tuned for longer inter-whisker stimulation intervals , yielding a topographic map for the inter-whisker deflection interval across the surface of L2/3 . This map constitutes a neural place code for the relative timing of sensory stimuli .
A fundamental question in computational neuroscience asks how the brain represents the relative timing of stimuli as they move between sensory receptors , e . g . as a light source moves relative to the retina , or as contact moves between touch sensors on the fingertip . For over 60 years Jeffress’ place theory [1] has remained the dominant model . The idea is that coincidence detector neurons receive input from sensors after delays governed by the distance of the neuron from either sensor . The inter-sensor time difference is encoded by the location of neurons that are active because their connection delays exactly compensate the inter-sensor stimulation interval . The place theory therefore suggests an important role for neural geometry in computing the motion of sensory stimuli . Strong support for Jeffress’ place theory has been provided by a number of studies of midbrain neurons in auditory specialists like the barn owl , who locate sound sources by resolving small differences in the arrival time of sounds at either ear ( see ref . [2] for a review ) . Evidence from the mammalian auditory system is less conclusive because , for example , rabbit auditory cortex neurons are tuned to inter-ear time differences that are too long to attribute to inter-neuron distances alone [3] ( see also refs . [4] , [5] , and ref . [6] for an alternative mechanism based on slow lateral connections ) . However few studies have investigated how inter-sensor time-differences might be resolved in specialist mammalian sensory systems . Tactile specialists like rats , mice , shrews , and seals determine the form and motion of tactile stimuli using prominent arrays of whiskers ( vibrissae ) on the face [7] , [8] . For example , shrews hunting in the dark can use their whiskers to localise particular body-part shapes on fast-moving prey animals [9] . Specific to the whisker system is a precise topographic correspondence between the individual sensor and its neural representation . Deflection of adjacent whiskers A and B on the face evokes the largest amplitude and shortest latency responses in adjacent cortical columns A and B in the somatosensory ( barrel ) cortex . This precise mapping , as well as observations of sub-millisecond temporal precision throughout [10]–[12] , makes the whisker-barrel system ideal for exploring the impact of neural geometry on neural computation . A consistent finding across studies in the rat and mouse somatosensory cortex is that responses vary with the time interval between adjacent whisker stimulation [13]–[24] . A useful metric for comparing the response to a two-whisker stimulus to the response to the individual whisker deflection is the facilitation index [17] , defined as ‘the response to paired deflection of whiskers A and B divided by the sum of the response to deflection of whisker A deflected alone and the response to whisker B deflected alone’ or . In layer 2/3 barrel cortex ( L2/3 ) in particular , paired stimuli in which the adjacent whisker deflection precedes by 20– typically evoke sublinear responses ( ) . For a range of near-simultaneous deflections , a number of studies have also reported supralinear responses ( ) , again particularly in L2/3 neurons [16]–[18] , [22] , [23] ( but see ref . [25] ) . Interestingly Shimegi et al . [18] reported that septa-related neurons in L2/3 , located at the midline area between two barrels , were more likely to show response facilitation for short-interval stimuli , whereas barrel-related neurons were more likely to show response suppression by prior deflection of the distal whisker at longer intervals ( see Figure 1 ) . Plots of the relationship between the inter-whisker-interval and the response magnitude for individual neurons showed evidence of tuning to particular short intervals . Together these results suggest that the location of the L2/3 neuron relative to the underlying barrel geometry is important in determining its response to a two-whisker stimulus . One explanation for the different responses of barrel-related and septa-related neurons , as summarised in Table 1 , is that they reflect the operation of different mechanisms for integrating adjacent-whisker signals in distinct barrel and septal circuits ( see refs . [26]–[28] ) . However an alternative hypothesis , inspired by the place theory , is that the differences reflect an underlying continuum of responses , which are determined by the location of the neuron with respect to the two cortical columns . This hypothesis would allow for , although it would not require , an essentially homogeneous population in L2/3 . According to this alternative hypothesis , the relationship between the inter-whisker deflection interval and the facilitation index in L2/3 neurons may be determined by differences in the arrival times of synaptic inputs that originate from either barrel . These differences may be attributed to inter-soma distance-dependent delays in the feed-forward projection from the major input in layer 4 barrel cortex ( L4 ) . This hypothesis is supported by estimates of the speed of the projection between L4 and L2/3 neuron pairs that are relatively slow , around 0 . 2 meters per second for excitatory and inhibitory post-synaptic neurons [29] , [30] . In this paper we show that simulated barrel cortex neurons that receive synaptic inputs with onset times constrained to embody this hypothesis can account for all of the trends relating to the stimulus interval in the data of ref . [18] . We show that a natural prediction of the model is the existence of a topographic mapping of the inter-whisker deflection interval across the surface of L2/3 . Specifically , supralinear population responses will peak at short non-zero intervals in neurons located closer to the barrel representing the later of the two deflected whiskers . The responses of individual L2/3 neurons satisfy the basic requirements for a motion detector , and across the population these responses encode a range of stimulus motion velocities . Results therefore suggest that two-whisker timing is represented by a place code in L2/3 barrel cortex . More generally , the lateral displacement of active neurons due to distance-dependent delays on projections between cortical columns can be used to compute the sequence and timing of events between the sensory stimuli represented by activity in those columns . The results are interpreted as evidence in support of the place theory as a general model of cortical processing of spatiotemporal information .
We hypothesise that distance-dependent delays associated with inter-columar projections in sensory cortex can be used to extract the relative timing of sensory events . Specifically , delays in the projection from layer 4 ( L4 ) to layer 2/3 ( L2/3 ) barrel cortex might generate selectivity to the inter-whisker deflection interval for adjacent whiskers . To test the hypothesis , the latencies of synaptic inputs to a leaky integrate and fire neuron were constrained to reflect the range of geometries that characterise the L4 to L2/3 projection . To validate the model , we recreated an adjacent-whisker paired-deflection study [18] , and compared responses of neurons in different cortical locations to stimuli in which the whiskers were deflected through a range of intervals . The simplified model is based on three main assumptions , which are described with respect to the validation data in terms of adjacent whiskers A and B , but which in principle apply to a general model of cortical responses to arbitrarily complex multi-whisker deflection patterns . The first assumption is that , upon whisker stimulation , inputs to L2/3 tend to originate from L4 neurons at the center of the corresponding barrel in L4 . Therefore , in the model , the input layer L4 is collapsed down to just two point sources , with activity at each source representing the deflection of the corresponding whisker A or B . The second assumption is that the excitatory and inhibitory synaptic inputs evoked by deflection of whisker A and by deflection of whisker B arrive at a population of L2/3 neurons situated above and between corresponding barrels A and B . Therefore , in the model , each L2/3 neuron receives just four inputs , although each represents the total contribution of many similar synaptic contacts . The third assumption is that the time taken for a L2/3 neuron to register a synaptic input is proportional to the straight-line distance between the L4 and L2/3 neuron . Therefore , in the model , we assume that the time of arrival of each synaptic input is a linear function of the distance of the L2/3 neuron from either point source in L4 , and we refer to the associated constant of proportionality as the connection speed . This simplified model of the neural geometry may deviate from the true situation . For example , if the signalling delays are due to the axonal propagation speeds , then delays could be modified by the morphology of L4 axons , which branch vertically and laterally into L2/3 [31] , [32] . Delays could also be modified by particular branching patterns that vary systematically with the location of the neuron in the home barrel [33] . We choose not to explicitly model the variety of axonal morphologies , firstly to keep the model formulation simple , secondly because L4 to L2/3 signalling delays are well predicted by the straight-line inter-soma distance [29] , [30] , [34] , and thirdly because post-hoc simulations which considered a laterally-branching axonal morphology did not significantly alter the results . Furthermore , recurrent interactions within L2/3 are not modelled explicitly , because they would occur subsequent to the initial activation of L2/3 , and thus could only affect the afferent response after the critical first spike response has been determined ( see Discussion ) . Similarly , modelling each L4 input source as a discrete representation of one whisker is justified because multi-whisker responses in L4 are thought to be due to latent contributions from intra-cortical mechanisms [19] ( see Discussion ) . The following sections outline how each assumption is represented formally in a model that we refer to as the distance-dependent delay hypothesis . The plausibility of each assumption , the impact of each simplification , and the alternatives to each are considered in Discussion . The thalamocortical volley of excitation from thalamus to L4 and then up into L2/3 [34] , [35] is closely followed by a volley of disynaptic inhibition , mediated by a small number of interneurons in L4 [36] , with a diverse range of morphologies [32] . We posit that the main excitatory input to L2/3 is derived from direct synaptic connections from excitatory neurons in L4 , and the main inhibitory inputs are derived indirectly from excitation of L4 inhibitory interneurons . The circuit therefore consists of three connections: an excitatory connection from L4 to L2/3 , an excitatory connection onto the L4 inhibitory interneuron , and an inhibitory connection from the L4 interneuron to the L2/3 neuron . According to the distance-dependent delay hypothesis each connection has an associated delay . The onset time of the direct excitatory synaptic input at the L2/3 neuron is proportional to its distance from the barrel center . To model the indirect inhibition through an inhibitory interneuron we use a time delay proportional to the L4 to L2/3 distance plus a constant time delay accounting for the distance of the interneuron and its spike generation time . The circuit therefore has three parameters: the speed of the excitatory pathway between L4 and the L2/3 target neuron ( ) , the speed of the inhibitory pathway between L4 and the L2/3 target neuron ( ) , and a fixed latency representing the delayed onset of the spike in the inhibitory interneuron ( ) relative to the onset of excitation in L4 . For neurons in the barrel cortex , the principal whisker is typically defined as the one which , upon deflection , elicits the shortest latency and/or the largest-amplitude response . Neurons of a particular barrel column tend to share the same principal whisker , the one which on the face is isomorphic with the position of the barrel in the grid of barrels . For a given neuron all three criteria usually select the same whisker . These constraints can be built into the model if , for progressively longer inter-soma distances , whisker-evoked inhibition arrives progressively earlier than excitation . This pattern of delays requires that inhibitory connections are faster than excitatory connections , and that the onset of inhibition is delayed relative to the excitation . This is achieved in the model by setting and . In the analysis presented by Shimegi et al . [18] , against which the model will be validated , L2/3 neurons were characterised by their horizontal location with respect to two underlying barrel columns . The geometry is shown in Figure 2 . In the model axes and refer to orthogonal axes of the plane tangent to the pia matter of the brain ( i . e . , the plane tangential to the cortical surface ) [37]; specifically is aligned with barrels that correspond to a row of whiskers on the face , and is orthogonal in the ‘tangential plane’ . The axis is normal to the tangential plane . Axes and will henceforth be referred to as the horizontal and vertical axes respectively . In the model , L2/3 neurons will be parameterised only by their horizontal location relative to the two input sources in L4 . In effect , this means reducing the three spatial dimensions in which intra-cortical connections are defined to just two spatial dimensions by setting . In this way we can define the position of two sources in L4 at . Similarly we can describe L2/3 as a one-dimensional string and uniquely describe the location of individual L2/3 neurons along the string in terms of . For example the neurons at , , and are L2/3 neurons located directly above barrel A , above barrel B , and above the midline respectively . The Euclidean distance of each L2/3 neuron from the two sources can now be written in terms of : ( 1 ) ( 2 ) For the analyses presented in Results , the input sources were located at and the two layers were separated by vertical distance . We will henceforth refer to and as inter-soma distances . Reducing the description of the neural geometry in this way makes interpretation of the behaviour of the model tractable , and it allows for a direct comparison with the available electrophysiological data . We note that using an alternative geometry has little impact on the main results , as considered in detail in Discussion . The L2/3 neuron receives excitatory and inhibitory synaptic inputs from each stimulated whisker . Thus , under two-whisker stimulation , the time of each input is given by: ( 3 ) ( 4 ) ( 5 ) ( 6 ) The inter-whisker interval ( ) is the time of deflection of whisker A , relative to whisker B , which is always deflected at time 0 . Thus if whisker A was deflected before whisker B , if whisker B was deflected before whisker A , and if then the whiskers were deflected simultaneously . The relationship between the inter-soma distance and the onset time of excitation and inhibition is illustrated in Figure 3A . The connection speeds were chosen to be and , which are in the range of estimates derived from electrophysiological data [29] , [30] , but we note that similar analyses have estimated speeds as slow as [34] . The constant was chosen to delay the onset of inhibition relative to excitation by for the neuron located closest to either barrel center , i . e . , . With the inter-soma distance constrained by the geometry of Equations 1 and 2 , the input onset times , described by the linear functions in Figure 3A , become hyperbolic functions of the neuron location , as shown in Figure 3B . The model neuron is a simple integrate and fire neuron with inputs in the form of excitatory and inhibitory post-synaptic conductance changes ( EPSCs and IPSCs ) . Parameters followed those reported by Puccini et al . [38] as a guide for neurons in the barrel cortex . The time course of each input , following its onset at time , , or , was modelled as a normalised difference of two exponentials: ( 7 ) The normalisation term , where , ensures that the potential peaks at 1 . For excitatory synapses and simulating AMPA receptor channel opening [39] , and ensuring that excitatory inputs peak at . For inhibitory synapses and as used by Puccini et al . [38] to model GABA receptor channel opening , peaking later than the EPSC at as seen in electrophysiological data ( e . g . , ref . [40] ) . The maximum EPSC amplitude was and the maximum IPSC conductance amplitude was ( similar to ref . [38] ) . The relative amplitude and time course of the excitatory and inhibitory post-synaptic currents are illustrated in Figure 3C . For the L2/3 neuron we used a standard leaky integrate and fire neuron [41] , again with parameters guided by those from ref . [38]: ( 8 ) where the membrane time constant , the resting potential , the reversal potential for synapses of type inhibitory , and for excitatory synapses . The leak conductance was and hence the membrane resistance . Gaussian noise with standard deviation was added to the membrane potential at each time step . Integration was by the forward Euler method ( ) . When the membrane potential reached a spike was recorded , and the membrane potential was set to .
To anticipate how a L2/3 neuron might respond to independent deflections of either whisker , we first determine when the onset times of the EPSC and IPSC evoked by deflection of that whisker will be coincident . We derive the time of coincidence by setting the onset times to be equal and rearranging: ( 9 ) ( 10 ) Therefore we can determine that when and hence we would expect to see the largest responses to deflection of whisker A because the excitatory input precedes the inhibitory input . To test this , neurons through the range of locations were stimulated by applying a deflection to either whisker A or whisker B in isolation . Analogous to the experimental procedure of ref . [18] , each trial began prior to the onset of the first whisker deflection and ended after the onset of the second deflection . Spike counts were calculated over this time window for the results of all simulations , however we note that spikes were precisely timed to the whisker stimuli and so this choice of time window is not critical for the behaviour of the model ( see Figure S1 ) . The spike rate is shown as an average over 50 trials in Figure 4A to allow direct comparison with the results of ref . [18] , and averaged over 5000 trials for clarity in 4B . As expected , neurons located closer to a particular barrel spike more often in response to deflection of the corresponding whisker . As the distance of the neuron from either source increases , the excitatory and inhibitory inputs evoked by the corresponding neuron register at the neuron closer together in time and thus the window of opportunity in which the EPSC can cause a spike decreases . At longer inter-soma distances , the IPSC precedes the EPSC , and effectively silences the neuron . These observations agree with the notion of the principal whisker as that represented by the barrel closest to the neuron , and which evokes the shortest latency and largest amplitude response . Figure 4 shows the linear sum of the response to independent deflection of both whiskers . These values for the linear sum are later used to construct facilitation index scores from the average spike counts obtained in paired whisker-deflection trials . For independent deflections of either whisker , we have seen that the spike rate is dictated by the sequence and relative timing of the synaptic inputs . Responses to paired whisker deflection stimuli are more complex because they are dictated by four PSCs rather than two and also by the . However similar analysis of the relative arrival times of PSCs can be used to anticipate these responses . To this end it is useful to consider regions of the space of possible neuron location and inter-whisker deflection intervals ( henceforth – space , see Figure 5A ) that are delineated by different ordering of arrival times of the four PSCs . These regions are delineated by loci representing coincident arrival of each possible pair amongst the four PSCs . Equations 9–10 represent two such pairs . As their solutions are not dependent on the , Equations 9–10 describe four loci , which when plotted are straight lines at constant values of that divide – space into five columns in Figure 5A . Solutions for the other four pairs of PSCs can be written as functions of as follows: ( 11 ) ( 12 ) ( 13 ) ( 14 ) The solutions to Equations 11–14 are also plotted in Figure 5A , and they further divide the columns into ‘rows’ . For each region of the graph we can use the equations to state the sequence of inputs for each synaptic pair . This is done by setting all signs to signs in Equations 9–14 . The eight inequalities that define each region of the graph can then be combined to give the order of all four synaptic PSCs , and the twenty-four possible PSC orderings take the form , for example , in the top-left region of – space shown in Figure 5A . Considering now only whether each synaptic event in the input sequence is excitatory or inhibitory , we can describe the input to the L2/3 neuron more simply . This effectively reduces the twenty-four PSC sequences to just six different orders in which excitation and inhibition can arrive at the neuron . Figure 5B shows how each of the six orderings delineates a zone in – space . For a range of short interval stimuli , neurons situated near the midline receive both excitatory inputs before both inhibitory inputs . They receive inputs in the order , which can be read as ‘two excitations followed by two inhibitions’ . This zone is coloured dark blue in Figure 5B . It is in this zone that we would expect to observe the greatest spike rate because neither IPSC precedes the EPSCs . Notice that this zone is oriented diagonally in – space , and therefore neurons in different locations near the midline will prefer a range of ( short ) . Similarly we can expect that the greatest suppressive interactions will be displayed in the yellow ( ) , brown ( ) , and orange zones ( ) , in which an IPSC event is always registered first . Of these zones the orange will be expected to yield the smallest suppression as the second IPSC is preceded by both EPSCs . In the blue zone ( ) we might expect just one of the whisker deflections to evoke a response , as the second EPSC will be silenced by two preceding IPSCs . In the cyan zones ( ) both EPSCs are followed immediately by an IPSC . Therefore we might expect that if the two EPSC/IPSC pairs are separated sufficiently in time for the neuron to respond to them independently , i . e . , if the first inhibition has little effect on the second excitation , then the response will resemble the linear sum of that evoked by either whisker deflected independently , and hence the facilitation index score here will be around one . Neurons through the range of locations were stimulated by applying paired deflections to whisker A and whisker B in sequence . By analogy with the experimental procedure of ref . [18] , each trial began prior to the onset of the first whisker deflection and ended after the onset of the second . The spike rate is shown as an average over 50 trials in Figure 5C . As anticipated , the greatest activity was evoked in neurons around the midline ( ) when the whiskers were deflected through a range of short inter-whisker intervals ( ) . Within this range neurons located left of the midline and therefore closer to barrel A responded maximally to slightly positive inter-whisker intervals where whisker B was deflected before whisker A . Neurons to the right of the midline and therefore closer to barrel B responded maximally when whisker A was deflected before whisker B at short intervals . For intervals longer than around in either direction , and for neurons further from the midline than around half a millimetre , responses were much smaller . In a region of – space roughly corresponding with the light blue zone in Figure 5B , responses were more variable at around 0 . 2 spikes per stimulus . These results from the full spiking model fit well those expected based on the relative timing of the synaptic inputs . Thus changing the relative timing of the synaptic inputs with distance-dependent delays alters the response of the neuron to paired whisker stimuli in a predictable way . A major feature predicted by the simulation data is a mapping of short interval stimuli to the location of the most active L2/3 neuron . The simulation data presented thus far suggest that distance-dependent delays in the L4 to L2/3 projection can generate a spatial encoding of the relative timing of whisker inputs for short interval stimuli . But to what extent do these observations match up with experimental data ? To answer this question we look first at the responses of individual model neurons to the range of different interval stimuli . Figures 6A and 6B show the average spike rate for an individual neuron located either close to barrel B or between barrels A and B respectively . The neuron in Figure 6A was located approximately to the right of the midline . Also indicated in the figure is the linear sum of the response of this neuron to either whisker deflected in isolation . Where paired stimuli evoke responses equal to this value , a facilitation index of 1 would be measured and we would conclude that no facilitatory interaction had occurred . Where it is less , suppression would have been measured , and where it is greater facilitation would have been measured . The neuron in Figure 6A shows no facilitatory interaction when whisker B ( the principal whisker ) is deflected prior to the adjacent whisker A . However for slightly negative intervals strong facilitation was measured , with the average spike count exceeding the linear sum baseline three-fold or more around a peak when whisker A is deflected before whisker B . When whisker A precedes by more than the response is strongly suppressed and almost no spikes are evoked . The suppression recovers towards the linear sum baseline for intervals exceeding . For the example midline neuron shown in Figure 6B facilitation appears more symmetrical around the zero inter-whisker interval . Facilitation peaks for simultaneous intervals and fluctuates around baseline for longer intervals in either direction . The peak in the average spike count is larger than that for the previous neuron , as is the linear sum response used to compute the strength of its facilitatory interaction . Equivalent plots for individual L2/3 neurons , found in refs . [17] , [18] , [23] , [24] , display similar qualitative trends to those in Figure 6A and 6B , in terms of both the facilitatory interactions and of the average spike counts for independent and paired whisker stimuli . In Figure 6C we group the L2/3 neurons by location as either above barrel A , above barrel B or in the septal region between the barrels . This allows for a direct comparison between the simulation data ( Figure 6C ) and the available experimental data of ref . [18] ( compare with Figure 1 ) . The simulation data share many of the qualities of the experimental data , as summarised in Table 1 . Septal neurons show a large facilitatory peak for near simultaneous paired whisker deflections and for longer intervals in either direction respond with an average , equivalent to the response to either independently deflected whisker . Neurons located above barrel B display on average a lesser facilitatory peak at interval stimuli , are suppressed by prior deflection of whisker B , and display no facilitatory interactions when whisker B is deflected first . Geometry in the model is symmetrical about the midline and therefore the responses are symmetrical about the zero inter-whisker interval . Therefore the above barrel B population display the exact opposite interactions with respect to the interval compared with the above barrel A population . This includes a lesser peak for interval stimuli not apparent in the electrophysiological data . Notice too that the peak of the septal group in the experimental data is for a slightly negative inter-whisker interval . We will shortly demonstrate how an extension to the model , which introduces asymmetries related to the direction in which each whisker is deflected , may account for these differences . For now we note that the population response predicted by the model affords a good match to the experimental data . Instead of asking how L2/3 neurons in particular locations respond to different interval stimuli , we can ask how particular interval stimuli are represented across the population of L2/3 . It is particularly important to consider the population response because even the most effective stimuli typically elicit less than one spike per stimulus in any particular neuron , and so individual spikes yield ambiguous information about the stimulus [42] . Figure 7 shows the distribution of average responses across the population for a range of positive intervals . Each of the short inter-whisker deflection intervals is clearly associated with a tuning curve across the population , with a peak that shifts to the left ( negative ) and scales systematically with the increase in interval . Negative intervals also evoke symmetrical results , i . e . , a shift in peak responses towards neurons on the right , but we do not show them in the figure for clarity . Viewed in this way , it is clear that the model predicts the existence of a topographic map for the inter-whisker deflection interval across the surface of L2/3 barrel cortex . According to the model , paired whisker stimuli should elicit supralinear responses and display a systematic shift in tuning across the population for stimulus intervals ranging to . As well as the representation of the inter-whisker interval across cortical space , it is useful to consider how the stimulus is represented in the timing of spikes . Inspection of maps for the spike timing revealed that in paired-whisker stimulations , spikes were precisely timed to the whisker stimuli . Moreover the largest responses reflected a combination of the delayed response to the principal whisker , as well as the superposition of excitatory influences from both whiskers ( see Figure S1 ) . Therefore the model predicts that the effects measured by ref . [18] primarily operate on the first somatosensory-evoked spikes in L2/3 . Barrel cortex neurons are selective for the direction in which the whiskers are deflected . The mechanism thought to underlie directional selectivity in L4 neurons is similar to that which we have outlined for two-whisker timing , but with distances measured in degrees from the preferred stimulus direction [38] , [40] . Several studies have suggested that direction preferences vary systematically within the barrel column , such that deflection of the principal whisker to the left or right is correlated with increased activity in neurons located to the equivalent left or right of the barrel column [43] , [44] . Therefore we can model the effect of deflecting the whisker in either direction by moving the L4 point source for that whisker in either direction in L4 . Accordingly , to represent a deflection of whisker A to the left ( away from whisker B ) we offset the point source in L4 that corresponds to whisker A by a fixed distance to obtain a new source location at . Deflecting whisker A to the right means moving the point source to and similarly deflecting whisker B to the left or right means moving the second source to . For two whiskers and two deflection directions , possible combinations are both deflections to the left ( leftwards ) , both right ( rightwards ) , A left & B right ( outwards ) , and A right & B left ( inwards ) . Results obtained from the model in these conditions are summarised in Figure 8 . For the analysis shown in Figure 1 , Shimegi et al . [18] deflected both whiskers to the left , and so we consider the leftwards condition first ( Figure 8A ) . Conditions leftwards and rightwards produce symmetrical effects and so we only show results for the former . In the leftwards condition , the relative projections , distances , and geometry are identical to the case where the stimulus originates from the barrel centers . However , each projection is shifted to the left , and so each neuron inherits the input timing of that located to the right . As a result the effects are still symmetrical but they are symmetrical about a new midline that is shifted to the right at . When we average the data across groups defined in terms of the original midline at , as in Figure 8A , we observe systematic asymmetries in the results . The facilitatory peak in the above A group is increased , that in the septal group is shifted towards negative inter-whisker intervals , and the peak in the above B group is decreased . Thus by introducing a topology associated with the stimulus deflection direction , the model can account for each of the previously unexplained observations in the original data . This account is also consistent with the observations of Shimegi et al . [18] and Kida et al . [23] ( but not ref . [45] ) , that preferences for the deflection direction of the principal whisker are strongly correlated with those for the adjacent whisker deflection direction , and with the deflection direction evoking facilitatory interactions when both are deflected in that same direction at short intervals . Predictions of the model for the two stimulus conditions not yet tested experimentally , inwards and outwards , are shown in Figure 8B and Figure 8C . Deflected towards one another ( Figure 8B ) , as may occur when the whiskers encounter a concave stimulus shape , the two stimuli should be represented in the two adjacent sides of the corresponding barrels . This configuration effectively shortens all connection distances , and expands the zone in which both excitatory inputs precede both inhibitory inputs across . Thus the facilitatory interactions are distributed more broadly across the population , and we would expect to see more similar facilitatory peaks amongst the three neuron groups . Conversely if the two whiskers are deflected away from one another ( Figure 8C ) , as may occur when the whiskers encounter a convex stimulus shape or during divergent whisking movements [46] , inputs originate from distal sides of the barrels . This configuration squeezes the zone in which we expect to see facilitatory interactions with respect to , and concentrates them under a single peak in the septal neuron group . Demonstration of effects to the contrary could be used to falsify this aspect of the model . The particular neuron model from which the previous results have been derived was chosen to allow comparison of the results with real biological neuron data . We have shown how the sequence of synaptic inputs due to distance-dependent delays can change the output of the neuron , but we have not yet determined the origin of the non-linear effects underlying the observed facilitatory interactions . To understand this better we tried to reproduce the effects using as simple a neuron model as possible . We found that all of the trends in the full model simulations could be reproduced using a simple linear filter neuron model . The reduced model is: ( 15 ) where or , with output squashed using the logistic output function: ( 16 ) The range of facilitatory interactions can be seen if we interpret either the maximum or the mean value of over time as the spiking probability for each stimulus trial . The logistic output function performs the role of the thresholding operation in the full model . Its form in the full model is affected primarily by the noise , which has a similar effect to the slope of the sigmoid ( slope parameter = 0 . 04 ) , and the relationship between the firing threshold and the synaptic weights and reversal potentials , which essentially sets the inflection point of the sigmoid ( inflection point = 0 . 2 ) . Because both neuron models yield comparable stimulus-evoked interactions , we can be confident that the thresholding non-linearity in the full neuron model , as approximated by the sigmoidal output function in the simpler neuron model , can account for the observed non-linear effects . Comparing the two models in terms of the spike probability is valid in this instance because we observed that in the full model neurons generate less than one spike per stimulus .
A major simplification we made in order to construct the model was to explicitly simulate only four synaptic contacts per neuron , whereas real L2/3 neurons receive hundreds of synaptic contacts originating from L4 [48] . Where possible , the parameters of the full neuron model were derived from existing models or electrophysiological data . However to compensate for the decrease in afferent drive , the spiking threshold was lowered from a realistic to a low . In the final results section we showed that the behaviour of the model is not sensitive to the form of the neuron model chosen , but that each of the trends in the electrophysiological data can be reproduced using a simple sigmoid output function neuron , as used in previous models of the barrel cortex [49] , [50] . Another simplification was to relate the delay on each projection to the straight-line distance between the input and its target . This choice was motivated by several studies reporting an approximately linear relationship between the straight-line inter-soma distance and the associated delay [29] , [30] , [34] . However , the axons of L4 neurons tend to project vertically into L2/3 before turning to branch laterally [51] . Therefore it may be appropriate to consider the Manhattan distance , the vertical plus the horizontal distance , defined in the model by rewriting Equations 1–2 to be of the form . This change has the effect of changing the hyperbolic relationship between and the synaptic onset latency into a piecewise linear relationship . Each of the zones of synaptic input sequence is maintained in – space; hence using the Manhattan distance to compute synaptic input latencies does not change the form of the main results when they are recalculated using this alternative geometry . The model relies implicitly on the assumption that connections between L4 and L2/3 are organised on a finer spatial scale than that defined by the column boundaries , such that the location of the L2/3 neuron determines its response properties . Evidence from several studies supports this assumption . For example calcium transients measured between pairs of neighbouring L2/3 neurons located above the barrel centers are more highly correlated than those between pairs of distant neurons located above the barrel borders [52] . These data suggest that L2/3 neurons receive input from particular regions of the L4 barrel according to their tangential location in the column [52] . More evidence for a sub-columnar spatial resolution of connections is provided by a correlation between the maximally effective direction of whisker deflection for L4 and L2/3 neuron pairs in vertically aligned sub-regions of the barrel column [43] . Similarly , connected thalamic and L4 neuron pairs share tuning to the whisker deflection direction [53] . The mechanism by which the model accounts for tuning to inter-whisker interval is essentially the same as that thought to underlie tuning for the deflection direction in L4 [38] , [40] , [54] , [55] . In both cases the relative latency of inhibition creates a short ‘window of opportunity’ in the post-synaptic neuron , in which excitatory input representing the preferred stimulus can evoke a response . The dependency of the preferred inter-whisker interval on the connection geometry raises the intriguing possibility that tuning for deflection direction in L4 is inherited from the geometry of the thalamo-cortical projection . A reported topographic organisation of directional preferences about the barrel center in L4 could be inherited from a map of direction preferences measured along the major anatomical axis of the thalamic input barreloid [56] . This idea seems plausible given that thalamocortical axon conduction times range from to [57] , and that latencies ranging to can account for responses to preferred and anti-preferred stimuli respectively [38] , [40] . To account for the data of ref . [18] , the model requires that at short inter-soma distances excitation precedes inhibition and for longer distances inhibition precedes excitation ( see Figure S2 ) . This we attributed to differences in axonal conduction velocity on excitatory and inhibitory projections into L2/3 . The origin of the faster inhibition is unlikely to be mediated by L2/3 interneurons , because excitatory connection speeds from L4 to L2/3 interneurons are similar to those from L4 to L2/3 excitatory targets ( compare ref . [30] and ref . [29] respectively ) . The origin is also unlikely to be thalamocortical , because L4 interneurons and L4 excitatory targets are excited after comparable latencies [58] , although interneurons are excited via slightly thicker , shorter , and thus faster thalamocortical axons [59] . Therefore we suggest that differences in speed may be attributable to morphological differences between the axons of L4 inhibitory and L4 excitatory neurons; L4 interneurons are known to branch into L2/3 and extend well beyond the boundary of the vertically aligned barrel [32] . To our knowledge , the axonal conduction velocities for this connection have not been directly measured . Therefore the critical quantitative prediction , that the L4 inhibitory axonal conduction speed must be faster than the L4 excitatory speed , can be used to validate the model in a future experiment . Because each input source in L4 represented the deflection of one whisker , the present model assumed no contribution of sub-cortical mechanisms to the integration of multi-whisker signals . To a first approximation , the barrels in L4 can be considered as functionally separate processing units [34] , [37] . Moreover , although non-linear multi-whisker responses can be evoked in L4 neurons [14] , [15] , [60] , much of the effect may be due to intra-cortical rather than thalamocortical mechanisms [19] , which are most pronounced in non-granular layers [24] , [61] , and which would shape responses only after the first stimulus-evoked spikes had been determined . However , the contribution of sub-cortical mechanisms to multi-whisker integration should not be overlooked; an extended version of the model will be required to explore this important issue in more detail . Tactile stimuli which include three or more whiskers cause suppressive interactions across barrel cortex which serve to enhance the representation of complex multi-whisker deflection patterns [16] , [19] , [24] , [61] . We investigated how additional whiskers are represented according to the model , by simulating the effect of a stimulus moving at various speeds through a row of whiskers which included two , three , four , or five whiskers ( see Figure S3 ) . When the whiskers were deflected simultaneously , the resulting activity across L2/3 was widespread and large and formed a symmetrical pattern , but when the whiskers were deflected consecutively the activity decreased across L2/3 in the direction corresponding to the stimulus motion . In agreement with previous studies the model predicts the existence of an activity gradient that is steeper for slower stimulus motions . A previous modelling study suggested that a spatial gradient in the afferent activation of L2/3 could represent the direction of stimulus motion through the whisker field , and that this representation in L2/3 would be sharpened by recurrent inhibitory interactions [44] . The present model did not consider recurrent inhibition , which is prevalent in L2/3 [35] , [62]–[64] , because it considered primarily how subthreshold inputs interact to generate the earliest spikes in L2/3 ( see Figure S1 ) . We are currently working on a model which extends the present study and that of ref . [44] , to test the hypothesis that regions of contrast in activity due to initial feed-forward interactions are enhanced by subsequent lateral inhibition . This model will also explore how stimulus coding might be affected by distance-dependent weights on synaptic connections , as suggested by recent experiments [62] , [64] . The present simulation results afford an existence proof for a more general hypothesis that the geometry of projections between neighbouring cortical columns could be useful for encoding relative inter-sensor motion speed and direction . In its weakest form the implication of the hypothesis is that interconnection geometry and connection speeds should be considered in detailed cortical microcircuit models if they are to accurately predict the response properties of individual cortical neurons . Given the remarkable spatial relationship between the whisker and its associated barrel column , it is surprising that , with the exception of refs . [65] , [66] and our own previous model [44] , connection geometry has not been an important factor in computational neuroscience models of the barrel system . In its strongest form the implication is that the cortex could carry out specific computations by reading out the tangential position of active cortical neurons . This is essentially the same idea as the place theory proposed by Jeffress [1] . The principle behind our model and the Jeffress model are essentially the same . In both , a bank of coincidence detectors receive input from spatially separated sources after delays governed by the distance from either source , and thus activity in detectors whose connection delays compensate that of the stimulus motion reports the stimulus velocity . It remains to be shown whether tactile specialists such as rats and mice can discriminate adjacent whisker contact times over the range generated in the model , although emerging techniques are allowing the link between barrel cortex activity and performance on tactile discrimination tasks to be explored in unprecedented detail [67] . Jeffress’ place theory can be thought of as a specific case of a more powerful computational principle , recently termed ‘polychronous wavefront computation’ ( PWC ) [68] . In PWC terminology , two sources in the Jeffress model specify a one-dimensional axis through a medium ( the axonal web ) , along which the placement of detector neurons determines their inter-stimulus interval selectivity . However , sources and detectors can be arranged in two- or higher- dimensional media , such as the barrel cortex , to perform non-trivial computations . The barrel cortex , with the precise correspondence between the grid of cortical columns and the grid of whisker sensors , is an ideal structure in which to investigate the role of neural geometry in neural computation . The simplicity of the current model affords its explanatory power . However , a future study will be required to verify under what conditions the behaviour of the model is retained , when many hundreds of neurons and thousands of synaptic contacts are modelled explicitly . The barrel column is currently the target of a number of detailed modelling efforts [39] , [69]–[71] . Complementing these approaches , the power of our simple geometric model to explain a series of complex observations suggests that the geometry of synaptic connections in and between barrel columns should be considered if we are to understand the function of cortical microcircuitry .
|
To perceive how stimuli move over sensor surfaces like the retina or the fingertips , neurons in the brain must report the relative timing of signals arriving at different locations on the sensor surface . The rat whisker system is ideal for exploring how the brain performs this computation , because the layout of a small number of sensors ( whiskers ) maps directly onto the layout of corresponding columns of neurons in the sensory cortex . Previous studies have found that neurons located between adjacent cortical columns are most likely to respond when the corresponding adjacent whiskers are stimulated in rapid succession . These results suggest a link between the location of the neuron and the relative timing of sensory signals reported by its activity . We hypothesized that , if the time taken for whisker signals to arrive at a neuron is related to its distance from each cortical column , then neurons closer to a particular column will report stimuli moving towards that particular whisker . In a model approximating the geometry of cortical connections , responses of artificial neurons matched those of real neurons on a wide range of details . These results suggest an important role for neural geometry in neural computation .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"circuit",
"models",
"auditory",
"system",
"computational",
"neuroscience",
"biology",
"computational",
"biology",
"sensory",
"systems",
"neuroscience",
"sensory",
"perception",
"coding",
"mechanisms"
] |
2011
|
Neural Computation via Neural Geometry: A Place Code for Inter-whisker Timing in the Barrel Cortex?
|
Inflammation plays an important role in the pathophysiology of Chagas disease , caused by Trypanosoma cruzi . Prostanoids are regulators of homeostasis and inflammation and are produced mainly by myeloid cells , being cyclooxygenases , COX-1 and COX-2 , the key enzymes in their biosynthesis from arachidonic acid ( AA ) . Here , we have investigated the expression of enzymes involved in AA metabolism during T . cruzi infection . Our results show an increase in the expression of several of these enzymes in acute T . cruzi infected heart . Interestingly , COX-2 was expressed by CD68+ myeloid heart-infiltrating cells . In addition , infiltrating myeloid CD11b+Ly6G- cells purified from infected heart tissue express COX-2 and produce prostaglandin E2 ( PGE2 ) ex vivo . T . cruzi infections in COX-2 or PGE2-dependent prostaglandin receptor EP-2 deficient mice indicate that both , COX-2 and EP-2 signaling contribute significantly to the heart leukocyte infiltration and to the release of chemokines and inflammatory cytokines in the heart of T . cruzi infected mice . In conclusion , COX-2 plays a detrimental role in acute Chagas disease myocarditis and points to COX-2 as a potential target for immune intervention .
Chagas disease is a multisystemic disorder caused by Trypanosoma cruzi infection that affects more than 8 million people worldwide , being endemic in Latin America . Due to the scarcity of preventive and therapeutic tools and population at risk , it is considered as a neglected tropical disease [1 , 2] . More than 40 , 000 new infected people and 12 , 550 deaths per year are estimated . The high rate of migration towards non-endemic countries has spread the boundaries of the infection to other continents . Non-vectorial transmission is possible through oral ingestion , blood transfusion , organ transplantation and during pregnancy . The risk of infection is related to the country of origin of the migrants and the rate of prevalence in a given country [3] . Chagas disease is characterized by acute and chronic phases . Death occurs occasionally in the acute phase ( <5–10% of symptomatic cases ) as a result of severe myocarditis , meningoencephalitis , or both . The experimental model of infection in mice recapitulates many clinical features observed in human infection , although different strains of mice and parasites produce different disease outcomes [4] . Heart inflammation during the acute phase of T . cruzi experimental infection is initiated by lymphoid and myeloid mononuclear cell infiltration [5] . We have isolated from infected hearts an infiltrating monocytic CD11b+Ly6C+Ly6G- population expressing both classical ( M1 ) and alternatively ( M2 ) activated macrophage markers that is able to suppress T cell proliferation ex vivo , characteristics that define them as myeloid-derived suppressor cells ( MDSCs ) [6 , 7] . Myeloid cells are thought to be the major source of prostanoids , end products of cell membrane arachidonic acid ( AA ) catabolism , which include prostaglandins , prostacyclin and thromboxanes [8] . Enzymes implicated in prostanoid production have been investigated for many years [9] . All these lipid mediators have important roles in homeostasis and immune response regulation [10] . Cyclooxygenases , COX-1 and COX-2 , are principal enzymes in prostanoid production . COX-1 expression is involved in homeostasis while COX-2 is induced by several factors , including infection [9] . However , the specific role of COX-2 and downstream enzymes in the context of infection varies depending on the infectious agent [11–13] . PGE2 , a product of terminal PGE2 synthases ( PGES ) has pro-inflammatory properties [14] but also immunosuppressive properties [15] by signaling through G-protein coupled PGE receptors ( EP ) , termed EP-1 , EP-2 , EP-3 and EP-4 . PGE2 also decreases the ability of macrophages to phagocytize and kill microorganisms [16 , 17] , and is required for monocyte migration in response to chemokines [18 , 19] . There are few studies about the role of prostanoids in human chagasic pathology [20 , 21] , but it has been described that monocyte inflammatory mediators inhibit cellular proliferation and enhance cytokine production in patients [22] . In rodent models of acute infection , the levels of PGF2α , TXB2 , 6-oxo-PGF1α [23] and PGE2 [24] in plasma , were increased . Macrophages from infected rats show an increased number of lipid bodies , where COX-2 produces PGE2 [25] . Recently , it has been shown that the absence of Phospholipase A2γ , an enzyme implicated in AA release from membranes , decreases mice survival [26] . The role of COX in mice infected with T . cruzi has been studied using non-selective inhibitors of COX-1 and COX-2 , as well as COX-2-selective inhibitors ( NSAIDs ) , with conflicting results . Thus , it has been described that COX inhibitors cause an increase in mortality and parasitism [27] in T . cruzi infection , but contrarily , other reports claim that COX-2 inhibition decreases the level of parasitism [28 , 29] . In addition , both beneficial and adverse effects of COX inhibitors have been reported , depending on the phase of T . cruzi infection and the mice strain used [30] . Discrepancies between these studies could be explained by the different ability of BALB/c and C57BL/6 mouse strains to produce PGE2 [31]; the presence of distinct levels of cytokines in serum [32] or because of differences in cardiac cytokine expression profile [6] . Thus , in order to clarify the role of prostanoids in acute cardiac inflammation , we infected susceptible and non-susceptible mice , as well as COX-2 and EP-2 deficient mice with T . cruzi and analyzed cardiac inflammation , leukocyte infiltration and expression of cytokines , chemokines and inflammatory mediators in the infected mice .
We infected mice with the Y strain of T . cruzi . Immunopathology caused by this parasite strain is characterized by cardiac inflammatory damage . As we previously reported [33] , C57BL/6 , but not BALB/c , infected mice recovered from infection and survived ( S1A Fig ) . Parasitemia was detectable between 9 and 21 d . p . i . and was higher in BALB/c mice ( S1B Fig ) . Hearts of BALB/c infected mice showed more leukocyte infiltration and parasite nests than C57BL/6 mice ( S1C Fig ) . We next studied the expression level of enzymes involved in the AA pathway in the heart of both strains after infection ( Fig 1A ) . COX-2 gene expression ( Ptgs2 ) , but not COX-1 ( Ptgs1 ) , was increased in heart tissue during the acute phase of T . cruzi infection similarly in both mouse strains . Ptges ( microsomal prostaglandin E2 synthase , mPGES-1 ) , Hpgds ( leucocyte type PGD synthase ) and Tbxas1 ( thromboxane synthase ) mRNA expression levels were also incremented . However , Ptgds ( lipocalin-type prostaglandin D synthase ) mRNA basal level of expression in heart tissue did not change upon T . cruzi infection . These results indicate that T . cruzi infection promoted the selective up-regulation of some of the enzymes involved in prostanoid production in heart tissue , including COX-2 and mPGES-1 . Since COX-2 plays a key role in the synthesis of PGs in inflammatory processes , we aimed to identify the cells expressing this enzyme in the heart of T cruzi infected mice . Hearts from C57BL/6 ( Fig 1B ) and BALB/c ( Fig 1C ) mice were immunostained for COX-2 and myeloid and lymphoid markers , and imaged by confocal microscopy . Cells expressing COX-2 were abundant in the infected hearts of both mice strains , showing a strong staining in the perinuclear region of both myeloid CD68 positive and non-myeloid CD68 negative infiltrating cells . Interestingly , there was no co-localization of COX-2 and Arg-1 , a marker of M2 macrophages and MDSCs ( Fig 1D ) . Although COX-2 expression by activated lymphocytes has been previously described [34] , CD4 staining was not detected in infiltrating COX-2+ cells in hearts of infected C57BL/6 ( S2A Fig ) nor BALB/c mice ( S2C Fig ) . No staining was observed in negative control sections incubated with secondary antibodies alone ( S2B , S2D and S2E Fig ) . We next isolated myeloid cells from hearts of T . cruzi infected C57BL/6 and BALB/c mice at the times , 14 and 21 d . p . i . respectively , when maximum Arg-1 and inducible nitric oxide synthase ( iNOS ) expression is observed [7] . Using anti-Ly6G antibody labeled magnetic microbeads we obtained the Ly6G+ population . CD11b+ cells were selected from the remaining Ly6G- population , ( Fig 2A ) . As previously described [7] , the CD11b+Ly6G- cell population expressed iNOS and Arg-1 , and here we show that they also expressed COX-2 ( Fig 2A ) . Interestingly , COX-2 gene expression was much higher in CD11b+ cells obtained from infected cardiac tissue than those from the blood ( Fig 2B ) , pointing to infiltrating myeloid cells in inflamed tissue as the source of COX-2 . In agreement with the increase in COX-2 expression , a significant increase in the production of prostanoids as PGE2 and 6-oxo PGF1α in infected hearts was detected by mass spectrometry analysis on total heart extracts ( Fig 2C ) . Further ex vivo analysis on Ly6G+ and CD11b+Ly6G- purified heart infiltrating myeloid cells cultured in the presence of radiolabeled AA , showed that Ly6G+ cells did not produce any detectable prostanoid ( Fig 2D ) . In contrast , CD11b+Ly6G- cells produced high levels of PGE2 and low amounts of PGF2α ( Fig 2E ) . These results indicate that the CD11b+Ly6G- myeloid population is able to synthesize high levels of PGE2 from AA , while other cell types in heart tissue are likely producing PGE2 and 6-oxo PGF1α . In order to study the role of COX-2 in the development of cardiac leukocyte infiltration , we infected COX-2+/+ and COX-2-/- mice with the Y strain of the parasite . COX-2-/- mice showed 30% reduction in blood parasite number compared to COX-2+/+ mice at the peak of parasitemia ( Fig 3A ) . However , COX-2 deficiency did not significantly affect cardiac parasite burden compared to COX-2+/+ infected mice ( Fig 3B ) . No mortality was observed neither in COX-2+/+ nor in COX-2-/- infected mice up to 42 d . p . i . Inflammatory infiltrates were analyzed and quantified in T . cruzi infected hearts from COX-2+/+ and COX-2-/- mice . Fig 3C shows the extent of leukocyte infiltration calculated from several tissue sections as described in Methods . We observed significant less inflammatory infiltration in infected COX-2-/- than in COX-2+/+ mice . Representative images corresponding to the quantification of cell infiltration are shown in Fig 3D . We next analyzed the cellular composition of the immune inflammatory infiltrate by determining gene expression of surface markers characteristic of various immune cell populations by qRT-PCR and normalizing the data from infected animals respect to non-infected controls . In agreement with histological findings , infection in COX-2-/- mice compared to COX-2+/+ mice , led to lower expression of the common leukocyte marker Ptprc ( CD45 ) as well as of Cd4 , Cd8 , Cd68 , and Itgax ( CD11c ) as markers of T helper cells , cytotoxic T cells , macrophages and dendritic cells , respectively ( Fig 4A ) . To characterize the immune response in hearts of COX-2-/- infected mice , gene expression of chemokines and cytokines were analyzed . mRNA levels of chemokines ( Ccl2 , Ccl5 and Cxcl9 ) and cytokines ( Ifng , Tnf , Il4 , Il6 and Il10 ) were significantly increased during T . cruzi infection in hearts of both COX-2+/+ and COX-2-/- mice ( Fig 4B and 4C ) . However , chemokine expression presented different patterns in COX-2+/+ and COX-2-/- mice . Ccl2 and Ccl5 expression , but not Cxcl9 , was significantly higher in COX-2+/+ mice than in COX-2-/- mice ( Fig 4B ) . Induction of pro-inflammatory cytokines Ifng , Tnf and Il6 was lower , whereas Il4 expression was higher , in COX-2-/- compared to COX-2+/+ mice ( Fig 4C ) . The anti-inflammatory cytokine Il10 showed lower expression in the COX-2-/- infected mice . There were no significant differences in Arg1 expression between COX-2+/+ or COX-2-/- mice ( Fig 4D ) , but induction of Nos2 mRNA ( iNOS ) was significantly lower in COX-2-/- infected mice ( Fig 4D ) . There was no induction of Ptgs1 ( COX-1 ) expression that could compensate for the COX-2-/- deficiency ( Fig 4D ) . Ptges ( mPGES-1 ) expression was increased upon infection , with lower levels in heart tissue from COX-2-/- mice compared to COX-2+/+ mice ( Fig 4D ) . Nevertheless , protein analysis by western blot showed lower expression of both iNOS and Arg-1 in infected COX-2-/- respect to COX-2+/+ mice ( Fig 5 ) . Analysis of TNF-α levels in plasma showed a similar increase in both COX-2+/+ and COX-2-/- infected mice , indicating that the effect of COX-2 deficiency is not systemic but specific of the heart ( S3A Fig ) . Basal levels of gene expression did not significantly change between COX-2+/+ and COX-2-/- mice ( S4 Fig ) . We found increased levels of PGE2 in heart tissue and cardiac infiltrating cells after T . cruzi infection . Since the effector function of PGE2 produced by myeloid cells depends on its binding to EP receptors , we studied gene expression of its 4 receptors , EP-1 ( Ptger1 ) , EP-2 ( Ptger2 ) , EP-3 ( Ptger3 ) and EP-4 ( Ptger4 ) , in hearts of mice during infection . The results show that in control infected mice the overall expression of EP receptors is higher than in non-infected hearts , except for Ptger3 ( Fig 6A ) . However , in C57BL/6 infected hearts Ptger2 expression showed the highest increases suggesting a potential role of this receptor in T . cruzi infection . Thus , we used mice deficient in the expression of the EP-2 ( in the C57BL/6 background ) , which has been involved in inflammation [35] , and also in an autocrine loop of macrophage activation by PGE2 [36] , in order to study the role of this receptor during T . cruzi infection . The results show that EP-2+/+ and EP-2-/- mice survived infection and no significant differences in parasitemia or in heart parasite burden were observed between them ( Fig 6B and 6C ) . These results suggest that EP-2 signaling does not play an essential role in parasite elimination . However , significant less heart inflammatory infiltrates were observed in infected EP-2-/- in comparison with EP-2+/+ mice at 14 d . p . i . ( Fig 6D ) . Representative images of cardiac tissue and inflammatory infiltration are shown in Fig 6E . The expression of the common leukocyte marker Ptprc ( CD45 ) was lower in the heart of infected EP-2-/- mice respect to EP-2+/+ , whereas mRNA levels of cell markers as Cd4 ( Th cells ) , Cd8 ( Tc cells ) , and Itgax- ( CD11c; DCs ) , did not show significant differences . However , the expression of Cd68 , a macrophage marker , significantly increased in EP-2-/- respect to EP-2+/+ mice ( Fig 7A ) . Regarding chemokines , Ccl2 expression , but not Ccl5 and Cxcl9 , was significantly reduced in the EP-2-/- compared to EP-2+/+ infected mice ( Fig 7B ) . Induction of pro-inflammatory cytokines Ifng and Il6 , the Th2 cytokine Il4 and the anti-inflammatory cytokine Il10 , but not Tnf , was lower in EP-2-/- compared with EP-2+/+ ( Fig 7C ) . Similarly to COX-2-/- , no differences were observed in TNF-α plasma levels in EP-2-/- as compared to EP-2+/+ infected mice ( S3B Fig ) . Ptgs2 ( COX-2 ) gene expression was significantly lower in EP-2-/- infected mice ( Fig 7D ) . There were no differences between mouse strains in Nos2 mRNA ( iNOS ) expression ( Fig 7D ) , but Arg1 mRNA expression was higher in EP-2-/- mice ( Fig 7D ) . Western blot analysis showed a significant increase in EP-2-/- respect to EP-2+/+ mice , in the protein levels of these enzymes involved in L-arginine metabolism ( Fig 8 ) . Basal levels of gene expression did not significantly change between EP-2+/+ and EP-2-/- mice ( S5 Fig ) .
BALB/c and C57BL/6 mice ( 6 to 8-week-old ) were purchased from Harlan , Interfauna Iberica . B6;129S7-Ptgs2tm1Jed/J ( COX-2-/- ) mice were purchased from The Jackson Laboratory . C57BL/6 Ptger2tm1Sna ( EP-2-/- ) mice were a gift form Dr . Shu Narumiya , ( Faculty of Medicine , University of Kyoto ) . Wild type B6/129S ( COX-2+/+ ) and C57BL/6 ( EP-2+/+ ) mice were obtained by breeding heterozygote pairs . In vivo infections were performed with Y T . cruzi strain as described [6 , 7] . Groups of 3–15 mice were infected with 2 , 000 trypomastigotes per mice by intraperitoneal injection . Groups of 3–6 non-infected mice were included in the experiments as a control . Survival was monitored daily and parasitemia levels were checked every 2–3 days . Mice blood and tissues were collected at 0 ( non-infected ) , 14 and 21 days post-infection ( d . p . i . ) , as indicated . This study was carried out in strict accordance with the European Commission legislation for the protection of animals used for scientific purposes ( Directives 86/609/EEC and 2010/63/EU ) . Mice were maintained under pathogen-free conditions at the Centro de Biología Molecular Severo Ochoa ( CSIC-UAM ) animal facility . The protocol for the treatment of the animals was approved by the ‘‘Comité de Ética de Investigación de la Universidad Autónoma de Madrid” , Spain ( permits CEI-14-283 and CEI-47-899 ) . Animals had unlimited access to food and water . They were euthanized in a CO2 chamber and all efforts were made to minimize their suffering . Hearts were perfused with Phosphate buffered saline ( PBS ) solution containing 1UI/ml of heparin , minced into small pieces with a sterile scalpel and DNA was isolated with High PurePCR Template preparation Kit ( Roche ) . For T . cruzi detection , we used the quantitative PCR ( qPCR ) assay described by Piron et al . [37] . 100 , 10 , 1 , 0 . 1 and 0 . 01 pg of DNA purified from Y strain epimastigotes were used to generate the standard curve . qPCR reactions were performed with 100 ng of genomic DNA and murine Tnf gene primers were used as DNA loading control . For RNA extraction , heart tissue was perfused with PBS containing 1UI/ml of heparin , cut in small pieces using a sterile scalpel blade , followed by mechanical disruption using a PT 1300 D homogenizer ( Kinematica Polytron , Fisher Scientific ) in TRIzol reagent ( Invitrogen ) as indicated by the manufacturer . Gene expression was analyzed by quantitative reverse transcription PCR ( qRT-PCR ) . Reverse transcription of total RNA was performed using the components of the High Capacity cDNA Archive Kit ( Applied Biosystems . Life Sciences ) or the SuperScript Enzyme ( Invitrogen , Life Sciences ) . Amplification were performed using TaqMan MGB probes ( S1 Table ) and the TaqMan Universal PCR Master Mix ( Applied Biosystems ) on an ABI PRISM 7900 HT instrument ( Applied Biosystems . Life Sciences ) . For cultured cells , samples were treated as mentioned above except for the mechanical disruption . All samples were assayed in triplicate . Quantification of gene expression by real-time PCR was calculated by the comparative threshold cycle ( CT ) method as described in [38] ( RQ = 2-ΔΔCT ) . All quantifications were normalized to the ribosomal 18S control to account for the variability in the initial concentration of RNA and in the conversion efficiency of the reverse transcription reaction . Protein extracts were prepared from heart tissue perfused with PBS containing 1UI/ml of heparin , cut in small pieces using a sterile scalpel blade followed by mechanical disruption in Triton X-100 based protein lysis buffer . Protein concentration was determined by the bicinchoninic acid method ( BCA , Pierce ) using Bovine Serum Albumin ( BSA ) for the standard curve . Western blot analyses were performed as follows: 15 or 50 μg of cell or tissue extract were fractionated on SDS polyacrylamide gel and transferred to a Nitrocellulose membrane Hybond-ECL ( Amersham Biosciences ) and blocked in 5% fat free dry milk or 5% BSA in 0 , 1% Tween-20 Tris Buffered Saline . Membranes were incubated overnight with diluted primary antibodies ( S2 Table ) at 4–8°C . The membranes were incubated with horseradish peroxidase conjugated secondary antibodies ( S2 Table ) and detection was carried out with Supersignal detection reagent ( Pierce ) followed by photographic film exposure . Fiji package software was used to quantify band intensity normalizing band areas of the sample to their respective loading control . Cardiac tissues from mice were placed in 10% neutral buffered formalin for at least 4 hours at room temperature followed by incubation in 70% ethanol overnight . Samples were then embedded in paraffin ( Tissue Embedding Station Leica EG1160 ) , and 5 μm tissue sections were prepared ( Microtome Leica RM2155 ) . Samples were deparaffinized and rehydrated using a Tissue Processing Station Leica TP1020 . Slides were stained with Masson´s Trichrome staining and mounted permanently in Eukitt´s quick hardening mounting medium ( Biochemika , Fluka analytical ) . The sections were microscopically analyzed in a Leica microscope ( DMD 108 , Leica microsystems Wetzlar GmbH , Germany ) using the 20x magnification objective lens and Lamp intensity 10 and f/Stop 12 . Ventricular myocardium micrographs were taken avoiding pericardium , endocardium , atria and big vessels . Nine pictures of different sections , separated by at least 50 μm , per heart were taken . The degree of inflammatory-cell infiltration was quantified using the Fiji package [39] and the plugin Trainable Weka Segmentation developed by Ignacio Arganda Carreras ( Versailles , France ) [39] ( Image J macro used for automated image analysis is detailed in S1 File ) and expressed as the percentage of the nuclei/tissue area ratio . Organs were removed from mice at different d . p . i . , cut and fixed in a 4% paraformaldehyde PBS buffered solution for 2h at room temperature , followed by incubation in a 30% sucrose solution at 4°C overnight . Tissues were then embedded in Tissue-Tek OCT in Cryomolds ( Sakura ) and frozen . 10 μm sections were cut using a cryostat Leica CM1900 . Slides were fixed in acetone for 10 min at room temperature and incubated 10 min with NH4Cl to reduce autofluorescence . Then , slides were washed with PBS , permeabilized with 0 , 1% Triton X-100 , blocked and incubated over night at 4°C with primary antibodies ( S2 Table ) in blocking buffer ( PBS 0 , 1% Triton X-100 , 5% BSA ) . The samples were washed with PBS and secondary antibodies ( S2 Table ) were added in blocking buffer and incubated overnight at 4°C . Blocking of unspecific secondary antibody binding was achieved by addition of 2% of normal serum of the species in which the secondary antibody was raised . As a negative control , sections were treated in the same manner , except that incubation with primary antibody was omitted . Nuclei were stained using 1 μg/ml of DAPI ( 268298 , Merck ) . Prolong Gold Antifade Reagent ( Invitrogen ) was used to mount the slides that were kept at 4°C until observation . Stained slides were observed with the confocal laser scanning microscope LSM710 , coupled to an AxioimagerM2 microscope ( Zeiss ) . The micrographs were processed using the software ZEN ( Zeiss ) or the Fiji Package . BALB/c ( n = 15 ) or C57BL/6 mice ( n = 15 ) were infected i . p . with 2 , 000 trypomastigotes of the Y strain . At 21 d . p . i . for BALB/c and 14 d . p . i . for C57BL/6 , mice were euthanized in a CO2 chamber and hearts were aseptically removed , perfused with 10 ml PBS containing 1UI/ml of heparin , and kept in cold Hank´s balanced saline solution ( HBSS ) . Then , hearts were pooled in a cell culture dish , washed thoroughly with HBSS and minced into small pieces with a sterile surgery blade . Mouse hearts ( maximum 4 per tube ) were transferred into the gentleMACS C tube containing 4 , 7 mL of HBSS . 300μL of Collagenase II solution ( 600 U/ml ) and 10 μl DNase I solution ( 60U/ml ) were added . Then tissue was disrupted with GentleMACS Dissociator ( Miltenyi Biotec GMbH ) . To obtain cell suspensions , a 70 μm cell strainer ( Falcon BD ) was used . After red blood cells lysis , the cells were magnetically sorted . For Ly6G+ cell sorting , anti Ly-6G MicroBead kit was used with MACS LS columns and MACS Separators ( Miltenyi Biotec GmbH ) , following manufacturer instructions . Ly6G- fraction of the cell suspension was afterwards processed for CD11b+ cell sorting using CD11b Microbeads kit . Prostanoid levels were determined in mouse tissue extracts from 0 ( non-infected ) and 21 d . p . i . by Metabolon Inc . , and expressed as ScaledImpData as previously described [40] . To determine in vitro prostanoid production , heart infiltrating cells were magnetically sorted as described above and incubated 30 minutes at 37°C in 500 μl of RPMI without Fetal bovine serum ( FBS ) in the presence of 25 μM [14C] AA , PerkinElmer ( Massachusetts , USA ) . 500 μl of 2% acetic acid in cold methanol was added to extract and preserve AA derivatives . Samples were vortexed and the air inside the tube was substituted by inert nitrogen gas . Samples were kept frozen at -80°C until HPLC was performed . HPLC device was composed by a Beckman Solvent Module 126 with the column Ultraphere ODS ( C-18 , Beckman-Coulter ) 5 μm diameter sphere particle , 4 . 5 mm and 25 cm column diameter and length respectively and a Beckman 171 Radioisotope Detector . Scintillation liquid Ecoscint H was purchased from National Diagnostics . Prostanoids were resolved with the isocratic flow ( 1ml/min ) of the mobile phase: Acetonitrile/water/acetic acid 33:67:0 . 1 v/v/v . Standards were produced using [14C] Arachidonic Acid and different cell types expressing the respective enzymes , and [14C] Arachidonic Acid incubated in medium was used as input control as described [41] . For in vivo experiments , data are shown as means ± SEM . All the in vitro experiments were performed at least three times . Significance was evaluated by Student’s t-test when two groups were compared . ANOVA one way followed by Tukey post-test was used when groups of samples from an experiment had different time points . ANOVA two way followed by Bonferroni post-test were used when the experiment included time and mice strain as variables . For survival analysis , we used Gehan-Breslow-Wilcoxon method . GraphPad Prism 5 . 00 software ( La Jolla , CA , USA ) was used for statistical analysis .
In order to clarify the role of prostanoids in the outcome of T . cruzi infection we first analyzed the expression of prostanoid-synthesizing enzymes in cardiac tissue from T . cruzi susceptible ( BALB/c ) and non-susceptible ( C57BL/6 ) mice . Our results showed an increase of COX-2/mPGES-1/PGE2 axis in heart tissue upon infection in both strains of mice , indicating that it has no direct effect on susceptibility to infection . Confocal microscopy analysis showed the presence of CD68+Arg-1+COX-2- cells and CD68+Arg-1-COX-2+ , suggesting that there are at least two subpopulations of monocytic infiltrating cells with mutually exclusive expression of those enzymes . Thus , our results show that the myeloid population infiltrating the heart in T . cruzi infection is more complex than previously described [7] , and suggests a difference in the function of these two myeloid populations . Macrophages can rapidly change their phenotype and function in response to local microenvironmental signals , playing key roles in the initiation and resolution of inflammation and tissue homeostasis [42] and could be involved in tissue repair and fibrosis [43] . Thus , myeloid cardiac infiltration could inhibit parasite replication and also facilitate the repair of damaged muscular tissue [44] . A suggestive hypothesis is that COX-2 expressing macrophages could be linked to inflammation meanwhile Arg-1+ macrophages could be involved in tissue repair . Immunostaining of heart tissue sections showed the presence of myeloid and non-myeloid cells positive for COX-2 in heart tissue sections of both BALB/c and C57BL/6 mice . However , after purification of myeloid cells from heart tissue , we found that only a particular subset expressed COX-2 , being the levels of COX-2 expression higher in C57BL/6 than BALB/c mice . Likely , non-myeloid COX-2 positive cells are lost in the purification process , a fact that might account for the apparent contrary results . Interestingly , we demonstrated that PGE2 and 6-oxo-PGF1α ( stable hydrolysis product of PGI2 ) were elevated in infected heart tissue . Furthermore , monocytes ( CD11b+Ly6G- ) isolated from infected heart express COX-2 , and are able to produce high levels of PGE2 ex vivo . The differences in metabolites detected in total heart extracts versus purified myeloid cells are likely due to their synthesis by other cell types and/or enrichment after cell purification . In agreement with this , we have previously reported that T . cruzi infection induces COX-2 in cardiomyocytes , leading to PGF2α and TXA2 production [45] . Although COX-2 expression can be induced in CD4+ T cells upon activation [34] , heart infiltrating CD4+ cells did not express detectable levels of COX-2 . On the other hand , the increase of Tbxas1 and Hpgds gene expression observed upon T . cruzi infection in both C57BL/6 and BALB/c mice suggests the production of their respective TXA2 and PGD2 metabolites in infected cardiac tissue . Tbxas1 was elevated up to 28 d . p . i . and its product , TXA2 , besides its vascular functions [46] , could have a pro-inflammatory role for monocytes [47] . In contrast , Hpgds expression showed a gradual increase during the acute phase , and its product , PGD2 , could be involved in resolution of inflammation , as described in other settings [48–50] . Resistant C57BL/6 mice showed significantly higher expression of Hpgds at 21 d . p . i . and lower expression of Tbxas1 at 28 d . p . i . than the susceptible BALB/c mice , suggesting that C57BL/6 may resolve inflammation earlier than BALB/c infected mice . However , TXA2 and PGD2 metabolites were not detected in purified Ly6G+ nor in CD11b+Ly6G- cells , suggesting that they could be produced by other infiltrating cell types or by infected cardiomyocytes [45] . Previous reports using COX-2 inhibitors in T . cruzi infection showed discordant results [30] . Moreover , COX-2 inhibitors may interfere with the immune response [34] , but more importantly , many COX-2 inhibitors have effects independent of their ability to inhibit cyclooxygenase activity [51–53] . For those reasons , we analyzed the contribution of COX-2 by using a mouse model deficient for its expression . We found a small variation in parasitemia ( 30% reduction at the peak of parasitemia ) in COX-2-/- respect to the COX-2+/+ mice , which cannot be taken as indicative of resistance . In addition , no changes were observed in cardiac parasite burden , in spite that COX-2-/- mice expressed less iNOS than COX-2+/+ mice , considered to be key for resistance in T . cruzi infection [54] , indicating that heart parasite load is not affected by the lack of COX-2 expression . Interestingly , we found that COX-2 was required for leukocyte infiltration and inflammation in the heart upon T . cruzi infection , but it did not affect systemic inflammation . However , COX-2-/- mice are resistant to death in sepsis , indicating that in this case COX-2 has a systemic pro-inflammatory role [55] . An important role of endogenous COX-2 derived PGs in migration of immune cells to infected tissues or lymphoid organs is becoming evident [19] . Thus , the decrease in cardiac inflammation and in local production of cytokines and chemokines observed in COX-2-/- infected mice , indicate a pro-inflammatory role of prostanoids as PGE2 in acute myocarditis . We have previously described that PGE2 induces COX-2 and mPGES-1 expression in an autocrine loop required for full activation of macrophages [36] . Thus , the fact that COX-2-/- mice showed a reduced Ptges mRNA expression ( mPGES-1 ) suggest a blockade of the autocrine loop that may impair full activation of macrophages , resulting in the reduced cardiac infiltration observed . In the same direction , lack of PGE2 signaling through EP-2 receptor in EP-2-/- infected mice , resulted in reduced Ptgs ( COX-2 ) mRNA expression that may also block the autocrine loop , impairing macrophages to infiltrate the cardiac tissue . This is in agreement with the observed decrease in Ccl2 mRNA expression in COX-2-/- and EP-2-/- infected mice respect to wild type infected mice , since this chemokine is required for migration of monocytes to the inflamed infected tissue [56] . Moreover , PGE2 also affects migration of myeloid cells potentiating CCL2 activity [19] . Therefore , the reduction of cardiac infiltration in both animal models suggest a detrimental pro-inflammatory role of COX-2 in the onset of cardiac inflammation . Strikingly , Arg-1 and iNOS expression , markers of MDSCs , was higher in hearts of infected EP-2-/- mice than in those from COX-2-/- infected mice hearts , indicating that infiltrating cells from EP-2-/- mice present a more marked MDSCs phenotype . In addition , the effect of EP-2 deficiency on cytokine and chemokine production in heart , was milder than the observed in COX-2-/- infected mice . In the hearts of C57BL/6 mice , infection caused a greater significant increase in the expression of Ptger2 , which validates the use of EP-2-/- mice in the C57BL/6 background . But still Ptger1 and Ptger4 were significantly increased although in a minor extent . Thus in EP-2-/- infected mice PGE2 can still signal through Ptger1 and Ptger4 causing this milder effect in EP-2-/- mice in comparison to the observed in COX-2-/- infected hearts . In contrast , in COX-2-/- infected mice , mPGES1 synthase expression is substantially reduced and PGH2 substrate for PGE2 production likely relies on constitutive COX-1 activity . Thus , the decreased levels of PGE2 may affect signaling through all PGE2 receptors , having a broader effect on leukocyte infiltration . The response to T . cruzi infection in mice deficient in other enzymes and products of the AA pathway has been scarcely studied [26 , 27] . Sharma et al . described that deficiency of iPLA2-γ ( Ca++ independent PLA2 isoform-γ ) , which is involved in AA membrane release , aggravated infection and decreased survival , while Mukherjee et al . described that COX-1-/- mice showed higher parasitemia than wild type infected mice , but no difference in survival was noted [27] . In our hands , interference within the AA pathway at a different level , as COX-2 mediated production of prostanoids or PGE2/EP-2 signaling , results in decreased inflammation in heart of T . cruzi infected mice , with low incidence in parasite burden and survival . Altogether , these results point to an essential role of the AA pathway in heart inflammation during T . cruzi infection . In the other hand , related with the prostanoid pathway , 5-lipoxygenase ( 5-LO ) has been shown to play a detrimental role during T . cruzi infection by potentiating heart parasitism and inflammation [57 , 58] through the regulation of iNOS activity [59] . However , further studies are needed to elucidate the crosstalk between LO and COX pathways during infection . ” Besides , we have previously showed that monocytic CD11b+Ly6G- heart infiltrating cells ( MDSCs ) , expressing iNOS and Arg-1 suppressed ex vivo T cell proliferation [7] . Since some of these infiltrating cells also express COX-2 and produce PGE2 it is possible that COX-2-derived PGs could contribute to immune suppression , a possibility that should be addressed in the future . In conclusion , during acute T . cruzi infection there is an increase in the expression of many enzymes of the AA metabolism , including COX-2 and mPGES-1 that leads to an increase in their metabolite PGE2 , partially due to infiltrating myeloid cells in the heart . Besides , we have identified a new myeloid infiltrating population characterized by the expression of COX-2 . Thus , so far there are at least three different myeloid populations infiltrating the T . cruzi infected heart: granulocytes , monocytic MDSCs expressing iNOS and Arg-1 and monocytic cells expressing COX-2 . COX-2 activity likely increases PGE2 levels in heart tissue , which play a pro-inflammatory role by signaling through EP-2 . However , the phenotype of EP-2-/- is not as strong as COX-2-/- infected mice probably due to PGE2 signaling through alternative EP receptors . Our findings suggest that COX-2 plays a detrimental role in acute Chagas disease myocarditis . Further research of the AA pathway is needed to completely understand its role during T . cruzi infection for immune intervention approaches .
|
The role of prostanoids , products of the arachidonic acid pathway , during Trypanosoma cruzi infection has been studied by inhibiting key enzymes in prostanoid synthesis as cyclooxygenases ( COX-1 and COX-2 ) , with opposed results . Here we analyzed the expression of cyclooxygenases , prostanoid synthases and receptors in the heart of mice susceptible and non-susceptible to T . cruzi infection and found that they were highly increased respect to non-infected mice . We previously identified the presence of myeloid-derived suppressor cells expressing arginase-1 ( Arg-1 ) . Further analysis showed that COX-2 was expressed in Arg-1- myeloid cells in heart tissue , suggesting the existence of different myeloid populations involved in the leukocyte infiltration ( COX-2+Arg-1- ) and tissue repair ( COX-2-Arg-1+ ) . Mice deficient in the expression of COX-2 and the prostaglandin PGE2 receptor EP-2 infected with T . cruzi showed a marked reduction in the cardiac inflammatory infiltration in comparison with infected wild type mice , indicating an adverse effect of COX-2 and PGE2 signaling through EP-2 receptor in the development of myocarditis during acute T . cruzi infection , suggesting the possibility of immune intervention using COX inhibitors .
|
[
"Abstract",
"Introduction",
"Results",
"Methods",
"Discussion"
] |
[] |
2015
|
Cyclooxygenase-2 and Prostaglandin E2 Signaling through Prostaglandin Receptor EP-2 Favor the Development of Myocarditis during Acute Trypanosoma cruzi Infection
|
Whether unique to humans or not , consciousness is a central aspect of our experience of the world . The neural fingerprint of this experience , however , remains one of the least understood aspects of the human brain . In this paper we employ graph-theoretic measures and support vector machine classification to assess , in 12 healthy volunteers , the dynamic reconfiguration of functional connectivity during wakefulness , propofol-induced sedation and loss of consciousness , and the recovery of wakefulness . Our main findings , based on resting-state fMRI , are three-fold . First , we find that propofol-induced anesthesia does not bear differently on long-range versus short-range connections . Second , our multi-stage design dissociated an initial phase of thalamo-cortical and cortico-cortical hyperconnectivity , present during sedation , from a phase of cortico-cortical hypoconnectivity , apparent during loss of consciousness . Finally , we show that while clustering is increased during loss of consciousness , as recently suggested , it also remains significantly elevated during wakefulness recovery . Conversely , the characteristic path length of brain networks ( i . e . , the average functional distance between any two regions of the brain ) appears significantly increased only during loss of consciousness , marking a decrease of global information-processing efficiency uniquely associated with unconsciousness . These findings suggest that propofol-induced loss of consciousness is mainly tied to cortico-cortical and not thalamo-cortical mechanisms , and that decreased efficiency of information flow is the main feature differentiating the conscious from the unconscious brain .
Despite the centrality of consciousness to our experience , no agreement has yet emerged on which aspects of brain function underlie its presence , and what changes are connected to its disappearance in the healthy brain ( e . g . , during sleep ) as well as in pathological conditions ( e . g . , coma ) . As a consequence , we are currently hard pressed to answer even basic questions concerning the presence , absence , degree and nature of the phenomenon of consciousness in humans and other species [1] . As experimental investigations into this domain have increased , a number of proposals have been put forth to characterize the neural fingerprint of consciousness . According to some views , the crucial feature underlying consciousness is the presence of specific patterns of activations , such as the presence of competing assembly of cells , or ‘neural coalitions’ [2] , synchronization of neural activity in specific frequency bands [3] , [4] , or the level of spontaneous oscillatory activity , at fast frequencies , in the thalamo-cortical system [5] . According to other proposals , consciousness is related to the spread and reverberation of information across the neural system , and in particular within specific regions in parietal and frontal cortices [6] , [7] – although the scope of this view is mostly relevant to the idea of conscious availability of content to a neural system , as compared to the more general “state of consciousness” of a neural system [8] . Finally , a recently proposed view [1] , [9] , stresses the importance of evaluating not the degree of correlation among different ( often long-range ) regions , but rather the degree of information present and the extent to which information is integrated across the nodes of a system . In the present work we look at spontaneous low-frequency fluctuations in the functional magnetic resonance imaging ( fMRI ) signal [10] , [11] , to assess the relationship between different states of consciousness and basic principles of information processing ( as captured by the blood oxygenation level dependent signal; i . e . , BOLD ) . The analysis of spontaneous fluctuations of the BOLD signal has been fruitfully employed to explore consciousness-related changes in clusters of temporally coherent regions during sedation [12] , [13] , sleep [14]–[16] , and in the pathological brain [17] , [18] . In particular , associations within specific networks of regions have been found to be monotonically modulated by consciousness [19]–[21] , consistent with some theoretical views [3]–[5] . This idea , however , clashes with reports of increased cross-regional correlation concurrent with decrease or loss of consciousness [22] , [23] , suggesting the importance of characterizing not just the strength but also the quality of information processing within a system [1] , [24] . Following this idea , we employ previously collected resting-state fMRI data [21] to assess , in 12 healthy volunteers , the dynamic change of governing principles of brain organization during wakefulness ( W ) , propofol-induced sedation ( S ) and loss of consciousness ( LOC ) , as well as after consciousness recovery ( R ) , a dynamic approach that has been recently advocated for [25] . In particular , we focus on the change of global and local topological metrics of information processing across conditions [26]–[28] , a technique that has been successfully employed to characterize and model dynamics within physical [29] , biological [30] and social systems [31] , and that has been shown to capture specific aspects of brain organization in the maturing , healthy adult , and pathological brain [32]–[36] . A particularly appealing aspect of this technique in the context of studies of consciousness is the parallel between the measures it offers , focused on characterizing how information is exchanged and propagated through a network , and theories of consciousness that stress the centrality of how information is treated and integrated within the brain [1] , [9] . As detailed below , we report three main findings . First , contrary to a recent report [25] , we find that long- and short-range connections are not differentially affected by sedation . Second , employing a support vector machine ( SVM ) classifier , we dissociate the thalamo-cortical and cortico-cortical hyperconnectivity observed during sedation from the cortico-cortical hypoconnectivity observed during loss of consciousness . Finally , contrary to results in other species [37] , we find significant global changes in the ( functional ) topological organization of the brain during sedation . However , we show that normalized clustering , the global metric that was previously reported to be sensitive to the loss of consciousness [25] , remains significantly elevated also through post-sedation recovery of wakefulness . Conversely , we find that a strong decrease in efficiency of information distribution ( defined as the inverse of the characteristic path length – see Materials and Methods ) is the only unambiguous marker of propofol-induced loss of consciousness .
The average connectivity matrices and the frequency distribution of ( average ) correlations for each condition are shown in Figure 1 and Figure 2a , respectively . According to a two-sample Kolmogorov-Smirnov goodness-of-fit test , the distribution of positive and negative correlations are significantly different for all pairwise comparisons ( ; ; ; ; all ) . In all four conditions about 80% of correlations were between 0 and 0 . 4 . LOC , however , exhibited a leftwards shift of the distribution , as shown by the median correlation value of 0 . 11 , as compared to 0 . 23 , 0 . 22 , and 0 . 19 for W , S , and R , respectively . Furthermore , 14% of correlations in the LOC condition were negative , as compared to about 2% in all other conditions , while only 6% were above 0 . 4 , versus 17% , 14% and 11% for W , S , and R , respectively . To assess whether correlations between areas at different distances were unequally affected by the level of consciousness , we employed a repeated measures ANCOVA with one within-subjects variable ( i . e . , condition ) with four levels ( W , S , LOC , R ) , and inter-ROI distance as a covariate ( with distance defined as the 3-dimensional Euclidean distance between the baricenter of each ROI; see Figure 2b ) to predict correlation strength . As expected , we found a significant effect of condition ( , ) , indicating that correlation strength systematically varied across conditions . Specifically , W consistently exhibited the strongest average correlation level , across all bins , followed by S and R , while LOC consistently exhibited the weakest average correlation across all bins . We also found a significant effect of distance ( , ) , indicating that , as shown in Figure 2b , the average correlation strength decreased with distance . In addition to the two main effects , we also found a significant interaction between condition and distance ( , ) , indicating an uneven effect of condition on links of different length . However , when we followed up this significant interaction with a set of separate repeated measures ANOVAs ( one per each bin ) we found that it was entirely driven by the absence of a significant difference between W and S for the first 3 bins ( out of 15; i . e . , regions closer than 3 . 4 cm ) . With this exception , the effect of propofol was remarkably consistent at all other connection lengths ( particularly with respect to the crucial condition – i . e . , loss of consciousness – where no difference was found across connection length ) . Indeed , at all other bins the four conditions were found to be significantly different from each other , based on estimated marginal means and a Sidak correction for multiple comparisons . The observation of a small effect of distance on connection strength across levels` of sedation is also consistent with the extremely low effect size observed for the interaction between condition and distance in the overall ANOVA ( ) , and strengthens the idea that , overall , connection size had a minimal effect on correlation strength – something that is immediately clear from Figure 2b . Results for the classification of brain networks ( i . e . , correlation matrices ) are reported in Table 1 and Figure 3 . At a global level , the SVM algorithm classified successfully states of wakefulness ( W & R ) versus states of sedation ( S & LOC ) with high accuracy , sensitivity and specificity ( all above 83 . 33%; ) . The same level of classification was also achieved when comparing contiguous brain states ( namely , W vs . S; S vs . LOC; and LOC vs . R; see Table 1 for a detailed report of accuracy , specificity , sensitivity and significance for each ) . Conversely , wakefulness ( W ) and wakefulness recovery ( R ) could not be successfully distinguished from each other ( ) . ( For completeness the two remaining classifications , namely W vs . LOC and S vs . R , are reported in Figure S1 . ) At the local level , accurate classification of each transition relied on different sets of edges within each brain graph ( see Tables S1 and S2 for full details ) . In particular , as depicted in Figure 3b , and more in detail in Figure 4a , the edges mostly contributing to correctly classifying S versus W included positive cortico-cortical ( 54 . 8% ) and thalamo-cortical ( 40 . 9% ) connections , as well as a minority of cerebello-cortical ( 0 . 5% ) and striato-cortical ( 3 . 8% ) connections . Conversely , as depicted in Figures 3c and 4b , the distribution of connections correctly classifying LOC , as compared to S , mostly included negative cortico-cortical connections ( 82 . 5% ) , as well as a minority of positive cortico-cortical ( 9 . 9% ) , thalamo-cortical ( 3 . 5% ) , cerebello-cortical ( 2 . 9% ) and thalamo-striatal ( 1 . 2% ) connections . Notably , when tested statistically , the allocation of classifying edges for these two transitions are significantly different ( , ) . Finally , as shown in Figures 3d and 4c , as compared to LOC , classification of R was almost entirely based on the re-emergence of positive cortico-cortical connections ( 98 . 4% ) as well as a small minority of cerebello-cortical connections ( 1 . 6% ) .
In this study we assessed propofol-induced changes in patterns of connectivity , as well as in global and local governing principles of brain organization , during wakefulness , sedation , loss of consciousness , and wakefulness recovery . Our results contribute to a growing literature addressing the topological organization of the human brain [26] , [38] , the changes in functional architecture accompanying the loss of consciousness [16] , [25] , [37] , as well as a specific hypothesis concerning the role of different subsystems in loss of consciousness [40] , [41] . Overall , our main findings are three-fold . First , despite the frequently voiced idea that long-range connections play a key role in anesthesia-induced unconsciousness [40] , we fail to find a substantial asymmetric decrease in cross-region correlation as a function of inter-regional distance . Average connectivity strength decreased monotonically with distance in approximately the same manner across conditions ( with the sole exception of extremely short connections , below 34 mm , but only during the initial phase of sedation , and not during loss of consciousness ) . This finding runs counter to a recent report demonstrating an uneven effect of propofol-induced unconsciousness on short-range ( i . e . , ) versus long-range ( i . e . , ) connections [25] . The only effect we detected concerned much shorter connections ( i . e . , ) , and was only found for the initial period of sedation , and not for the period of loss of consciousness . Whether the different result is to be attributed to methodological asymmetries ( e . g . , 2-timepoint versus 4-timepoint paradigms , the binning procedure , the use of different ROIs parcellation schemes ) or to un-modelled third factors remains to be determined . The second central aspect of our results directly addresses the discussion concerning the role of thalamo-cortical versus cortico-cortical circuits in propofol-induced unconsciousness [40] , [41] . In particular , our SVM classification isolated increased thalamo-cortical and cortico-cortical synchronization as being maximally informative in the wakefulness versus sedation classification , suggesting a prominent role of this circuit in the initial stages of sedation , before the onset of unconsciousness . Conversely , correct classification of the state of loss of consciousness , as compared to sedation , overwhelmingly relied on negative cortico-cortical correlations . These findings support the view that propofol-induced loss of consciousness is more closely linked to cortico-cortical mechanisms rather than thalamo-cortical ones , as also suggested in a recent EEG effective connectivity study [41] . It is important to point out that our SVM classification is entirely based on the full matrix of ROI-to-ROI correlations and is , therefore , entirely data driven and blind to the existence of particular neural circuits or opposing hypothesis concerning their role in propofol-induced loss of consciousness . The observed major role of negative cortico-cortical connectivity in propofol-induced unconsciousness should be differentiated , however , from studies on pathological loss of consciousness in severe brain injury where post-mortem [42] and in-vivo [43] evidence highlights the role of thalamus in loss and recovery of consciousness [44] , [45] . While further studies will have to directly address the issue , our findings are consistent with the suggestion that thalamus may be a necessary but not sufficient component in maintaining consciousness [41] consistent with the view that thalamic lesions might induce unconsciousness after severe brain injury by virtue of disconnecting an otherwise functioning cortex [46] , [47] . The third result of our study concerns changes in governing principles of information processing during loss and recovery of consciousness . Contrary to a recent study in other species [37] , we do find significant changes in global topological measures across levels of consciousness . Consistent with a previous report [25] , we find that loss of consciousness is marked by an increase in normalized clustering ( ) , which measures the ‘cliquishness’ of brain regions , potentially indicating an increase in localized processing and thus a decrease of information integration across the brain . Our multi-stage design , however reveals that clustering remains significantly elevated ( as compared to initial wakefulness and sedation ) during post-anesthesia wakefulness recovery . This result shows that while it is true that clustering increases once consciousness is lost , it is not a sufficient marker of consciousness , something that the two-point design ( i . e . , initial wakefulness versus loss of consciousness ) in Schröter et al . [25] could not reveal . On the other hand , we find that the normalized characteristic path length ( ) is significantly increased only during loss of consciousness , suggesting that during unconsciousness the efficiency of information distribution within the network is reduced ( a finding that is consistent with a very recent study on loss of consciousness in sleep [16] ) . Whether this state of increased “functional distance” between regions is causal or consequent to propofol-induced loss of consciousness will have to be addressed in future research . As previously reported , the small-world architecture of brain networks ( ) persisted ( and in fact increased ) in loss of consciousness [25] , confirming the robustness of this core principle of organization of biological networks despite profound state changes [32] . Mirroring , however , small-world architecture also remained significantly elevated during wakefulness recovery . Although much weaker , a similar effect of condition was also uncovered for normalized modularity ( ) . Finally , we remark that the presence of different results observed in the two propofol conditions ( sedation and loss of consciousness ) and , importantly , consciousness recovery , is consistent with the view that changes in global brain topology observed here and elsewhere [25] , [37] are not simply due to drug exposure , but rather reflect brain state changes relating to the loss of consciousness , supporting a previously expressed view [25] . Beyond the global reorganization of brain topology , we also observed changes in local network topology . With respect to nodal strength , selected frontal and parietal regions along the midline , as well some lateral and opercular ROIs , appeared to be modulated by changes in the level of consciousness . In particular , regions in medial frontal and parietal cortices , along with occipital and lateral parietal , exhibited less nodal strength during sedation and loss of consciousness . Other regions , on the other hand , in temporal cortex especially , but also in dorsal and ventro-medial prefrontal cortex , exhibited the reverse pattern . Mirroring the result for , local efficiency appeared to be modulated mostly across midline parietal and prefrontal regions . Overall , this pattern of reorganization of local network topology is consistent with the view that propofol affects specific hubs central to normal/wakeful connectivity [48] which are also known to play a critical role in consciousness [49]–[51] and self-consciousness [52] . Taken together , our findings support the idea that ( propofol-induced ) loss of consciousness correlates with a change in the quality of information processing , and not only a change in the strength of connectivity across regions [1] , [24] . In particular , dynamic reconfiguration of thalamo-cortical and cortico-cortical connections , and contemporaneous decrease of efficiency and increased local processing might affect the degree by which information can be effectively integrated across the brain [9] . In terms of theories of consciousness , these findings can be interpreted as making two contributions . First , the significant increase of cortico-cortical decorrelations during loss of consciousness is coherent with views of consciousness stressing the role of coherent reverberation and spread of neural activity [6] , [7] , particularly within fronto-parietal regions [5] . ( We point out that , as shown in Tables S1 and S2 , all fronto-parietal connections driving the correct classification of loss of consciousness , compared to sedation , are negative . ) Second , our graph theoretic analysis further indicates that , in terms of network information processing , propofol-induced loss of consciousness is marked by a specific change in the quality of information exchange ( i . e . , decreased efficiency ) , consistent with the view that the specific modality with which information is exchanged within brain networks is crucial to the maintenance of a state of consciousness [1] , [9] . Finally , it is important to stress that many of the methodological limitations expressed elsewhere concerning the interpretation of the blood oxygenation level dependent signal , as well as the current challenges tied to applying graph theory to brain measures previously discussed [25] , [28] , [32] , [34] , [37] , also apply to our study . In particular , with respect to the implementation of graph-theory measures in neuroscience , several issues are still in search of resolution . Here , we believe it is important to stress five methodological considerations . First , as we note in the Materials and Methods section , most topological measures require thresholding of adjacency matrices , a procedure that presently lacks a defined standard approach ( e . g . , how many and which thresholds to employ ) and might have important effects on the derived metrics [53] . While the real resolution of the issue will likely include measures that can be applied to fully connected matrices [28] , we stress that our results were robust to the choice of threshold . Second , in contrast to some previous studies [25] , we made use of weighted measures , a difference that might explain the divergence of results . For instance , we note that the observed between-group differences in our study were most pronounced at the lowest density thresholds ( corresponding to least sparse networks ) , in contrast to many binary brain-network studies , in which between-group differences are most pronounced at the highest density thresholds ( corresponding to most sparse networks ) [54] . Many binary-network studies discard as many as 90–95% of all possible connections to elucidate the observed between-group differences [53] and it is likely that these more radical thresholding approaches are associated with substantial loss of connectivity information [55] . High thresholds are needed in binary studies because when weak and strong links surviving thresholding are equally assigned a value of 1 , measures based on path length become susceptible to the creation of spurious long-distance short-cuts , which might obscure the architecture of strong connections and , thereby , important across-group differences [38] . It is therefore possible that the use of binary matrices in previous studies might have obscured the differences in characteristic path length that we have observed . Consistent with our findings , a recent study in the domain of sleep also uncovered loss of efficiency during unconsciousness [16] . Third , as discussed in the Materials and Methods section , because of the known effects of motion on graph theoretic analysis [56] , [57] , our sample was reduced to 12 volunteers . Although this sample size is within the boundaries of previous work on this same topic ( e . g . , N = 11 in [25] , N = 20 in [37] ) it does fall at the low end of the spectrum . Therefore , even though our analyses leverage on a statistically more powerful 4-point repeated measures design ( as compared to the more typical two groups across-subjects comparison and two-points within subject design ) , future studies will have to confirm their generality . Nonetheless , we do stress that the effect-size analysis , which is robust to small samples , shows that our effects are of large magnitude , and that our results are consistent with previous reports [16] . Fourth , while we adopt the presently accepted mainstream interpretation of characteristic path length and global efficiency as measures of functional integration , we acknowledge that these interpretations have not been directly validated and are less trivial to make in networks where edges represent correlations and hence do not necessarily represent causal interactions or information flow [28] . Finally , it is important to stress that a recognized source of variance across results is the choice of ROIs [58] , [59] . In particular , we employed more ROIs than in similar previous studies [25] , [37] , hence it is possible that some of the reported differences are due to the less granular parcellation schemes previously employed . Similarly , it is also possible that , if we had used an even greater number of ROIs , or based our networks on a voxel-wise analysis , results would have differed . However , it has been shown that simple binary decisions concerning the presence of certain network organizational parameters ( e . g . , small-worldness ) are robust across different parcellation granularity [58]–[60] . Consistent with this finding , a recent study evaluating network properties during sleep reported a loss of efficiency during loss of consciousness that paralleles our own findings , despite the fact that their networks featured more than 3 , 700 nodes [16] . It should be stressed , however , that high granularity parcellations might yield quantitatively very different estimates of network properties , as compared to low granularity parcellations , and might allow topological features to be displayed more prominently [58] , [59] . There is , however , an important conceptual difference that separates region-based networks from voxel-based networks [34] , [61] . In our report , as in all region-based analyses of brain connectivity , network locality is conceived at a specific scale , determined by the coarseness of the employed parcellation . Hence , when we investigate local network properties , we are investigating topological features calculated over proximal brain regions . Conversely , voxel-wise networks assess locality within regions of the brain , an approach which has the potential advantage of capturing differences across regions of the brain in within- and between-connectivity [34] , [61] . In this sense , region-based network analyses might be biased towards highlighting the properties of regions with widely distributed connections at a coarse scale , predominant in heteromodal association areas [62] , and blind to local hierarchical connections more predominant in sensory cortical areas [63] . Voxel-based network analysis , instead , allow for examining inter-regional as well as intra-regional connectivity [34] . Nonetheless voxelwise parcellations might however pose conceptual difficulties with respect to computing global network properties because grid-like subdivisions do not generally respect boundaries or sizes of heterogeneous functional areas , an approach that might lead to mischaracterization of brain network function [64] . In conclusion , in interpreting our results ( as any region-based network analysis with comparably sized , or larger , ROIs ) it is thus important to keep in mind that our statements concerning changes in local topological features are intended as network-local , and do not necessarily reflect local changes at the brain physical level . In sum , our findings show that changes in the level of consciousness induced by propofol affect basic organization principles and dynamics of information processing across the whole brain as well as within specific regions known to be involved in consciousness . In particular , we find that propofol-induced loss of consciousness is mostly associated with cortico-cortical mechanisms , as opposed to thalamo-cortical ones , and with a substantial decrease in the efficiency of information flow within the network . Future research will have to assess whether different anesthetic agents and pathology ( e . g . , brain trauma , seizures ) induce loss of consciousness via the same mechanisms .
The study was approved by the Ethics Committee of the Medical School of the University of Liège ( University Hospital , Liège , Belgium ) . Data analysis was carried out in three stages: initial preprocessing , support vector machine ( SVM ) matrix classification , and computation of global and local graph-theoretic measures .
|
One of the most elusive aspects of the human brain is the neural fingerprint of the subjective feeling of consciousness . While a growing body of experimental evidence is starting to address this issue , to date we are still hard pressed to answer even basic questions concerning the nature of consciousness in humans as well as other species . In the present study we follow a recent theoretical construct according to which the crucial factor underlying consciousness is the modality with which information is exchanged across different parts of the brain . In particular , we represent the brain as a network of regions exchanging information ( as is typically done in a comparatively young branch of mathematics referred to as graph theory ) , and assess how different levels of consciousness induced by anesthetic agent affect the quality of information exchange across regions of the network . Overall , our findings show that what makes the state of propofol-induced loss of consciousness different from all other conditions ( namely , wakefulness , light sedation , and consciousness recovery ) is the fact that all regions of the brain appear to be functionally further apart , reducing the efficiency with which information can be exchanged across different parts of the network .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
Dynamic Change of Global and Local Information Processing in Propofol-Induced Loss and Recovery of Consciousness
|
Disseminated histoplasmosis , a disease that often resembles and is mistaken for tuberculosis , is a major cause of death in patients with advanced HIV disease . Histoplasma antigen detection tests are an important addition to the diagnostic arsenal for patients with advanced HIV disease and should be considered for inclusion on the World Health Organization Essential Diagnostics List . Our objective was to systematically review the literature to evaluate the diagnostic accuracy of Histoplasma antigen tests in the context of advanced HIV disease , with a focus on low- and middle-income countries . A systematic review of the published literature extracted data on comparator groups , type of histoplasmosis , HIV status , performance results , patient numbers , whether patients were consecutively enrolled or if the study used biobank samples . PubMed , Scopus , Lilacs and Scielo databases were searched for published articles between 1981 and 2018 . There was no language restriction . Of 1327 screened abstracts we included a total of 16 studies in humans for further analysis . Most studies included used a heterogeneousgroup of patients , often without HIV or mixing HIV and non HIV patients , with disseminated or non-disseminated forms of histoplasmosis . Six studies did not systematically use mycologically confirmed cases as a gold standard but compared antigen detection tests against another antigen detection test . Patient numbers were generally small ( 19–65 ) in individual studies and , in most ( 7/10 ) , no confidence intervals were given . The post test probability of a positive or negative test were good suggesting that this non invasive diagnostic tool would be very useful for HIV care givers at the level of reference hospitals or hospitals with the infrastructure to perform ELISA tests . The first results evaluating point of care antigen detection tests using a lateral flow assay were promising with high sensitivity and specificity . Antigen detection tests are promising tools to improve detection of and ultimately reduce the burden of histoplasmosis mortality in patients with advanced HIV disease .
Histoplasmosis was first described in Panama in 1906 in a patient who appeared to have miliary tuberculosis , a confusion that is still very much present today . [1] Disseminated histoplasmosis has been an AIDS-defining infection since 1987 . [2 , 3] Recent estimates using different methods converge on the conclusion that , in Latin America , each year , over 22000 HIV-infected patients get disseminated histoplasmosis and that between 5000 and 10000 HIV-infected persons die from it , mostly for lack of diagnosis . [4] A multicentre study in Latin America reported a 79% increase in the hazard of dying among culture negative “tuberculosis” cases , the author concluded that the patients had another diagnosis , presumably a good illustration of the confusion that is so frequent between tuberculosis and histoplasmosis . [5] In French Guiana , among consecutive HIV patients hospitalized for infectious symptoms ( fever , isolated or with other symptoms ) , 42% of those with CD4<200 had culture-confirmed histoplasmosis , and 85% of those with CD4<50 had histoplasmosis[6]; In Fortaleza , of 378 consecutively admitted HIV patients , 164 ( 43% ) had microscopically confirmed histoplasmosis[7]; In Panama , 7 . 65% of patients with an HIV infection had culture-positive H . capsulatum[8] , in Guatemala , and in Colombia histoplamosis and TB are the main opportunistic infections;[9 , 10] In Venezuela , in patients with AIDS , histoplasmosis was documented in 29 of 66 ( 44% ) autopsies performed[11 , 12] . Most histoplasmin prevalence studies took place before the AIDS epidemic was identified , because histoplasmin testing is no longer performed . They however are a good marker for the current endemicity of Histoplasma in the countries where they were performed . [13] In all the above regions , these past studies showed that approximately 30% of the population was reactive to a histoplasmin skin test , thereby illustrating the ubiquity of the fungus[14] . These few studies illustrate the importance of histoplasmosis among HIV-infected patients in Latin America , however much of the continent has no data and often no diagnostic capacity . [4 , 15 , 16] But the data gap is even worse in other parts of the world . Cases have been reported in many parts of Africa . [17] In Cameroun a prospective study found that 13% had microscopically confirmed histoplasmosis[18] , preliminary unpublished results in South Africa using antigen detection showed that 10–15% of hospitalized HIV-infected patients with a CD4 count <100 had positive antigenemia . In South Asia , South East Asia , China histoplasmosis has also been increasingly reported in HIV patients[19 , 20] . Overall , there are enormous data gaps regarding the global burden of histoplasmosis , and the lack of simple diagnostic tools creates a vicious circle where the absence of data perpetuates the low awareness of the importance of the problem and the lack of research on the topic . [21] Diagnostic tests are thus crucial to improve public health and to improve the care of patients with advanced HIV disease who are most at risk of developing disseminated histoplasmosis . Cohort studies have shown that incidence increases rapidly as CD4 counts fall below 200 cells per mm3 . [22] The isolation of the pathogen may be performed using direct examination of tissue biopsy ( identification of yeast-like forms in tissue from bone marrow , lymph nodes , liver , intestine and other organs… ) , allowing strong suspicion of the diagnosis ultimately confirmed by culture , which often takes over a month and requires expertise and a BSL2/3 laboratory because of the hazard of inhaling the fungal microconidia . [23] Treatment should be presumptive ( Amphotericin B or itraconazole depending on the severity ) given the delays of fungal culture , and waiting for results to start treatment may delay care for weeks and lead to the patient’s death . [24 , 25] The invasive procedures , the mycological expertise required and the important delays have led to search for alternative diagnostic methods using non invasive , diagnostic tests giving rapid results to the physician . Histoplasma antigen detection tests have been used in the USA for over 30 years[26] , but rarely in tropical regions[27–29] . Antigen detection tests are the simplest diagnostic method that could be implemented in low and middle income countries in order to diagnose HIV patients with advanced disease and/or severe illness including a TB-like presentation . There have been 2 systematic reviews on antigen detection tests . [30 , 31] One included a meta-analysis but with some methodological issues . First , this metaanalysis included a study on antibody detection with the antigen detection studies . Secondly , the meta-analysis pooled sensitivity and specificity of different Histoplasma antigen detection tests , which is debatable . Overall , both systematic reviews concluded that antigen detection tests in general were sensitive and specific diagnostic tools . However , the publication of new studies and concerns for some study limitations and the variation they introduce led us to review the literature in the specific context of advanced HIV disease , with a special focus on low and middle income countries . The systematic review aimed at providing evidence-based arguments for the essentiality of Histoplasma antigen detection test in persons with advanced HIV in low and middle income countries .
This analysis did not deal with individual patient data but with published data , which does not require regulatory approval . A systematic review of published articles and conference abstracts evaluating Histoplasma antigen detection was conducted . PubMed , Scopus , Lilacs and Scielo databases were searched for published articles in English , Spanish and Portuguese between 1981 and 2018 . The search terms usedwere: “histoplasmosis” and “antigen detection” , or “histoplasmosis” and “antigenuria” , or “histoplasmosis” and “antigenaemia” . Articles reporting studies in humans and the diagnostic accuracy of antigen detection tests were retained . To ensure completeness , we crosschecked with 2 recent systematic reviews to determine whether we had not missed important studies[30 , 31] . Finally we looked at Food and Drug administration reports of the Immuno-Mycologics ( IMMY ) Alpha study . [32] Then , because we were evaluating rigorous studies in HIV patients with disseminated histoplasmosis , articles with culture as a gold standard comparator , and including consecutive severely ill HIV-infected patients with disseminated histoplasmosis were retained . This decision was based on our focus on patients with advanced HIV disease and our assumption that among these patients disseminated histoplasmosis is a major killer . Thus , ideally , all patients ( cases and controls ) should be HIV-infected and all histoplasmosis patients should be microbiologically confirmed histoplasmosis ( gold standard ) , because from a clinician’s point of view this would reflect the diagnostic challenge when facing a patients with unspecific signs of infection and advanced HIV disease . The studies retained used different diagnostic tests therefore we decided not to synthesize their results in a meta-analysis . Supporting information files show the PRISMA checklist and the PRISMA flow diagram . From the reported sensitivities and specificities we computed pre and posttest odds , positive and negative likelihood ratios , and posttest probabilities for different hypothesized proportions of histoplasmosis prevalence and for both positive and negative test results . The posttest odds was the multiplication of the pretest odds ( hypothesized prevalence ) by the positive likelihood ratio; the posttest probability was posttest odds/ ( 1+posttest odds ) . Pretest and posttest odds , positive and negative likelihood ratios , and posttest probabilities were calculated for scenarios of Histoplasma prevalence of 1% , 10% , 20% , and 40% in patients with advanced HIV disease . When missing in the published article , we calculated 95% confidence intervals . We computed the point estimates but also the lowest and highest values derived from the 95% confidence interval . Data were analyzed using STATA 13 ( STATA Corporation , college station , Texas ) .
Table 1 shows the 16 studies retained and their characteristics . Studies on Histoplasma antigen detection tests included heterogeneous groups of patients , some ( n = 2 ) without HIV or others ( n = 3 ) mixing HIV and non HIV patients [32–34] , with disseminated or non-disseminated forms of histoplasmosis , which may introduce great variation in measuring the diagnostic accuracy of a test . In 4 studies it was not clear whether patients were HIV-infected or not ( Table 1 ) . The control groups , when present , were also different: to calculate specificity some studies used non fungal controls , while others did not , thus potentially introducing variability . Four studies did not use mycologically confirmed cases as a gold standard but compared antigen detection tests against another antigen detection test , calculating sensitivity and specificity in the absence of a gold standard[35–38] , instead of looking for agreement using split samples . Patient numbers were generally small ( usually ranging from 19 to 65 histoplasmosis cases , the largest compiling 158 cases ) and , in most individual studies , no confidence intervals were given . Finally , many studies used stored samples and failed to include consecutive patients . The first study in AIDS patients by Wheat et al . performed a comparison with a gold standard of 61 AIDS cases with microscopically proven disseminated histoplasmosis and 30 AIDS controls ( total obtained by adding controls with different opportunistic infections ) . The assay had a very high sensitivity ( 96 . 7% ( 95% CI = 88 . 6–99 . 6 ) ) and specificity ( 100% ( 95% CI = 88 . 4–100 ) ) . [26 , 39] Receiver operator characteristic ( ROC ) curves or likelihood ratios were not calculated in the paper . When calculating the post-test probabilities with the lower bounds of the 95% confidence intervals for sensitivity and specificity , a worst case scenario , the posterior probability of a positive test was 5% if prevalence was 1% , 38% for a 10% prevalence , 58% for a 20% prevalence and 78% for a 40% prevalence . The calculations using 96 . 7% and 100% point estimates yielded a 98 . 9% posterior probability for prevalence at 1% . The patients , however , were apparently not included consecutively . The second generation test from MiraVista was apparently tested using consecutive AIDS patients with disseminated histoplasmosis and 100 controls without fungal infections , who did not have HIV or AIDS[40] . We thus excluded the study from our analysis . Regarding the IMMY Alpha , although the samples ( in data submitted to the FDA ) were tested using microscopically confirmed histoplasmosis as a reference standard[32] , it is not clear what proportion of patients were HIV-positive , or what proportion were disseminated histoplasmosis . The study was thus not further analyzed . After retaining studies which consecutively enrolled HIV-seropositive patients with a comparison against the gold standard of culture only 3 studies remained . These studies compared 2 different antigen tests ( the CDC polyclonal antigen test and the Immy monoclonal antigen test ) to culture , all 3 studies having taken place in 2 populations in Latin America . [27–29] The CDC test in urine had a sensitivity of 81% ( 95%CI = 67–91% ) and a specificity of 95% ( 95%CI = 91–98% ) . The area under the ROC curve was 0 . 87 ( 95% CI , 0 . 80 to 0 . 95 ) , and positive and negative likelihood ratios were 16 . 1 ( 7 . 4–45 . 5 ) ) and 0 . 2 ( 0 . 09–0 . 36 ) , respectively . Thus for areas of low prevalence , there was a low post test probability when the CDC test was positive , but the post test probability increased rapidly with prevalence ( Fig 1 ) . On the contrary in areas with high prevalence a negative CDC test was still associated with a 0 . 11 probability of having the disease . Overall the IMMY monoclonal test performed better than the CDC test in a similar Latin American context ( Fig 2 ) . For the IMMY monoclonal antigen detection when using the manufacturer instructions test sensitivity was 98% ( 95%CI = 95–100% ) and specificity was 97% ( 95%CI = 96–99% ) . The areas under the ROC curve were 0 . 99 for the quantitative determination and 0 . 97 for a semi quantitative adaptation . The results in the patients from Guatemala and from the patients in Colombia showed similar performances . The positive likelihood ratiowas 32 . 6 ( 95% CI = 19–100 ) and the negative likelihood ratio was 0 . 02 ( 95%CI = 0–0 . 05 ) . Finally , an abstract presented at the 2018 ISHAM conference compared a new lateral flow assay from MiraVista diagnostics used on culture confirmed HIV-associated histoplasmosis and non-HIV infected controls . The point estimates for this new point of care test were 95% sensitivity and 82% specificity .
The first antigen detection test was developed in the USA in 1986 , using polyclonal antibodies against Histoplasma galactomannan . [26] It was initially a radio immune assay and was subsequently modified as an EIA . [34 , 40] This test has very good reported sensitivity and specificity but it is only performed in Indianapolis at MiraVista Diagnostics and it is not FDA approved . In the context of low and middle income countries this test is thus apparently not a viable option . [25] In 2007 , a polyclonal FDA approved EIA was commercialized ( IMMY Alpha ) . There were conflicting results after comparisons with the MiraVista test with a variety of methodological issues ranging from the types of patients selected in the study , their HIV status , the test procedure , and other biases . If we focus on item 18 of the recent PRISMA DTA guidelines , it is noteworthy that there was a conflict of interest ( many of the authors pointing the IMMY alpha’s lack of sensitivity were linked to MiraVista a company with a large share of the United States’ market of histoplamosis diagnosis ) . [44] Despite these methodological issues , the early controversy has stood in the way of a widespread use of the IMMY alpha . [32 , 35 , 36] The IMMY test has been recently modified using monoclonal antibodies which have greatly improved its sensitivity . [29] Given the lack of diagnostic test for low and middle income countries the CDC’s mycotic diseases branch developed a polyclonal EIA that was evaluated in consecutive HIV patients in Colombia and Guatemala , and was successfully implemented in Brazil , Suriname , and French Guiana . [28] Although there is an agreement that antigen detection was a very sensitive and specific non-invasive test , [23] the published studies have greatly varied in design and in comparison groups . The burden of histoplasmosis in patients with advanced HIV-disease has been greatly underestimated , [4 , 20] in this specific context looking at consecutive HIV-infected patients with advanced disease , the antigen tests seem to have great value . The CDC test and the MiraVista EIA are not likely to be submitted for FDA approval , and the IMMY monoclonal EIA is apparently in the process of FDA approval for commercialization . There is currently 1 test with FDA approval: IMMY alpha . [32] We did not retain the analysis from the FDA data because it was not clear that all cases only had disseminated histoplasmosis or a mix of clinical forms , and if any or some cases had HIV . Beyond diagnostic accuracy , the epidemiological context is important to keep in mind when interpreting the results of a test . We do not precisely know the burden of histoplasmosis in HIV-positive patients in many parts of the world . Until now , antigen detection tests have been used successfully for research in a number of tropical countries including , Colombia , Guatemala , Panama , Brazil , French Guiana , and Suriname . This suggests that antigen detection is appropriate for different regional clades . [27–29 , 42] Costing will be an important aspect since diagnosis is mostly a problem in low and middle income countries with high HIV prevalence . [25] Although it has been argued that doing the antigen test on urine and serum increased sensitivity , [45] recent studies have shown there was no significant benefit to repeat the test . [46] In resource limited countries , this would unnecessarily increase costs . There are still discussions on whether urine is more sensitive than serum , this should be further evaluated using a proper study design in consecutive patients with advanced HIV disease . The test is an EIA format and therefore requires a minimum infrastructure with an EIA reader , electricity , refrigeration , and organization with sample transport , conservation , batching EIA runs … Therefore , reference hospitals and large hospitals may be equipped to perform such a test . The cost of an EIA for low and middle income countries is not available yet . However , costs often increase dramatically with multilayered distribution networks that are necessary for a manufacturer in a high income country to reach the end users in countries where they have no presence . In this situation , each additional intermediary level will amplify costs , whatever the manufacturers’ initial cost . The fact that reference laboratories are targeted would possibly make it easier for manufacturers to streamline distribution and avoid unnecessary cost hikes . When point of care antigen detection format becomes available this would allow further scale up of diagnosis to the most remote health care facilities . The recent communication by Caceres et al . on the miravista lateral flow assay seems very promising and could radically change things if the test confirms its diagnostic performance and if it becomes available in all endemic countries at an affordable price . The WHO has published an essential diagnostics list[47] . We believe that Histoplasma antigen detection tests should be included because this would greatly raise awareness of clinicians and public health authorities , a crucial first step to reduce unnecessary AIDS deaths for lack of a proper diagnosis . [4] While the performance measurements of the test are a first step , the most important question would be will these tests make a difference in terms of saving lives ? Thus , as recommended by WHO , a PICO ( Population , Intervention , Comparison , Outcome ) question regarding antigen detection tests could be: Among patients with advanced HIV-disease , are Histoplasma antigen detection tests better than the present standard of diagnosis to improve the diagnosis of histoplasmosis and reduce mortality ? Since the most frequent standard of diagnosis is no diagnosis for histoplasmosis , the answer to this question may seem straightforward: it is better to make a diagnosis than to miss a diagnosis . In terms of outcome , for diagnosis and mortality , data from Colombia showed that the availability of antigen testsand training dramatically increased the number of diagnoses . [48] In French Guiana , increased awareness and diagnostic progress led to very important increases in the number of diagnoses and a 4 fold reduction of case fatality at one month . [49] Increased awareness alone will have huge benefits on the number of patients diagnosed and treated , an antigen detection test may even show more of the hidden part of the Histoplasmosis “iceberg” . In conclusion , the studies on Histoplasma antigen detection methods have suffered from great heterogeneity , partly because it is challenging to get sufficient numbers of consecutive HIV-infected culture proven disseminated histoplasmosis cases . Excellence centers in low and middle income countries seem better positioned to perform these studies . As shown by the pioneering Colombian and Guatemalan collaboration studies , the evaluation of future antigen test upgrades ( lateral flow assays ) should rest on well-designed studies in consecutive patients with HIV and confirmed histoplasmosis test that can best be achieved through a North-South collaboration . Furthermore , 30 years after the first test , Histoplasma antigen detection tests manufacturers should go through the regulatory approval process in order to make tests available in low and middle income countries which have been an underappreciated potential market . Meanwhile , the available antigen detection tests , should be included in the essential diagnostics list to start mapping the global burden of disseminated histoplasmosis . This would greatly accelerate the goal of having diagnostic tests and effective drugs for disseminated histoplasmosis in most hospitals in endemic countries be achievable . [50]
|
Disseminated histoplasmosis , a disease that often resembles and is mistaken for tuberculosis , is a major cause of death in patients with advanced HIV disease . Histoplasma antigen detection tests are an important addition to the diagnostic arsenal for patients with advanced HIV disease and should be considered for inclusion on the World Health Organization Essential Diagnostics List . Our objective was to systematically review the literature to evaluate the diagnostic accuracy of Histoplasma antigen tests in the context of advanced HIV disease , with a focus on low- and middle-income countries . Systematic review of the published literature extracted data on comparator groups , type of histoplasmosis , HIV status , performance results , patient numbers , whether patients were consecutively enrolled or if the study used biobank samples . At the end of the screening process we included a total of 16 studies in humans for further analysis . Most studies included used a heterogeneous group of patients , often without HIV or mixing HIV and non HIV patients , with disseminated or non-disseminated forms of histoplasmosis . Patient numbers were generally small in individual studies and , in most of these , no confidence intervals were given . When considering the diagnostic accuracy of Histoplasma antigen detection tests evaluated among consecutive HIV-infected patients with confirmed histoplasmosis , the performance of the tests was good . These non invasive diagnostic tools would be very useful for HIV care givers at the level of reference hospitals or hospitals with the infrastructure to perform ELISA tests . Antigen detection tests are promising tools to improve detection of and ultimately reduce the burden of histoplasmosis mortality in patients with advanced HIV disease .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"histoplasmosis",
"pathogens",
"microbiology",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"drug",
"delivery",
"system",
"preparation",
"antigen",
"encapsulation",
"rna",
"viruses",
"pharmaceutics",
"fungal",
"diseases",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"medical",
"microbiology",
"hiv",
"aids",
"microbial",
"pathogens",
"research",
"assessment",
"hiv",
"diagnosis",
"and",
"management",
"diagnostic",
"medicine",
"tuberculosis",
"diagnosis",
"and",
"management",
"viral",
"pathogens",
"systematic",
"reviews",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"lentivirus",
"pharmaceutical",
"processing",
"technology",
"organisms"
] |
2018
|
Histoplasma capsulatum antigen detection tests as an essential diagnostic tool for patients with advanced HIV disease in low and middle income countries: A systematic review of diagnostic accuracy studies
|
Recent studies have found that extracellular vesicles ( EVs ) play an important role in normal and disease processes . In the present study , we isolated and characterized EVs from the brains of rhesus macaques , both with and without simian immunodeficiency virus ( SIV ) induced central nervous system ( CNS ) disease . Small RNA sequencing revealed increased miR-21 levels in EVs from SIV encephalitic ( SIVE ) brains . In situ hybridization revealed increased miR-21 expression in neurons and macrophage/microglial cells/nodules during SIV induced CNS disease . In vitro culture of macrophages revealed that miR-21 is released into EVs and is neurotoxic when compared to EVs derived from miR-21-/- knockout animals . A mutation of the sequence within miR-21 , predicted to bind TLR7 , eliminates this neurotoxicity . Indeed miR-21 in EV activates TLR7 in a reporter cell line , and the neurotoxicity is dependent upon TLR7 , as neurons isolated from TLR7-/- knockout mice are protected from neurotoxicity . Further , we show that EVs isolated from the brains of monkeys with SIV induced CNS disease activates TLR7 and were neurotoxic when compared to EVs from control animals . Finally , we show that EV-miR-21 induced neurotoxicity was unaffected by apoptosis inhibition but could be prevented by a necroptosis inhibitor , necrostatin-1 , highlighting the actions of this pathway in a growing number of CNS disorders .
HIV-associated neurocognitive disorder ( HAND ) is a central nervous system ( CNS ) associated neurological disease where neurodegeneration is a consequence of CNS inflammation . The pathological characteristics of the most extreme form of this disease include astrogliosis , microgliosis , presence of multinucleated giant cells , and loss of dendrites and synapses [1–3] , collectively termed HIV encephalitis ( HIVE ) . These features are recapitulated in its nonhuman primate equivalent rhesus macaque model , simian immunodeficiency virus encephalitis ( SIVE ) [4] . In the CNS , HIV primarily infects microglia and macrophages but not the neurons . However , inflammatory molecules , as well as HIV gene products that are released from infected cells , have damaging affects on neurons [5–8] . Previously , others and we identified that SIV/HIV infection upregulated microRNAs ( miRNAs ) in macaque and human brains [9–11] . These studies have shown that upregulation of miRNAs can also lead to neuronal dysfunction by targeting crucial genes and by repressing their expression in the CNS . Further , we also identified that some of these miRNAs can be released extracellularly in extracellular vesicles ( EVs ) [12] . EVs are small membrane-bound structures . They play a significant role in cell-cell communication [13–16] , in progression of cancer [17] and in viral infections [18–20] . In the brain , astrocytes [21] , microglia [22] and neurons [23] have been shown to release EVs such as exosomes under physiological conditions . There is growing evidence for intercellular EV transfer within the CNS . EVs have been repeatedly discussed as potential carriers in the dissemination of disease pathology in neurodegenerative disorders , as they harbor proteins and RNA that can be transferred from the originating cell to a target cell [24] . We have previously identified that miR-21 is significantly upregulated during SIV/HIV infection in the brain [11] . Thus , we hypothesized that miR-21 may be present within EVs during SIV/HIV associated neuroinflammation and therefore , can be damaging to neurons . Intriguingly , a recent study indicated that certain extracellular miRNAs could bind to toll-like receptors ( TLRs ) in neurons and cause neurodegeneration [25] . These miRNAs had a G/U rich region capable of activating TLR7/TLR8 . Interestingly , miR-21 is one such miRNA . The overall goal of this study was to investigate whether miR-21 was significantly enriched in EVs in SIVE pathogenesis and if such an increase induces deleterious signaling pathways downstream . Here , for the first time , we report the miRNA profiling of EVs from the brain . We find that miR-21 is increased in EVs during SIVE pathogenesis and that it is deleterious to neurons by activating TLR7 dependent downstream cell death pathways . Hence , our data provide insight into the evolving EV-biology field and further expands our knowledge on understanding the molecular mechanism underlying the cause for neuronal damage during SIV/HIV-infection of the brain .
Previously , we determined that miR-21 is upregulated in SIVE and HIVE [11] . Recent studies reported that certain miRNAs such as miR-21 , if present extracellulary or in extracellular vesicles ( EVs ) could trigger TLR signaling pathways by acting as a ligand leading to cell injury [25–27] . Hence , we questioned whether miR-21 in association with EVs in SIVE neuropathology and whether this EV miR-21 ( EV-miR-21 ) causes neuronal damage . EVs were isolated from SIVE and uninfected macaques brain regions using a sucrose gradient protocol [28] . Transmission electron microscopy ( TEM ) was used to characterize the EVs . The results revealed a size of ~100–150 nm with an appearance ( cup-like ) of vesicles that were previously described as exosomes ( Fig 1A , left ) . Western blotting confirmed the presence of proteins associated with EVs: Flottilin , CD9 . CD63 , CD81 , HSP70 and TSG101 ( Fig 1A , right ) . Next , we extracted RNA from EVs , and small RNA sequencing was conducted . The results revealed that miR-21 was significantly upregulated in EVs derived from the SIVE brain samples when compared to uninfected animals , as well as to SIV infected animals that did not have CNS disease ( Fig 1B and S1 Table ) . Additionally , we also found two other miRNAs to increase at much lower levels of change and significance , miR-100-5p and miR-146-5p , and one miRNA to be decreased , miR-126-5p . The change in expression of miR-21 was then validated by quantitative real time polymerase chain reaction ( qRT-PCR ) on the EV samples for miR-21 , revealing significantly elevated expression of miR-21 in SIVE samples ( Fig 1C ) . Our initial studies found that in SIVE miR-21 is upregulated in neurons [11] . Trans migration of cargo from EVs has been shown to enter neurons from non-neuronal cells such as macrophages , microglia and astrocytes [16] . During HIV and SIV infection , macrophages infiltrate the brain , and activated macrophages as well as microglia and astrocytes are found . In order to examine whether such non-neuronal cells in the brain express miR-21 during infection , we performed fluorescence in situ hybridization ( FISH ) coupled with immunofluorescent ( IF ) labeling on brain tissue sections of SIVE and uninfected macaques . As a positive control , U6 , a noncoding snRNA , showed abundant signals in most cells in the tissue; as a negative control , a scrambled miRNA probe did not show any hybridization in these sections . Interestingly , miR-21 signal was seen in CD163 positive macrophages/activated microglial cells and cells with the phenotypic appearance of neurons , whereas minimal signaling is seen in GFAP positive astrocytes ( Fig 2 , SIVE-Mag panel ) . In uninfected controls , miR-21 expression was below the detection limit , although U6 could still be detected ( Fig 2 , Uninfected panels ) . Therefore , it is possible that during infection macrophages could secrete EVs containing miR-21 that could then affect neurons . Given the prime role of macrophages in neuropathogenesis of HIV/SIV and the presence of miR-21 in macrophages in the infected brain , we used macrophages as the cellular model for EV release in our experiments . Recent studies have found that certain microRNAs containing a GU-rich sequence could activate TLR7 . Neurotoxicity and neuronal and non-neuronal cell activation has been found with such free microRNAs and with synthetic EVs of lipid-encapsulated microRNAs [25–27] . First , we asked if the presence of extracellular miR-21 could render neurotoxicity . To do so , we used miRNA oligonucleotides ( oligos ) of wildtype miR-21 ( miR-21-WT ) , a mutant miR-21 ( miR-21-Mut ) containing a point mutation in one of the uridine residues in a small G/U sequence in the TLR binding motif ( U to G , since uridines are more crucial ligands to TLRs [29] ) . Another characterized microRNA , the TLR7 ligand let-7b , was used as a positive control . First , we added the free “naked” oligos directly to the hippocampal neuronal cultures . Results indicated no significant cell death observed either in miR-21-WT , miR-21-Mut , or let-7b , assessed with NeuN counting or LDH assay ( Fig 3A , middle and right ) . Next , we tested whether these microRNAs , when encased in EV-like vesicles , could have an effect on neurons . Interestingly , when the neuronal cultures were treated with these synthetic EVs , significant neuronal cell death was observed with miR-21-WT and let-7b but not with miR-21-Mut , again demonstrated by both NeuN cell counting assay and LDH assay ( Fig 3B ) . Staining with the neuronal marker MAP2 also revealed a loss in neurites ( Fig 3C ) . In clear distinction to what we saw with free miR-21 , the delivery of miR-21 in EV-like vesicles is essential to elicit neurotoxicity . To further examine whether EV-miR-21 activates the TLR7 pathway , we isolated EVs from bone marrow derived macrophage cultures prepared from wildtype ( WT ) and miR-21-/- mice and used these , differing in the presence of miR-21 , to examine potential neurotoxicity ( Fig 4A ) . Indeed , there is a significant increase in neuronal cell death when cultures were treated with EVs derived from WT than from miR-21-/- macrophage cultures ( Fig 4B ) . In order to examine if this neurotoxicity is dependent on TLR7 , we performed the neurotoxicity studies on neurons derived from TLR7-/- animals . To confirm that TLR7 -/- neurons do not respond to ligands , we treated the hippocampal neurons isolated from WT and TLR7 -/- mice with TLR7 agonist CL075 . Quantitative RT-PCR on confirms the expression of pro-inflammatory cytokine genes such as IL6 and TNFα only in WT neurons confirming that TLR7 -/- neurons did not respond to TLR7 ligand stimulation ( Fig 4C ) . Treating the TLR7 -/- neurons with WT-EVs and miR-21-/- EVs demonstrated that toxicity depended not only on the presence of miR-21 in the EVs but also upon the presence of TLR7 in the neurons ( Fig 4D ) . These results clearly indicate that both miR-21 and TLR7 are required for the activation of neurotoxic pathways . Since miR-21 is increased in EVs from the brains of monkeys with SIVE , and EV associated miR-21 ( EV-miR-21 ) is associated with neurotoxicity , we then assessed whether EVs isolated from the SIVE ( SIVE-EV ) and uninfected ( control-EV ) brains would show differences in neurotoxicity . Indeed , treatment of neuronal cultures with SIVE-EV significantly increased neuronal death as compared to control-EV ( Fig 5A ) . Next , we asked if the TLR7 pathway is activated by EV-miR-21 . Using HEK ( human embryonic kidney ) cell lines that expressed , or not , TLR7 , in addition to a reporter gene ( secreted alkaline phosphatase ) , we first examined the signaling of the EVs derived from SIVE brains ( as well as use of CL264 , a TLR7 agonist ) . The results indicated a dose dependent signaling with TLR7 , which was not seen with EVs from uninfected brains ( Fig 5B and 5C ) . In order to determine if the miR-21 induced TLR7 signaling , HEK-TLR7 cells were treated with EV-like vesicles . Results indicate that miR-21 induced signaling but not the vehicle control or the miR-21 mutant ( miR-21-Mut ) ( Fig 6A ) . We next examined if the cell death observed in the EV-miR-21 treated neuronal cultures occurs via apoptosis . Since the trigger of apoptosis involves activation of the mitogen activated protein kinase ( MAPK ) signaling pathway , that transduces signals to the nuclear transcription factor NF-κB , we first looked at the expression of these proteins . Western blot analysis revealed that none of the signaling proteins such as p-ERK1/2 , p-JNK and p-p38 changed by treatment with miR-21 WT EVs ( Fig 6B ) . Next , we treated the hippocampal neurons with a pan-caspase inhibitor , z-VAD-fmk , which have been shown previously to the neurotoxicity resulting from let-7b treatment [25] . However treatment of hippocampal neuronal cultures with z-VAD-fmk did not prevent neuronal cell death ( Fig 6C ) . A caspase-independent form of programmed cell death , termed necroptosis , has been recently identified to play a role in disorders of the central nervous system and elsewhere [30] . Necroptosis occurs through a signaling cascade dependent upon receptor interacting protein kinase-1 ( RIPK-1 ) . To determine if necroptosis was involved in the neurotoxicity induced by miR-21 , we treated the cultures with necrostatin-1 , which specifically inhibits RIPK-1 . Indeed the LDH assay results indicate that Nec-1 was able to prevent EV-miR-21 induced neurotoxicity in hippocampal neurons ( Fig 6D ) . Hence the necroptotic , rather than apoptotic , pathway is active in EV-miR-21 induced neurotoxicity .
In this present study , we showed that EV-miR-21 could activate the TLR7 signaling pathway thus leading to neurotoxicity in SIV neuropathogenesis . Through RNA sequencing on EVs isolated from control and SIVE brains , we found differences in several miRNAs , the most striking being miR-21 . Previously , we showed that miR-21 is significantly increased in neurons . Here , we significantly expand this to reveal the presence of miR-21 in brain EVs from macaques with SIV neuropathogenesis . In the diseased brain , microglial/macrophages express miR-21; and in vitro , macrophage produces EVs containing miR-21 . We found that miR-21 when associated with EVs exhibit neurotoxicity , and this neurotoxicity is dependent upon neuronal expression of TLR7 . Furthermore , we also discovered that neurotoxicity by EV-miR-21 is not caused by an apoptotic mechanism but through the activation of a programmed necrotic pathway termed necroptosis . Brain macrophages are the most likely source for EV-miR-21 , although we cannot exclude the possibility that neurons to secrete miR-21 associated EVs as well . Several lines of evidence suggest that miR-21 is upregulated during inflammation in the brain [24 , 31 , 32] . For the first time , we report that miR-21 is upregulated in EVs in the diseased brain and can activate TLR signaling in neurons during SIV infection . TLR7 , similar to other TLRs , is a pattern recognition receptor , and plays a role in pathogen recognition as part of the innate immune system . TLR7 is endosomally located and recognizes single stranded RNA ( ssRNA ) in mice and humans; TLR8 also recognizes ssRNA in humans . TLR7 and TLR8 are related phylogenetically and functionally and have been identified as important sensors of ssRNA from the viral genomes of influenza and vesicular stomatitis virus as well as HIV itself [29 , 33 , 34] . These sequences can specifically activate TLR7 in mice and TLR7/TLR8 in humans [33] . A number of studies have revealed that several miRNAs , such as miR-21 , miR-29a and let-7b , can even serve as physiological ligands of the ssRNA-sensing [25–27] . Ours is the first study so far that has tested this possibility in the context of SIV infection in the brain . Through the repression of its targets , miR-21 was shown previously to act as both pro-apoptotic [35] and anti-apoptotic miRNA [36] . Previously , we showed that miR-21 causes alterations in neuronal physiology by acting through its target gene MEF2C [11] . Expanding upon its pathogenic actions , in this study we found that miR-21 is released via EVs and that it can directly activate neurotoxic signaling pathways by activating TLR7 receptors in the neuron . Using in vitro constructed EVs , EVs from mouse macrophages , and EVs isolated from primate brains , we provide multiple lines of evidence revealing EV-miR-21 signaling through TLR7 , resulting in neuronal demise . Previously , it was shown that “naked” let-7b synthetic oligonucleotide elicited neurotoxicity [25 , 27] . However , in our cultures , we could not see significant neurotoxicity by naked let-7b ( Fig 3A ) . Enclosing let-7b in DOTAP as an EV-like particle , however , resulted in neurotoxicity . A recent study on the role of let-7b in activation of nociceptor dorsal root ganglion ( DRG ) neurons indicated that cell surface expression of TLR7 and another receptor ( TRPV ) were necessary for the effect [27] . Hence , the localization of the TLR7 in the cells , and its interaction with other receptors , might be important for miRNA-mediated activation of signaling pathways such as neurotoxicity and the potential actions of free versus EV-miRNA . Additionally , several other factors present in EVs were shown to mediate inflammatory responses and neurotoxic pathways , and EVs may contain proinflammatory mediators that could contribute to pathogenesis and progression of HAND [37–39] . In neurodegenerative diseases such as Prion disease , Parkinson’s and Alzheimer’s , toxic factors such as prions , tau , amyloid β , α-synucleins , aggregates of superoxide dismutase 1 were shown to be present in EVs eliciting neurotoxicity [40] . It is also unclear as to why miR-21 is localized to specific cell types in the brain , either through its production or its uptake from EVs . Intriguingly , temporal differences in expression patterns have been detected in neurons and astrocytes after ischemic injury , where the miR-21 increase in neurons was much later when compared to astrocytes , which occurred 12 hr post injury [41] . Given the more chronic nature of SIV infection , such temporal differences in the response could not be detected in our experiments . To study pathways potentially activated upon treatment with EV-miR-21 leading to neurotoxicity , we first looked at changes in the phosphorylation of signaling proteins such as ERK , JNK and p-38 in the MAPK pathway . The MAPKs are a family of kinases that transduce signals from the cell membrane to the nucleus in response to a wide range of stimuli , including stress ( reviewed in [42] ) . Interestingly , we did not find any significant changes in the protein expression of signaling proteins belonging to the MAPK pathway . MAPK activation is linked to apoptosis accompanied by caspase activation , in parallel with not finding activation of MAPK members treatment with a pan caspase inhibitor , z-VAD-fmk , did not rescue the neurons from undergoing death indicating that EV-miR-21 caused neurotoxicity by activating a different cell death pathway . Intriguingly , a novel cell death pathway has been reported recently that causes cell death by a regulated necrosis , termed necroptosis [43–46] . Death receptors [47] , interferons , toll-like receptors ( TLRs ) [48] , intracellular RNA and DNA sensors [49] , and probably other mediators induce this pathway . Necroptosis is a programmed necrosis that requires a number of regulatory proteins and a key protein , RIPK1 . RIPK1 has important kinase-dependent and scaffolding functions that inhibit or trigger necroptosis and apoptosis . The development of the RIPK1 inhibitor Nec-1 has been a major breakthrough in research on necroptosis , and the first disease model in which the role of necroptosis was investigated was ischemic brain injury [50] . Studies in several other disease models revealed that Nec-1 was able to prevent cell death in cells undergoing necroptosis [30] . Hence , we tested to see if Nec-1 will be able to rescue neuronal death triggered by EV-miR-21 . Indeed we observed that pretreatment with Nec-1 was able to prevent neurons from undergoing death . Hence for the first time we report that a miRNA ( miR-21 ) in EVs could cause cell death through a necroptotic cell death pathway . Further studies need to be conducted to ascertain the pathway components activated or involved in initiating necroptosis , and whether necroptosis inhibitors may be useful in vivo to lead to clinical studies . In the era of combination antiretroviral therapy , HAND continues to be a common morbidity among individuals infected with HIV . While the severity of the disease has decreased dramatically , it is still poorly understood as to why the milder forms of HAND are prevalent in HIV-1 infected individuals . The inflammatory condition in the brain due to the continued viral presence is one possible explanation for CNS damage [51] . It is interesting that a significant change in miR-21 levels was not seen in animals without CNS disease , which is in support with studies referring to miR-21 as a critical player in inflammation . It was shown previously that miR-21 levels markedly increased during tissue injury and inflammation in the heart [52] , spinal cord [53] , neurons and astrocytes [41] , and in traumatic brain injury [54–58] . Furthermore , it has been already shown that pro-inflammatory cytokine signaling , such as IL6 via the activation of STAT3 promoter , increases miR-21 [59] . In SIVE brains , there is a marked inflammatory cytokine response to the presence of the virus; and therefore , up regulation in miR-21 levels could be expected . In summary , our study for the first time provides evidence of differences in EV derived miRNAs in CNS disease . We found increased miR-21 expression in EVs derived from SIVE brains when compared to controls . We also report for the first time that EV-miR-21 causes neurotoxicity by activating necroptosis , a novel cell death pathway . The studies presented here are novel findings in neuroAIDS research , and the results implicate EVs as crucial communicators between various cells in the brain . In the context of HIV infection , they are mediators of many neurotoxic factors , miR-21 , being one of them . This study will further form a premise for therapeutic studies for prevention of long-term neuronal damage as seen in HAND .
Materials used in these studies were from animal work performed under Institutional Animal Care and Use Committee approval ( Protocol #: 08-034-07-FC and 11-032-05-FC ) from the University of Nebraska Medical Center . Animal welfare was maintained by following the National Institutes of Health Guide for the Care and Use of Laboratory Animals ( National Research Council of the US National Academy of Sciences ) and US Department of Agriculture policies by trained veterinary staff and researchers under Association for Assessment and Accreditation of Laboratory Animal Care certification , insuring standards for housing , health care , nutrition , environmental enrichment and psychological well-being . Primary enclosures consisted of stainless steel primate caging provided by a commercial vendor . Animal body weights and cage dimensions were regularly monitored . Overall dimensions of primary enclosures ( floor area and height ) met the specifications of The Guide for the Care and Use of Laboratory Animals , and the Animal Welfare Regulations ( AWR’s ) . Light cycle was controlled at 12/12 hours daily . All animals were fed standard monkey chow diet supplemented daily with fruit and vegetables and water ad libitum . Social enrichment was delivered and overseen by veterinary staff and overall animal health was monitored daily . Animals showing significant signs of weight loss , disease or distress were evaluated clinically and then provided dietary supplementation , analgesics and/or therapeutics as necessary . These met or exceeded those set forth in the Guide for the Care and Use of Laboratory Animals from the National Research Council of the US National Academy of Sciences . Archived tissue used in these studies was from animal ( Macaca mulatta ) studies performed under Institutional Animal Care and Use Committee approval from the University of Nebraska Medical Center . Animal welfare was maintained by following the National Institutes of Health Guide for the Care and Use of Laboratory Animals . All efforts were made to ameliorate suffering of the animals , including the use of anesthesia with ketamine , xylazine and phenobarbital at necropsy . The following oligoribonucleotides were synthesized by Integrated DNA Technologies ( Coralville , IA , USA ) using all phosphorothioate linkages to protect from degradation , and methyl groups on the 5’ and 3’ nucleotides . The changed base in miR21-mut ( U to G at position 20 ) is underlined . All were used following HPLC purification . miR21-WT: 5'- UAG CUU AUC AGA CUG AUG UUG A -3'; miR21-Mut: 5'- UAG CUU AUC AGA CUG AUG UGG A -3'; and let-7b , 5’- UGA GGU AGU AGG UUG UGU GGU U -3′ . Necrotstatin-1 , a necroptosis inhibitor and z-VAD-fmk , pan-caspase inhibitor , were purchased from Enzo lifesciences ( Farmingdale , NY , USA ) . miR-21−/− and Tlr7−/− mice were purchased from Jackson Laboratories ( Bar Harbor , Maine ) and bred in the UNMC animal facility . Pregnant WT mice were purchased from Charles River ( Wilmington , MA , USA ) . HEK-Blue TLR7 cells designed for studying the stimulation of TLR7 by monitoring the activation of NF-κB and AP-1 were cultured in DMEM supplemented with 10% FBS , normocin ( 50 μg/ml ) , blasticidin ( 10 μg/ml ) , zeocin ( 100 μg/ml ) ( InvivoGen , San Diego , CA ) . Cells were grown at 37° C in humidified air with 5% CO2 . Control HEK-Blue Null cells were cultured similarly except without zeocin . Samples from SIV-infected rhesus monkeys that developed SIVE , and from uninfected control monkeys , were obtained from previous studies . For animals used in this study , the infection was allowed to follow its natural course , and animals were euthanized when they showed signs of simian AIDS . At necropsy , all animals were perfused with PBS containing 1 U/ml heparin to remove blood-borne cells from the brain , and samples were taken and stored at -80° C . Those in which pathological examination revealed multinucleated giant cells , microglial nodules and infiltration of macrophages into the brain were classified as having SIVE . Samples from these animals , as well as uninfected control animals were prepared in a similar manner , were used for this study . FISH and IF were performed as described previously [60] . First , formalin-fixed paraffin-embedded sections were deparaffinized . For combined FISH and IF , this was followed by antigen retrieval using 0 . 01 M citrate buffer and postfixation using 0 . 16 M l-ethyl-3- ( 3-dimethylaminopropyl ) carbodiimide ( EDC; Sigma-Aldrich , St . Louis , MO , USA ) to prevent loss of small RNAs . The sections were incubated with hybridization buffer ( 50% formamide; 10 mM Tris-HCl , pH 8 . 0; 200 μg/ml yeast tRNA; 1× Denhardt's solution; 600 mM NaCl; 0 . 25% SDS; 1 mM EDTA; and 10% dextran sulfate ) for 1 hr at 37° C in a humidified chamber for prehybridization . They were then incubated overnight at 37° C with locked nucleic acid ( LNA ) -modified DNA probes , all labeled with digoxigenin at the 5′- and 3′-termini ( Exiqon , Woburn , MA , USA ) . Probes were used at a concentration of 4 pmol of probe per 100 μl of hybridization buffer . The sequences of the probes are; U6: CAC GAA TTT GCG TGT CAT CCT Y; miR-21: 5’- TCA ACA TCA GTC TGA TAA GCT A -3’; Scramble ( Scr ) 5’- GTG TAA CAC GTC TAT ACG CCC A -3’ . Stringency washes were performed with 2× and 0 . 2× SSC ( Invitrogen , Carlsbad , CA , USA ) at 42° C . The hybridization and wash temperatures were optimized in preliminary experiments . The sections were then blocked with a solution of 1% BSA , 3% normal goat serum in 1× PBS for 1 hr at room temperature , followed by incubation with anti-digoxigenin peroxidase antibody ( 1:100 in blocking buffer; Roche Applied Science , Mannheim , Germany ) overnight at 4° C . For combined FISH and IF , co-incubation with either anti-CD163 ( 1:100; Vector Labs , Burlingame , CA , USA ) or anti-glial fibrillary acidic protein ( GFAP; 1:2000; Dako , Glostrup , Denmark ) was performed at this step . The following secondary antibodies were used: 568 donkey anti-rabbit and 488 goat anti-mouse IgG ( 1:400; Invitrogen ) . This was followed by signal amplification using tyramide signal amplification Cy5 kit ( Perkin Elmer , Waltham , MA , USA ) according to the manufacturer's protocol . The slides were mounted in Prolong gold antifade reagent with DAPI ( Invitrogen ) . The sections were imaged in Zeiss Observer . Z1 microscope equipped with a monochromatic Axiocam MRm camera using Axiovision 40 v . 4 . 8 . 0 . 0 software ( Carl Zeiss , Oberkochen , Germany ) . The following colors were assigned to the fluorescent signals using the Axiovision software: Green for CD163 , Red for GFAP , Magenta for Cy5 , Blue for DAPI . EV isolations from the brains were carried out as described previously with modifications [28] . Previously , dissected and frozen macaque brain tissues ( weighing approximately 500 mg each ) were dissected and treated with 20 units/ml papain ( Worthington , Lakewood , NJ ) in Hibernate A solution ( 5 ml/hemi-brain; BrainBits , Springfield , IL , USA ) and rocked for 15 min at 37° C . The brain tissue was gently homogenized in 10 ml/brain of cold Hibernate A solution . The brain homogenate was sequentially filtered through a 40 μm mesh filter ( BD Biosciences , San Jose , CA ) , a 5 μm filter ( Pall Corporation , Port Washington , NY ) and a 0 . 2 μm syringe filter ( Thermo Scientific , Waltham , MA ) . EVs were isolated from the filtrate as described previously [15 , 28] . Briefly , the filtrate was sequentially centrifuged at 300 × g for 10 min at 4° C , 2000 × g for 10 min at 4° C , and 10 , 000 × g for 30 min at 4° C to discard cells , membranes and debris . The supernatant was centrifuged at 100 , 000 × g for 60 min at 4° C to pellet EVs . The EV pellet was resuspended in 37 ml of cold PBS ( Thermo Scientific , Waltham , MA ) , and the EV solution was centrifuged at 100 , 000 × g for 60 min at 4° C . The washed EV pellet was resuspended in 2 mL of 0 . 95 M sucrose solution and inserted inside a sucrose step gradient column ( six 2 ml steps starting from 2 . 0 M sucrose down to 0 . 25 M sucrose in 0 . 35 M increments , with the 0 . 95 M sucrose step containing the EVs ) . The sucrose step gradient was centrifuged at 200 , 000 × g for 16 hr at 4° C . A 1 ml fraction was collected from the top of the gradient and discarded , and 6 mL of the gradient were collected in the EV rich layers containing material with density higher than 1 . 07 ( 0 . 60 M sucrose layer ) and lower than 1 . 17 ( 1 . 30 M sucrose layer ) [28] . Pooled fractions were diluted to 30 ml with cold PBS . 25 ml of this volume was taken for RNA extraction , and 5 ml used for Western blot studies . Sample volumes were brought up to appropriate volumes with cold PBS and centrifuged at 100 , 000 × g at 4° C for 60 min . PBS was pipetted off both pellets . Protein pellet was suspended in 50 to 100 μl of PBS depending on pellet size . EV isolations from BMDM preparations were carried out by Exoquick ( SBI ) according to manufacturer’s instructions . For transmission electron microscope ( TEM ) , a 10 μl drop of EV sample was placed on the grid ( 200 mesh copper grids coated with Formvar and silicon monoxide ) and allowed to sit for 2 min . The excess solution was drawn off by filter paper , and the remaining thin film of sample was allowed to dry for 2 min . A drop of NanoVan negative stain was placed on the grid for 1 min . The excess negative stain was then drawn off by filter paper and allowed to dry for at least 1 min before being placed in the TEM . Grids were examined on a Tecnai G2 Transmission Electron Microscope ( built by FEI , Hillsboro , Oregon , USA ) operated at 80Kv . Small RNAseq was performed by LC Sciences ( Houston , TX , USA ) . Using the RNA isolated from EVs , a small RNA library was generated using the Illumina Truseq Small RNA Preparation kit following the manufacturer’s guidelines . The cDNA library was purified and used for cluster generation on Illumina’s Cluster Station and then sequenced on the Illumina GAIIx ( Ilumina , San Diego , CA ) . Raw sequencing reads were obtained using Illumina’s Sequencing Control Studio software ( version 2 . 8 ) following real-time sequencing image analyses and base-calling by Illumina's Real-Time Analysis ( version 1 . 8 . 70 ) . A pipeline script , ACGT101-miR v4 . 2 ( LC Sciences ) , was used for sequencing data analyses [61–63] . Sequences were then mapped to miRbase ( version 20 . 0 ) [64] . 636 unique sequences mapped to both Macaca mullata mirs in miRbase and the Macaca mullata genome . Many of these had very low normalized counts ( median/mean for Control , SIV , and SIVE were 6 . 34/444 . 28 , 5 . 17/435 . 8 , and 6 . 04/554 . 50 respectively ) ; thus , only those with >500 counts in any one group ( comprising 75 mirs ) were chosen for statistical analyses . To assess differences between the groups , normalized sequence counts were subjected to a Bayes-regularized one-way ANOVA using analysis conducted using the Cyber-T web server ( http://cybert . ics . uci . edu ) [65 , 66] . The sliding window size was set at 101 , the Bayesian confidence value was 11 , and analysis was performed on the natural logarithm of the values . Significant changes were assigned if the Bonferroni corrected p value was <0 . 05 . For quantification of miRNA in EVs by qRT-PCR , TaqMan mature miR assays ( Applied Biosystems , Carlsbad , CA , USA ) were used according to the manufacturer's protocol . The relative amount of miR-21 was determined by comparison to a standard dilution curve , made from a cDNA preparation from a miR-21 overexpressing cell line . The Ct values of the samples were extrapolated into the standard curve to calculate the relative copy number . We used the formula [RNA/DNA] = 10^Ct-b/m ( where Ct = threshold Ct value , b = Y-intercept and m = slope ) to calculate the amount of miRNA in each sample . Exosomal lysates were prepared using RIPA buffer ( 50 mM Tris/HCl , pH 8; 150 mM NaCl; 1% Nonidet P-40; 0 . 5% sodium deoxycholate; and 0 . 1% SDS ) , and protein quantification was carried out using Pierce BCA protein assay ( Thermo Scientific , Rockford , IL , USA ) . Protein ( 5–15 μg ) was loaded in each lane of NuPAGE 4–12% Bis-Tris gels ( Invitrogen ) . For EV proteins , all the blots for tetraspanins ( CD9 , CD63 , CD81 ) were run on non-reducing gels as described previously [67–69] and flotillin , HSP70 and TSG101 were run under reducing conditions . Separated proteins were transferred onto nitrocellulose membranes using iBlot ( Invitrogen ) . The membranes were blocked in SuperBlock ( TBS ) blocking buffer ( Thermo Scientific ) and then incubated overnight at 4° C with primary antibody . The following primary antibodies were used: rabbit polyclonal anti-Flottilin ( Abcam , Cambridge , MA , USA ) , anti-CD9 ( Systems Biosciences ( SBI ) , Mountain view , CA , USA ) , anti-CD63 ( SBI ) , anti-CD81 ( SBI ) , TSG101 ( SBI ) , HSP70 ( SBI ) Synaptophysin ( Thermo Scientific ) , SNAP-25 ( Cell signaling technology ( CST ) , Boston , MA , USA ) and all the signaling antibodies ( ERK1/2 , pERK1/2 , JNK , p-JNK , p-38 , p38 ) and the positive control lysates were purchased from CST . This was followed by incubation with secondary antibody: HRP conjugated anti-rabbit IgG ( 1:20 , 000; Thermo Scientific ) or anti-mouse IgG ( 1:20 , 000; Thermo Scientific ) for 1 hr at room temperature . Blots were developed using SuperSignal West Pico Chemiluminescent Substrate ( Thermo Scientific ) , imaged and quantified using Carestream MI software ( Carestream Health Inc , Rochester , NY , USA ) . Primary hippocampal cultures were isolated from P0/P1 mice as described previously [70] . In brief , hippocampi were dissected and washed 3x with ice-cold calcium–magnesium-free Hanks' balanced salt solution followed by incubation with 0 . 25% trypsin for 20 min in a 37° C water bath . Followed by subsequent washes with HBSS and complete neuronal media ( neurobasal medium containing 0 . 5 mM l-glutamine and B27 supplement ( Life Technologies , Grand Island , NY ) ) . Individual cells were mechanically isolated by trituration in complete neuronal media with a fire-polished glass pipette . The cells were plated on poly-d-lysine-coated coverslips/plates and cultured in at 37° C in a humidified atmosphere of 5% CO2 incubator . Neurotoxicity assays were performed as described previously with modifications [25] . Synthetic RNAs were diluted in HBS buffer ( 20 mM HEPES , 150 mM NaCl , pH 7 . 4 ) or encased in artificial EVs using N-[1- ( 2 , 3-Dioleoyloxy ) propyl]-N , N , Ntrimethylammonium methylsulfate ( DOTAP ) ( 1811177; Roche , Basel , Switzerland ) . DOTAP was first diluted in HBS- for 5 min before mixing with an equal volume of HBS containing the RNA . The resulting mix was incubated for 20 min and 50 μl were added per well of a 24-well plate , resulting in a final volume of 200 μl . Transfections were conducted in triplicate in all experiments . The ratio of DOTAP to ssRNA was 3:1 ( 3 μl DOTAP to 1 μg RNA ) . For toxicity studies , reagents were added to cell cultures for 24 hrs . For EV toxicity assay , 1 μg of EV preparations were added to 2 × 105 neurons/well in a 24 well plate . Control cultures were incubated with phosphate-buffered saline . After 24 hr , lactate dehydrogenase ( LDH ) assay was conducted on the media according to the manufacture’s instructions ( Cytotoxicity detection kit ( LDH ) , Roche , Basel , Switzerland ) . Briefly , 100 μL of culture medium were transferred to a new 96 well plate . 100 μL of the reaction solution from the kit , containing the detection dye and the catalyst , were then added; absorption was measured after 30 min at 490 nm with 655 nm as reference wavelength . A positive control , 2% triton was used leading to 100% cytotoxicity by lysing the cells completely . The background values from wells without cells were subtracted and average values for the triplicates calculated . Cytotoxicity was then calculated according to the following equation: Cytotoxicity ( % ) = ( experimental value–media control ) / ( positive control–media control ) × 100 . The cells were also immunostained with antibody to NeuN ( Millipore , Billerica , MA , USA ) . For each condition , experiments were performed in duplicates . NeuN-positive cells were counted by analyzing five high-power fields per coverslip . The viability of control cells was set to 100% . The numbers of NeuN-positive cells observed for each condition were compared with control conditions , and results were expressed as relative neuronal viability . For rescue experiments , the hippocampal neurons isolated from WT and TLR7-/- mice were pretreated with necrostatin -1 ( 5 μM ) and z-VAD-fmk ( 10 μM ) or a vehicle control for 1 hr followed by treatment with DOTAP formulations . The inhibitors were stayed in the culture along with DOTAP formulations . LDH assay was performed to assess the neuronal viability . HEK Blue murineTLR7 293 cells and control HEK-Null ( control ) cells were seeded at the concentration of 2 × 105 cells/well . 1 μg of EV preparations were added to the cells and incubated for another 24 hr . After 24 hr , detection medium ( QUANTI-Blue ) was added to the plate . The levels of SEAP ( secreted alkaline phosphatase ) produced by the activation of TLR7 quantitatively using a spectrophotometer at 630 nm .
|
HIV associated neurocognitive disorder ( HAND ) are neurological disorders caused due to the entry of HIV infection in the brain . HIV-1 does not directly infect central or peripheral neurons , however , virus-infected cells of the monocyte/macrophage lineage maintain a low-level HIV infection in the CNS . "Indirect effects" of macrophage activation–such as dysregulation of cytokines and chemokines , free-radical ( oxidative stress ) injury , and secretion of soluble factors that are potently neurotoxic–have been implicated as effectors of nervous system injury in HIV . Here , we report that extracellular vesicles released from macrophages can enhance neurotoxicity . Using a nonhuman primate model of HAND , simian immunodeficiency virus encephalitis ( SIVE ) , we find that exosomes isolated from SIVE brains contain , microRNAs , including miR-21 , that can serve as ligands to the key immune regulatory receptors , toll-like receptors , and can elicit neurotoxicity . We provide in vitro evidence for such an effect , and that the toxicity can be mediated by necroptosis . Thus , our study provides insights into other potential neurotoxic mechanisms by which HIV infection in the brain could harm neuronal health .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
MiR-21 in Extracellular Vesicles Leads to Neurotoxicity via TLR7 Signaling in SIV Neurological Disease
|
To further its pathogenesis , S . Typhimurium delivers effector proteins into host cells , including the novel E3 ubiquitin ligase ( NEL ) effector SspH2 . Using model systems in a cross-kingdom approach we gained further insight into the molecular function of this effector . Here , we show that SspH2 modulates innate immunity in both mammalian and plant cells . In mammalian cell culture , SspH2 significantly enhanced Nod1-mediated IL-8 secretion when transiently expressed or bacterially delivered . In addition , SspH2 also enhanced an Rx-dependent hypersensitive response in planta . In both of these nucleotide-binding leucine rich repeat receptor ( NLR ) model systems , SspH2-mediated phenotypes required its catalytic E3 ubiquitin ligase activity and interaction with the conserved host protein SGT1 . SGT1 has an essential cell cycle function and an additional function as an NLR co-chaperone in animal and plant cells . Interaction between SspH2 and SGT1 was restricted to SGT1 proteins that have NLR co-chaperone function and accordingly , SspH2 did not affect SGT1 cell cycle functions . Mechanistic studies revealed that SspH2 interacted with , and ubiquitinated Nod1 and could induce Nod1 activity in an agonist-independent manner if catalytically active . Interestingly , SspH2 in vitro ubiquitination activity and protein stability were enhanced by SGT1 . Overall , this work adds to our understanding of the sophisticated mechanisms used by bacterial effectors to co-opt host pathways by demonstrating that SspH2 can subvert immune responses by selectively exploiting the functions of a conserved host co-chaperone .
Salmonella enterica serovar Typhimurium ( S . Typhimurium ) , a causative agent of gastroenteritis and typhoid-like fever in mammals , encodes two type III secretion systems ( T3SS ) on Salmonella pathogenicity islands ( SPI ) -1 and SPI-2 . T3SSs are proteinaceous channels encoded by Gram negative bacterial pathogens that transport effector proteins directly into infected host cells [1] where they can interface with , and subvert , host cellular processes [2] . SPI-2 effectors are critical for S . Typhimurium pathogenesis as they are required for systemic infection [3] although the functions of individual effectors in the host milieu are not fully understood . Among the suite of effectors delivered by S . Typhimurium into the host cell are members of the NEL family of E3 ubiquitin ligases comprised of SspH1 , SspH2 and SlrP [4] . This class of E3 ubiquitin ligase has an amino terminal leucine rich repeat ( LRR ) domain and a carboxy terminal catalytic domain [5] . NEL E3 ubiquitin ligases covalently bind ubiquitin through a conserved catalytic cysteine residue and it is postulated that they are autoinhibited by structural constraint that is relieved upon substrate binding [5] . Host binding partners have been identified and putative functions during pathogenesis assigned to each S . Typhimurium NEL effector [4] . We recently identified a novel interaction between SspH2 and the human protein SGT1 ( hereafter referred to as HsSGT1 ) [6] . Originally identified in Saccharomyces cerevisiae as an essential cell cycle protein , yeast SGT1 ( hereafter referred to as ScSgt1p ) interacts with Skp1p , which is a component of the conserved eukaryotic Skp1/Cullin/F-box ( SCF ) E3 ubiquitin ligase . ScSgt1p is required for progression through both the G1/S and G2/M checkpoints of the cell cycle [7] . Homologs of SGT1 that retain these essential G1 and G2 cell cycle functions have been identified in both animal and plant kingdoms . Intriguingly , mammalian and plant SGT1 orthologs have gained additional functional roles in innate immunity as nucleotide-binding leucine rich repeat receptor ( NLR ) co-chaperones through their association with HSP90 [8] , [9] , [10] , [11] . In animal innate immunity Nod1 is a canonical NLR . Upon recognition of the bacterial peptidoglycan dipeptide component , iE-DAP , Nod1 is activated and signals through the transcription factor NF-κB to express pro-inflammatory chemokines such as IL-8 [12] . Importantly , HsSGT1 was shown to co-chaperone Nod1 and HsSGT1-silencing abrogates Nod1-mediated IL-8 secretion in vitro [10] . In contrast to mammalian NLRs that recognize general microbial structures , plant NLRs recognize pathogen-specific components , and upon activation , can induce programmed cell death as part of the hypersensitive response ( HR ) [13] , [14] . For example , the potato NLR , Rx , recognizes the coat protein of potato virus X ( PVX ) , and subsequently inhibits viral replication or induces an HR to stop viral spread [15] . Several plant NLRs , including Rx , are functionally dependent on SGT1 co-chaperone activity , in conjunction with HSP90 and RAR1 [16] . Model assays have been developed for SGT1 NLR co-chaperone function in both animal and plant systems [10] , [17] . In this study we used animal , plant and yeast model systems to gain insights into the biological and mechanistic function of the enigmatic S . Typhimurium NEL type III effector , SspH2 . We identified a novel functional role for SspH2 in host cells where catalytically active SspH2 enhanced SGT1-dependent NLR-mediated immunity in both animal and plant model systems . SspH2 selectively bound NLR co-chaperone-competent SGT1 and the mammalian NLR , Nod1 , in a trimeric complex and monoubiquitinated Nod1 . SspH2 did not affect essential SGT1 G1 and G2 cell cycle functions , however SspH2 protein stability and activity were enhanced by SGT1 in vitro . Thus , these cross-kingdom investigations have helped define a functional role for S . Typhimurium SspH2 that centers on the subversion of immune-specific functions of the host protein SGT1 .
We sought to characterize the interaction between SspH2 and HsSGT1 and determine the functional consequences of this interaction . The specificity of the interaction was confirmed because SspH2 , but not SspH1 that shares approximately 70% sequence identity [4] , co-precipitated with HsSGT1 when expressed in human embryonic kidney ( HEK ) 293T cells ( Fig . 1A ) . SGT1 has three domains , an amino terminal tetratricopeptide repeat ( TPR ) domain , a central Chord and SGT1 ( CS ) domain , and a carboxy terminal SGT1-specific ( SGS ) domain [16] . In humans SGT1 has two isoforms that are splice variants , HsSGT1A and HsSGT1B [16] , and both interacted with SspH2 ( Fig . 1A ) . Similarly , when expressed in cell culture , SspH2 co-localized with endogenous HsSGT1 when examined by immunofluorescence ( Fig . 1B ) . Given the conservation of SGT1 proteins among eukaryotes we tested for interactions between SspH2 and SGT1 proteins from plant ( Arabidopsis thaliana ) and yeast ( S . cerevisiae ) using co-expression in 293T cells . A . thaliana also has two SGT1 isoforms , AtSGT1A and AtSGT1B that each share approximately 40% identity/60% similarity with HsSGT1A [16] . Intriguingly , a similar specific interaction was detected between AtSGT1A or AtSGT1B and SspH2 , but not SspH1 ( Fig . 1C and Fig . S4A ) . By contrast , SspH2 did not interact with ScSgt1p , which shares only ∼25% identity with HsSGT1A ( Fig . 1D ) . To further characterize the interaction between SspH2 and SGT1 , truncation variants of HsSGT1A and SspH2 were generated and tested in the interaction assay described above ( Fig . 1E; primary data is provided in Fig . S1A and S1B ) . Expression of the SspH2 LRR domain , but not the NEL domain , retained interaction with HsSGT1 . An HsSGT1 binding region was identified in the distal portion of the SspH2 LRR domain since a construct lacking residues 430–527 of SspH2 ( designated SspH2ΔBD ) did not interact with HsSGT1 . By contrast , the SspH2 ubiquitin ligase catalytic mutant ( C580A ) [5] retained interaction with HsSGT1 . Similar truncation analyses with HsSGT1 indicated that both the CS and SGS domains of HsSGT1 were required for binding to SspH2 ( Fig . 1E ) . Notably , the CS and SGS domains of HsSGT1 were previously reported to facilitate its co-chaperone function of the mammalian NLR , Nod1 , in animal innate immunity [10] , [11] . Given that interaction with SspH2 involved the same two HsSGT1 domains , an in vitro Nod1-dependent IL-8 secretion assay was developed in HeLa cells ( see Fig . 2A and [10] ) and used to test the impact of SspH2 on this SGT1-dependent Nod1 response . Unexpectedly , SspH2 expression caused a significant increase in IL-8 secretion ( Fig . 2B ) . SspH2ΔBD and SspH2C580A failed to enhance IL-8 secretion in comparison to wild type SspH2 ( Fig . 2B ) suggesting that both HsSGT1 binding , as well as a functional ubiquitin ligase domain , are required for this phenotype . Similarly , expression of the SspH2 LRR or NEL domains individually failed to recapitulate the phenotype of enhanced IL-8 secretion induced by full length SspH2 ( Fig . 2B ) . Interestingly , SspH2ΔN , an SspH2 variant that lacks the first 163 amino acids including the palmitoylation site and membrane-targeting sequence recently identified [18] also significantly enhanced IL-8 secretion , albeit not to the same extent as the full-length protein . Lysate from cell culture was analyzed by Western blot to verify expression of all relevant constructs ( Fig . 2C ) . Collectively , these data suggest that SspH2 binds HsSGT1 and stimulates Nod1 through a mechanism that involves ubiquitin transfer , though SspH2 palmitoylation and/or membrane localization is not strictly required . In addition , IL-8 secretion was also measured in HeLa cells infected with S . Typhimurium strains containing a deletion of sspH2 . Infection with the sspH2 deletion strain complemented with sspH2 , but not empty vector , showed a significant increase in IL-8 secretion compared to the deletion strain ( Fig . 2D ) . Taken together , these data indicate that SspH2 either expressed directly in host cells , or bacterially delivered , can enhance Nod1 response . In this work we have used SspH2 expression constructs that are epitope-tagged at the amino or carboxyl terminal . Given the recent identification of a palmitoylation site at the amino terminus of SspH2 [18] we tested to what extent the position of the epitope tag impacted on SspH2 phenotypes . Both amino- and carboxyl-terminal tagged versions of SspH2 showed associations with HsSGT1 via co-immunoprecipitation or co-localization , respectively ( Fig . 1A and B ) . We also verified that amino-terminal tagged SspH2 was palmitoylated in host cells ( Fig . S2A ) and that the enhancement of Nod1 activity was indistinguishable between both amino- and carboxyl-tagged SspH2 constructs ( Fig . S2B ) . On the basis of these results we conclude that amino- and carboxyl-terminal tagged SspH2 constructs are functionally equivalent . Since both animal and plant SGT1 proteins are critical NLR co-chaperones , we examined whether SspH2 specifically enhances Nod1 response , or whether SspH2 generally alters SGT1-dependent immune responses . Accordingly , the impact of SspH2 was tested using the established SGT1-dependent Rx/PVX immunity assay in planta [19] . This assay employs a PVX variant with genomically encoded GFP to readily visualize PVX replication and has been reconstituted in transgenic Rx-expressing tobacco ( Nicotiana benthamiana ) to facilitate transient gene silencing and expression [19] . A schematic depiction of the assay that outlines the procedure , experimental and control constructs , and the functional readouts from the assay is shown in Figure S3 . Endogenous N . benthamiana SGT1 ( NbSGT1 ) was silenced [19] and complemented with AtSGT1A because its co-chaperone activity confers Rx stability ( thereby mediating resistance to PVX [20] ) and because of its robust interaction with SspH2 ( Fig . 1C ) . As controls for the Rx/PVX immunity assay we verified ( i ) that NbSGT1 depletion was required for PVX replication in Rx-transgenic N . benthamiana ( Fig . 3A; compare Nb Rx TRV:SGT and Nb Rx TRV:00 ) , ( ii ) that immune responses mounted to PVX were dependent on Rx ( Fig . 3A; compare Nb Rx TRV:SGT and Nb WT TRV:SGT ) and ( iii ) that all constructs expressed appropriately ( Fig . S4B ) . As expected in the Rx/PVX immunity assay , complementation of NbSGT1 with β-glucuronidase ( GUS ) as a negative control resulted in strong fluorescence ( Fig . 3A ) . The fluorescent signal was attributed to elevated PVX-GFP levels since lysate from this infiltration contained readily detectable GFP ( Fig . 3B ) . These data verify that PVX replication is robust in the absence of SGT1 . Interestingly , co-expression of AtSGT1A and SspH2 , but not SspH1 or SspH2C580A , also yielded a strong fluorescent signal that was observed in approximately 70% of leaves ( Fig . 3A ) . However , GFP levels were undetectable in this infiltration site ( Fig . 3B ) indicating that PVX replication was still suppressed in the AtSGT1A+SspH2 infiltration . This suggested that the observed phenotype might be caused by autofluorescence , which has been previously described as a marker of the HR immune response [21] . To assess the effect of SspH2 on HR in the transgenic N . benthamiana assay , infiltrations were performed as outlined above and the formation of necrotic-like lesions was examined . SspH2 , but not SspH1 or SspH2C580A , elicited enhanced HR in comparison to AtSGT1A expression alone ( Fig . 3C ) . Importantly , co-expression of SspH2 , AtSGT1A and Rx , did not elicit HR lesions in the absence of PVX ( Fig . S4C ) , indicating that SspH2 toxicity was not contributing to HR development . Similarly , co-expression of SspH2 , Rx and PVX did not elicit HR lesions in the absence of AtSGT1A , confirming the SGT1-dependence of the phenotype ( Fig . S4C ) . These data indicate that active SspH2 interacts with AtSGT1A to enhance plant innate immunity , and taken together with the Nod1 data , implies that SspH2 utilizes a conserved mechanism to alter SGT1-dependent immune responses across kingdoms . Our data indicates that SspH2 interacts with SGT1 proteins from human and plant , but not yeast . Since both human and plant SGT1 proteins complement ScSgt1p cell cycle mutants in yeast [7] , [9] we hypothesized that the basis for SspH2 and SGT1 interaction is the evolution of NLR co-chaperone function in human and plant SGT1 . Accordingly , SGT1 proteins from plant , human and yeast were assayed for Rx co-chaperone function in planta . HsSGT1A/B and AtSGT1A functionally restored Rx-mediated PVX immunity in planta as indicated by diminished fluorescence ( Fig . 4A ) and little detectable GFP in these infiltration sites ( Fig . 4B ) . By contrast , fluorescence and GFP levels were virtually indistinguishable between ScSgt1p and the GUS negative control ( Fig . 4A and B ) confirming that ScSgt1p does not function as an NLR co-chaperone in planta . Whole leaf infiltrations verified expression of all SGT1 constructs ( Fig . 4C ) . These data support the hypothesis that SspH2 specifically exploits SGT1 NLR co-chaperone functional determinants for binding . Having shown that SspH2 enhances SGT1-dependent NLR signaling in both mammalian and plant model systems , we next examined if SspH2 altered SGT1-dependent cell cycle progression . Yeast provides a robust model with which to study the essential cell cycle function of SGT1 and accordingly SspH2 was expressed in yeast with no apparent affect on cell viability ( Fig . 5A ) . The absence of an SspH2 phenotype in this assay might have been due to the lack of interaction between SspH2 and ScSgt1p ( Fig . 1D ) . Therefore , we re-examined the impact of SspH2 on SGT1 G2 cell cycle function using a characterized ScSgt1p temperature sensitive yeast mutant ( sgt1-3 ) that , unless complemented by a functional SGT1 allele [7] , arrests in G2 resulting in cell death when grown at the restrictive temperature . Expression of either HsSGT1A or HsSGT1B rescued yeast viability of the sgt1-3 mutant at the non-permissive temperature , and this complementation was unaffected by the presence or absence of SspH2 ( Fig . 5B and C ) indicating that SspH2 does not impinge upon the G2 cell cycle function of SGT1 . Similar results were also obtained with the sgt1-5 yeast mutant strain that arrests in the G1 phase of the cell cycle at restrictive temperatures ( data not shown ) . Nevertheless , the yeast sgt1-3 complementation assay has some limitations that warrant caution in the interpretation of these results . For example , the overexpression of human SGT1 might mask any potential SspH2-mediated inhibition of SGT1-cell cycle function . To address this issue SspH2 was transiently transfected into mammalian cultured cells expressing endogenous levels of human SGT1 . We reasoned that if SspH2 impinged on the essential cell cycle function of SGT1 this would manifest as an increased cytotoxic response because SGT1 silencing leads to cell death in mammalian cultured cells [22] . Consistent with the yeast sgt1-3 complementation assay , SspH2 expression in cultured mammalian cells showed no impairment in this assay compared to the empty vector control ( Fig . 5C ) . Thus , these data suggest that SspH2 discriminates between the cell cycle and NLR co-chaperone functions of SGT1 , selectively targeting the latter through interactions dictated by the functional diversification of SGT1 as an NLR co-chaperone . Having observed that SspH2 phenotypes were restricted to the enhancement of SGT1-dependent immune responses we sought to further characterize this phenotype in the Nod1 assay . SGT1 has been reported to bind Nod1 [10] , [11] and given that SspH2 also interacts with SGT1 we examined if SspH2 could interact with Nod1 . We observed that SspH2 , but not SspH1 , could specifically interact with Nod1 via reciprocal co-immunoprecipitation in cell culture lysate ( Fig . 6A ) . Further characterization of this interaction identified a putative ternary complex between SspH2 , Nod1 and SGT1 in cultured mammalian cells ( Fig . 6B ) . SspH2 E3 ubiquitin ligase activity was not required for its interaction with Nod1 since SspH2C580A also co-immunoprecipitated with Nod1 and SGT1 ( Fig . 6B ) . Having identified a putative ternary complex between SspH2 , Nod1 and SGT1 we sought to functionally characterize these interactions . We postulated that this interaction might be the basis for the SspH2-dependent activation of Nod1 , perhaps through the induction of a conformational change in Nod1 . Accordingly , Nod1 activity assays were repeated in the absence of the Nod1 agonist ( iE-DAP/hIFN-γ ) . Intriguingly , even in the absence of agonist , catalytically active SspH2 induced a significant increase in IL-8 secretion in the assay ( Fig . 6C ) . SGT1 was again implicated in this SspH2 phenotype , since SspH2ΔBD did not significantly enhance IL-8 secretion . To study the role of SspH2 catalytic activity on Nod1 activation , IL-8 secretion assays were performed in the presence of the proteasome inhibitor MG132 . Despite a marked decrease in IL-8 secretion , the presence of MG132 did not abrogate the significant increase in IL-8 secretion elicited by SspH2 compared to SspH2C580A ( Figure 6D ) . These data suggest that the proteasomal degradation of an SspH2 substrate ( s ) does not contribute to the enhancement of Nod1 activity . To further characterize the requirement of SspH2 catalytic activity in the enhanced NLR response phenotype , mammalian cultured cells were transiently co-transfected with epitope-tagged SspH2 , Nod1 and ubiquitin constructs , followed by Nod1 immunoprecipitation . Notably , transfection of catalytically active SspH2 , but not SspH2C580A , resulted in the apparent ubiquitination of Nod1 ( Fig . 6E ) . Dual detection of Nod1 and ubiquitin in the Nod1-immunoprecipitate indicated co-migration of the major ubiquitin and Nod1 signals ( indicated by an asterisk in Fig . 6E ) . The apparent molecular weight of the Nod1-ubiquitin species suggests that SspH2 monoubiquitinates Nod1 . The results obtained above indicated that SspH2 required interaction with co-chaperone-competent SGT1 proteins to enhance NLR responses in both plant and animal systems . However , the nature of the functional interaction between SspH2 and SGT1 remained unclear . Given that Nod1 appeared to be ubiquitinated by SspH2 we tested if SGT1 might also be an SspH2 substrate using recombinant proteins in an in vitro ubiquitination assay . SspH2 was catalytically active in this assay with SspH2-dependent ubiquitin polymerization evident after one hour ( Fig . 7A ) . Interestingly , SspH2 E3 ubiquitin ligase activity appeared to be enhanced in the presence of HsSGT1A over the course of the assay as indicated by the intensity of the ubiquitin signal . The enhanced activity was most likely due to increased stability of SspH2 in the presence of HsSGT1A ( Fig . 7A ) . Importantly , SspH2-dependent ubiquitination of HsSGT1A was not observed ( Fig . 7A ) suggesting that HsSGT1A is not a substrate of SspH2 . To further study the apparent stabilization of SspH2 by HsSGT1A , reaction components were pre-incubated together with and without HsSGT1A and initiated by the addition of ubiquitin . Pre-incubation of reaction components in the presence of HsSGT1A showed both increased amounts of SspH2 protein , and increased ubiquitin signal intensity ( Fig . 7B ) . These data are consistent with a model where SspH2 interacts with the host co-chaperone SGT1 in order to stabilize its active conformation . Since SspH2 NEL activity is required for subversion of the immune responses in our assays , these in vitro results offer mechanistic insights into the SGT1-dependence of SspH2 phenotypes .
In this study we have uncovered novel aspects of bacterial effector biology through cross-kingdom analyses of model systems . We report that the S . Typhimurium E3 ubiquitin ligase SspH2 enhances SGT1-dependent innate immune responses in both animals and plants . Our data show a strong physical and genetic interaction between SspH2 and SGT1 that is required for SspH2-mediated modulation of innate immune responses . Furthermore , biochemical assay suggests that the interaction between these proteins increases SspH2 stability and E3 ubiquitin ligase activity . Interaction with SspH2 is mediated by the SGT1 CS and SGS domains , where the CS domain mediates SGT1 co-chaperone function through interaction with Hsp90 and the SGS domain mediates substrate specificity by binding LRR domain-containing proteins , such as NLRs [23] . This suggests that the LRR domain of SspH2 , but not SspH1 or SlrP , may have specifically evolved to mimic an SGT1 client . By contrast , the SGT1 TPR domain , which is dispensable for interaction with SspH2 , has been shown to mediate kinetechore assembly and allow progression through the G2/M checkpoint of the yeast cell cycle [16] . Notably , in the assays tested here , SspH2 did not affect SGT1-dependent cell cycle functions in yeast or mammalian cells . The host SGT1/Hsp90 chaperone machinery represents a unique target for subversion by a bacterial effector . This conserved machinery represents one of the only protein networks that is common to the convergently evolved plant and animal innate immune systems [16] , [24] . Previously it was reported that the plant pathogen Pseudomonas syringae effector AvrB weakly interacted with AtSGT1B , though AvrB function was mediated by the plant-specific co-chaperone RAR1 [25] . Another P . syringae effector , AvrPtoB , showed a genetic , but not physical , interaction with SGT1 and RAR1 , requiring these co-chaperones to suppress plant immunity [26] . In plants , HSP70 has been reported to interact with the SGT1/HSP90 chaperone complex and participate in immune signaling [27] . Yet another P . syringae effector , HopI1 , requires interaction with plant HSP70 to suppress plant immunity presumably by altering HSP70 chaperone dynamics [28] . By contrast , the SspH2 interaction with SGT1 appears to alter the dynamics of effector stability and activity . Since SspH2 catalytic activity is required for its modulation of SGT1-dependent NLR responses , this co-option of the host chaperone machinery may represent a novel mechanism of effector action . Here we report an intriguing functional link between SspH2 and the enhancement of SGT1-dependent innate immunity in animals and plants . In mammalian cell culture , expression of SspH2 significantly increased Nod1-dependent IL-8 secretion . Similar results were found in bacterial infection studies using a complemented sspH2 deletion strain . In that experiment no significant difference in IL-8 secretion was detected between the sspH2 deletion mutant and wild type strain . However , the amount of translocated SspH2 required to elicit the phenotype is unknown and it is possible that this threshold amount is not translocated by the wild type strain . Indeed , it was reported that sspH2 expression in S . Typhimurium was categorized as low when expression of SPI-2 effectors was stratified into very high , high , medium and low categories [29] . In planta SspH2 expression caused enhanced intrinsic HR induced by the SGT1-dependent NLR , Rx . Though unusual that an effector would enhance cell death , it was reported very recently that the Xanthomonas campestris effector , XopL , also an LRR-containing E3 ubiquitin ligase , causes activity-dependent cell death in planta [30] . It is now becoming appreciated that the induction of specific immune responses by bacterial effectors can be beneficial for pathogenic microbes in both plant and animal systems , a concept recently termed ‘effector triggered immune pathology’ ( ETIP ) [31] . In this context the interaction between SspH2 and Nod1 may well represent the first example of direct recognition of an effector in an animal host . ETIP blurs the conventional paradigms of host-pathogen interactions . SspH2 targets SGT1 and induces an immune response from an SGT1-dependent NLR in both plant and animal model systems . Similarly , it was recently reported that the P . syringae effector AvrRps4 targets the A . thaliana immune regulator EDS1 and induces an immune response from the EDS1-dependent NLR , RPS4 [32] , [33] . In the RPS4/EDS1/AvrRps4 pathosystem this was suggested to be the mechanism by which RPS4 recognizes and mounts an immune response to the effector AvrRps4 . This may also be the case with the SspH2/SGT1/Nod1 pathosystem . Alternatively , SspH2 might exploit SGT1 and its interaction with NLRs to specifically induce ETIP . Particularly fascinating is the discovery that SspH2 can induce SGT1-dependent NLR responses across kingdoms . To our knowledge , SspH2 constitutes the first reported bacterial effector to modulate innate immunity in both plant and animal systems . Based upon our findings we speculate that SspH2 enhances Nod1 activity through the interaction and modification of this SGT1-dependent NLR . We identified the formation of a trimeric complex of SspH2 , Nod1 and SGT1 in immunoprecipitates from cultured mammalian cells; however , the role of SGT1 in complex formation should be further studied since both SspH2 and Nod1 interact with the same SGT1 domains [10] , [11] . Catalytically active SspH2 mediated ubiquitination of Nod1 , with the apparent molecular weight suggesting that one ubiquitin molecule was transferred to Nod1 . Monoubiquitination has been shown to impact cellular processes and signaling , rather than causing proteosomal degradation [34] and this is consistent with our finding that the SspH2-mediated enhancement of Nod1 activity is not altered by proteasomal inhibition . To our knowledge , ubiquitination of Nod1 has not been previously reported and this work sets the stage for further research aimed at uncovering the mechanisms underlying Nod1 ubiquitination and its agonist-independent activation .
HEK 293T and HeLa cells were cultured in DMEM containing 10% FBS , 1% GlutaMAX , and 1% non-essential amino acids ( Life Technologies ) at 37°C and 5% CO2 . For immunoprecipitations HEK 293T cells were seeded at 106 cells per 10 cm dish and transfected the next day using 8–12 µg of Ca3 ( PO4 ) 2-complexed DNA . Immunoprecipitations were performed as outlined in [6] . For Nod1 activity assays 6×104 cells were seeded per well of a 24-well dish . Cells were transfected using Fugene HD or XtremeGENE 9 ( Roche ) with a total of 1 µg of DNA [50 ng pcDNA3-Nod1-FLAG ( kindly provided by Dana Philpott , University of Toronto ) , 100–300 ng of effector construct and pcDNA3 . 1 carrier DNA] and stimulated overnight with Nod1 agonist [1 µg/ml C12-iE-DAP ( InvivoGen ) +10 ng/ml human interferon gamma ( AbD serotec ) ] in DMEM supplemented with 0 . 5% FBS on subsequent days . Supernatants were quantified using a human IL-8 ELISA Kit ( BD Bioscience ) . For proteasome inhibition experiments 10 µM MG-132 ( or DMSO vehicle ) was co-administered with Nod1 agonist . Cell lysate was analyzed with anti-HA and anti-FLAG antibodies to confirm expression of effectors and Nod1 , respectively . Cell toxicity was measured using an LDH Assay Kit . HeLa cells for immunofluorescence were plated on sterile 12 mm diameter number 1 . 5 coverslips in 24 well plates at 5×104 cells per well . Forty hours post-transfection cells were fixed for 10 minutes at 37°C in formaldehyde from a 2 . 5% ( w/v ) paraformaldehyde solution prior to permeabilisation and blocking in 0 . 1% ( v/v ) Triton-X 100 , 10% ( v/v ) normal goat serum . Rabbit anti-flag and mouse anti-SGT1 were each used at 1∶500 and alexa-fluor 488 labelled anti-rabbit and alexa-fluor 568 labelled anti-mouse were used at 1∶250 . Stacks were collected at 0 . 2 µm intervals using an Olympus water-immersion N . A . 1 . 2 objective and a Photometrics Coolpix HQ2 camera , using identical exposure times and gain settings . Stacks were processed by iterative deconvolution using a theoretical point spread function , followed by rolling ball background subtraction and intensity correlation analysis to determine colocalisation [35] . HeLa cells were plated at 1×105 cells per well in 24 well plates the day prior to infection . The S . Typhimurium SL1344 sspH2 deletion strain has been described elsewhere [36] . The sspH2 deletion strain was transformed with pWSK29 [37] or psspH2-HA . 1 ( see Text S1 for plasmid construction ) for complementation . HeLa cells were infected with SPI-1 induced , invasive Salmonella as described previously [38] with the following exceptions: DMEM containing 0 . 5% ( v/v ) fetal bovine serum was the medium used during the infection , and the media was changed at 6 h post-infection in order to limit SPI-1-mediated IL-8 secretion . Culture supernatants were collected at 16 h post-infection and stored at −80°C until quantified by ELISA . Agrobacterium tumefaciens C58C1 containing the pCH32 helper plasmid and pBIN derivatives were grown overnight in YEB medium , harvested and induced overnight in 10 mM MgCl2 , 1 mM MES , pH 5 . 6 and 0 . 1 mM acetosyringone ( Sigma ) . Virus-induced gene silencing ( VIGS ) of NbSGT1 was performed on 2–3 week old plants as outlined in [19] . PVX assays were conducted in 3-week VIGS-treated N . benthamiana by agro-infiltration at an OD600 of 1 comprised of SGT1- , effector- and PVX-GFP-expressing strains at an OD600 of 0 . 3 , 0 . 6 and 0 . 001 , respectively , and pBIN61 containing strain as carrier . pBIN61-GUS expressing strain was used as a negative control . VIGS- , AtSGT1- , and GUS-constructs were kindly provided by Ken Shirasu ( RIKEN ) . Agro-infiltrated N . benthamiana was grown for approximately 7 days . PVX data are representative of three or more independent experiments . Leaf lysate was analyzed with anti-GFP , anti-Myc and anti-FLAG antibodies to confirm expression of PVX , SGT1 and effector constructs , respectively . In vitro ubiquitination reactions ( 60 µl ) using purified recombinant GST-SspH2 ( 1 µg ) , His-HsSGT1A ( 1 µg ) , HA-Ubiquitin ( 1 . 5 µg ) , UBE1 ( 0 . 4 µg ) , UbcH5b ( 1 . 5 µg ) ( BostonBiochem ) were performed as in [5] . For pre-incubation experiments GST-SspH2 ( . 4 µg ) was pre-incubated with all reaction components as indicated above ( except ubiquitin ) at 37°C . Reactions were initiated by the addition of HA-Ubiquitin . The reaction buffer was: 80 mM Tris pH 7 . 5 , 50 mM NaCl , 10 mM MgCl2 , 5 mM ATP , 0 . 1 mM DTT . Reaction aliquots were quenched with SDS-PAGE loading buffer . Samples were immunoblotted as outlined above . For Nod1 ubiquitination studies 293T cells were seeded at 4×105 cells/well of a 6-well dish . The following day cells were transfected using XtremeGENE 9 reagent ( Roche ) according to the manufacturer's specifications with ( per well ) 200 ng Myc-ubiquitin construct , 200 or 400 ng Nod1-FLAG construct , 100 or 200 ng HA-SspH2 or HA-SspH2C580A construct , and empty vector carrier DNA to 1 µg total DNA . Cell lysate was harvested after 48 h and was immunoprecipitated with α-FLAG antibody as outlined above . Appropriate yeast strains were grown overnight at 25°C , normalized to OD600 = 0 . 1 , serially diluted five-fold , spotted onto plates and incubated at 25°C or 32°C for 3 days . For additional details refer to Text S1 .
|
Salmonella enterica serovar Typhimurium injects a suite of type III effectors into host cells , but their contributions to pathogenesis remain incompletely understood . We performed cross-kingdom analyses of the S . Typhimurium novel E3 ubiquitin ligase ( NEL ) effector SspH2 using animal , plant and yeast model systems to glean insights into the actions of this effector . We found that SspH2 could discriminately interact with the conserved host protein SGT1 , selectively binding human and plant SGT1 , which correlates with the evolution of an NLR co-chaperone function in innate immunity in these SGT1 proteins . We found that SspH2 enhanced these SGT1-dependent immune responses in both plant and animal model systems if SspH2 was catalytically active and could interact with SGT1 . We also showed that SspH2 interacted with , and modified , the animal NLR Nod1 becoming , to our knowledge , the first effector to directly bind and ubiquitinate this key component of the animal innate immune response . This interaction may underlie the finding that SspH2 activated Nod1 in the absence of its agonist , though this also required SspH2 activity and interaction with SGT1 . Intriguingly , SGT1 increased SspH2 protein stability and activity in vitro providing an example of a bacterial effector that co-opts the host chaperone machinery to subvert innate immunity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"model",
"organisms",
"plant",
"and",
"algal",
"models",
"molecular",
"cell",
"biology",
"cell",
"division",
"immunity",
"innate",
"immunity",
"microbial",
"pathogens",
"yeast",
"and",
"fungal",
"models",
"host-pathogen",
"interaction",
"biology",
"salmonella",
"microbiology",
"saccharomyces",
"cerevisiae",
"bacterial",
"pathogens",
"pathogenesis"
] |
2013
|
The Salmonella Type III Effector SspH2 Specifically Exploits the NLR Co-chaperone Activity of SGT1 to Subvert Immunity
|
Leptospirosis is one of the most important neglected tropical infectious diseases worldwide . Icterohaemorrhagiae has been throughout recent history , and still is , the predominant serogroup of this pathogen in China . However , very little in detail is known about the serovars or genotypes of this serogroup . In this study , 120 epidemic strains from five geographically diverse regions in China collected over a 50 year period ( 1958~2008 ) , and 8 international reference strains characterized by 16S rRNA sequencing and MLST analysis . 115 , 11 and 2 strains were identified as L . interrogans , L . borgpetersenii , and L . kirschneri , respectively . 17 different STs were identified including 69 ST1 strains , 18 ST17 , 18 ST128 , 9 ST143 and 2 ST209 . The remaining 12 strains belonged to 12 different STs . eBURST analysis demonstrated that , among the clonal complexes isolated ( CCs ) , CC1 accounted for 73 . 3% ( 88/120 ) strains representing three STs: ST1 , ST128 and ST98 . ST1 was the most likely ancestral strain of this CC , followed by singleton CC17 ( 17/120 ) and CC143 ( 11/120 ) . Further analysis of adding 116 serogroup Icterohaemorrhagiae strains in the MLST database and studies previously described using global eBURST analysis and MST dendrogram revealed relatively similar ST clustering patterns with five main CCs and 8 singletons among these 244 strains . CC17 was found to be the most prevalent clone of pathogenic Leptospira circulating worldwide . This is the first time , to our knowledge , that ST1 and ST17 strains were distributed among 4 distinct serovars , indicating a highly complicated relationship between serovars and STs . Our studies demonstrated a high level of genetic diversity in the serogroup Icterohaemorrhagiae strains . Distinct from ST17 or ST37 circulating elsewhere , ST1 included in CC1 , has over the past 50 years or so , proven to be the most prevalent ST of pathogenic leptospires isolated in China . Moreover , the complicated relationship between STs and serovars indicates an urgent need to develop an improved scheme for Leptospira serotyping .
Leptospirosis , caused by pathogenic Leptospira species , is emerging as one of the most widespread zoonosis with an estimated global burden of more than 500 , 000 cases of severe human leptospirosis and 100 , 000 deaths as well as great economic burden in farm and pet animals per year [1] . However , its actual prevalence might be still largely underestimated due to a lack of convenient and effective diagnostic approach resulting in underreporting and low awareness in medical and public health communities . The symptom of leptospirosis ranges from an asymptomatic or mild infection to severe manifestation causing multi-organ dysfunction and even deaths in humans [2 , 3] . Humans and animals can be infected through the direct or indirect exposure to urine of infected animals and urine-contaminated water or soil [2 , 4 , 5] . Nowadays , the classical taxonomy typing method of Leptospira spp . is mainly based on serological techniques including microscopic agglutination test ( MAT ) and cross-agglutinin absorption test ( CAAT ) . It is the practical taxon at the subspecies level and remains extremely valuable for epidemiology analysis of Leptospira . However , MAT or CAAT is laborious and time-consuming because these methods require the maintenance of a large range of reference strains and corresponding rabbit antisera . In addition , some serovars were found to have a across reaction . Therefore , MAT or CAAT is no longer sufficient to identify isolates to their species level . Recently , several molecular typing methods have been developed to discriminate Leptospira spp including PCR-restriction endonuclease analysis , pulsed-field gel electrophoresis ( PFGE ) [6–8] , multilocus variable number of tandem repeats analysis ( MLVA ) [9 , 10] . The most commonly used multilocus sequence typing ( MLST ) scheme has been recommended as a routine typing Leptospira species method and population phylogenetic analysis [11–14] . To date , Leptospira genus is now classified into 9 pathogenic , 5 Intermediate and 6 saprophytic species [11 , 15 , 16] . L . interrogans , L . borgpetersenii and L . kirschneri are the main pathogenic species of leptospirosis in humans and animals worldwide . Based on antigenic similarity , more than 300 antigenically related pathogenic serovars are clustered into 24 serogroups in the world , and 75 serovars belonging to 18 serogroups are reported in China . Among them , serogroup Icterohaemorrhagiae is the most predominant epidemic-causing strain in China , and is responsible for more than 60% reported cases of lepotospirosis [17] . However , to date , there is very limited information of the detailed predominant serovars or genotypes of serogroup Icterohaemorrhagiae in China , which plays a crucial role in the epidemiology of leptospirosis . Understanding this role may allow for the development of better control strategies of this disease . The aim of this work was to investigate the genetic diversity of predominantly epidemic serogroup icterohaemorrhagiae of pathogenic Leptospira in Mainland China . Therefore , we investigated the genetic characteristics of 120 serogroup Icterohaemorrhagiae strains isolated from leptospirosis patients or rodent sources in five Chinese provinces with the highest leptospirosis prevalence during the past 50 years by a combination of 16S rRNA sequencing and MLST . Our results could provide a more comprehensive overview of the predominant epidemic serogroup icterohaemorrhagiae in Mainland China and should contribute to understanding the changing epidemiological and evolutionary trends of this serogroup . To obtain a more overview of global population structure and microevolution of serogroup Icterohaemorrhagiae Leptospira strains , the available MLST data from MLST database and some previous studies described from other countries representing the international strains were introduced and further analyzed . The results in this study may be used as markers to trace pathogenic strains isolated from the environment and host in the near future , as well as to obtain a more complete overview of global population structure and microevolution of L . interrogans serogroup Icterohaemorrhagiae strains .
The information of these patients with leptospirosis in this study was anonymously obtained from national infectious disease surveillance system in China; only lots of the patients in the recent years were required to provide brief informed consent before blood sampling . All of the protocols in the study including collection and application of these anonymous serum specimens were conducted with approval by the ethical committee of the Chinese Center for Disease Control and Prevention ( China CDC , Beijing , China ) . A total of 128 non-epidemiologically related leptospiral isolates , including 120 Chinese strains isolated from five provinces and 8 international reference strains from seven countries ( Indonesia , Congo , Denmark , Japan , Zaire , Sri Lanka and Belgium ) were used ( S1 Table ) . The 120 Chinese strains were collected from human or rats over a 50 year period ( 1958~2008 ) . The 8 reference strains ( 56101 , 56102 , 56103 , 56104 , 56108 , 56166 , 56229 and 56233 ) were isolated between 1915~1966 ( except a Japanese strain without detailed information ) . Serogroup identification of these leptospiral strains was carried out by MAT with 15 Chinese standard serogroup-specific rabbit antisera from the National Institutes of Food and Drug Control , China , representing the most predominantly pathogenic serogroups of Leptospira spp . in China . The serogroup scoring the highest MAT titer of the test stains agglutinating 50% of live leptospiral against a given serogroup-specific rabbit antisera is defined as the presumptive corresponding one . All of the 128 strains were maintained by the National Institute for Communicable Disease Control and Prevention , China . Leptospires were stored long-term at −70°C and have been passaged every six months . When needed , they were subcultured at 30°C in 10ml Ellinghausen-McCullough-Johnson-Harris ( EMJH ) liquid medium to stationary phase , and genomic DNA was extracted using NucleoSpin Tissue kits ( Macherey-Nagel , Germany ) according to the manufacturer’s protocol . As a reference method of species identification , 16S rRNA gene sequencing was performed as previously described by Morey [18] for all the 128 epidemic strains . A total of 20 accessible Leptospira species reference sequences that represented pathogenic , intermediate pathogenic and non-pathogenic Leptospira species were obtained from GenBank database and Turneriella parva NCTC 11395T and Leptonema illini NCTC 11301T were set as outgroup ( S2 Table ) [16 , 18] . The sequences of all the Leptospira strains in this study and the 20 representative sequences from GenBank were compared using ClustalW multiple alignments . A Neighbor-joining tree was constructed with Mega software version 5 . 10 with a bootstrap value of 1 , 000 . MLST were performed based on 7 housekeeping genes including glmU , pntA , sucA , tpiA , pfkB , mreA and caiB as previously described [19] . PCR was conducted using the following parameters: an initial denature step at 94°C for 5 min , followed by 30 cycles of 94°C for 30 seconds , 46°C for 30 seconds , 72°C for 45 seconds , then 72°C for 10 min . The PCR products were sequenced by ABI PRISM 377 DNA sequencer . Each allele and the allelic profiles ( glum-pntA-sucA-tpiA-pfkB-mreA-caiB ) were submitted to the established internet Leptospira database ( http://leptospira . mlst . net ) to assign the sequence types ( STs ) . eBURST algorithm ( http://eburst . mlst . net/ ) was applied to determine the relationships among STs . Clonal complexes ( CCs ) were defined as multiple STs linked through single locus variants ( SLVs ) when they differed from each other at a single locus and named on the basis of the putative founder ST or the ST associated with the largest number of SLVs in the clonal complex . Singletons are defined as the STs differing at least three alleles from other STs . Phylogenetic analysis were performed using UPGMA by the BioNumerics software version 5 . 1 ( Applied Maths , Kortrijk , Belgium ) . Furthermore , multiple concatenated sequences of 7 housekeeping alleles were performed using CLUSTALW software and Phylogenetic analysis was conducted with MEGA 5 . 10 [20] . The Neighbor-joining tree was constructed using bootstrapping at 1 , 000 bootstrap replications . To explore the genetic diversity and evolutionary relationship between the isolates in China and other countries , 121 international isolates previously identified by MLST were added into our analysis ( S3 Table ) [21–23] . Among them , a total of 19 international strains belonging to serogroup Icterohaemorrhagiae from 10 countries , including 5 Chinese isolates in present study , were downloaded from the Leptospira MLST website . In addition , 102 sequence data related to Brazil , Argentina and Russia were obtained from three previous studies [21–23] . All of the 121 international strains are listed in S2 Table . The genetic relationship among the 128 isolates in our lab and the 116 isolates from MLST database and previous studies were further analyzed by a minimum spanning tree ( MST ) analysis using the BioNumerics software version 5 . 1 ( Applied Maths , Inc . , Austin , TX , USA ) .
All the 128 strains with the highest agglutinating MAT titer against serogroup Icterohaemorrhagiae of 15 standard serogroup-specific rabbit antisera were confirmed as serogroup Icterohaemorrhagiae in this study . Among the 128 strains , 115 strains were identified as L . interrogans , 11 strains as L . borgpetersenii , and two reference strains isolated from Congo and Zaire as L . kirschneri ( Fig 1 and S1 Table ) . Neighbor-joining trees were constructed for the 128 leptospiral isolates in this study and 20 international representative strains obtained from GenBank database ( Fig 1 ) . Three distinct groups representative of pathogenic , nonpathogenic , and intermediate Leptospira species were obtained . Turneriella parva NCTC 11395T and Leptonema illini NCTC 11301T formed a distinct basal out-group branch . Compared to the 20 representative sequences , 115 including 109 Chinese isolates from the five provinces ( Jiangxi , Sichuan , Anhui , Hunan and Anhui ) and 6 international strains isolated from five countries ( Belgium , Denmark , Indonesia , Japan and Sri Lanka ) were identified as pathogenic L . interrogans . Eleven isolates identified as the pathogenic L . borgpetersenii originated from Jiangxi province in China , and the remaining 2 strains isolated from Congo and Zaire were identified as the pathogenic L . kirschneri . All of the 128 isolates were successfully amplified and sequenced ( S1 Table ) . The discriminatory ability for different species ranged from 0 . 11 ST per isolate for L . interrogans to 1 . 0 ST per isolate for L . kirschneri ( Table 1 ) . Among 120 Chinese Leptospira strains , a total of 10 different STs were obtained , 5 of which were represented by multiple strains , while the remaining 5 STs were found as singleton ( Table 2 and S1 Table ) . The most predominant ST in China was ST1 ( 69/120 ) , followed by ST128 ( 18/120 ) , ST17 ( 17/120 ) , ST143 ( 9/120 ) , ST209 ( 2/120 ) and the remaining 5 isolates belonged to 5 different STs , respectively ( Table 2 ) . The most predominant genotype , ST1 , was temporally ( between 1958 and 2008 ) and geographically diverse ( 4 provinces distributed in Sichuan , Jiangxi , Anhui , Hunan ) . Furthermore , the distributions of STs in China were associated with special geographic regions . For example , ST17 , the second most frequent serotype , was found in Sichuan and Jiangxi provinces between 1969~2008 and ST128 was just found in Hunan province in 2007 . In addition , ST143 and ST209 were found in Jiangxi province between 2005~2007 . It was interesting that only ST143 and ST209 corresponded to L . borgpetersenii and all the other STs corresponded to L . interrogans in China . Whereas eight different STs were identified among 8 international strains , only ST17 was found in Chinese Leptospira isolates ( Table 2 and S1 Table ) . eBURST analysis based on the allelic profiles was first conducted to identify relationships among 10 Leptospira STs found in the 120 Chinese pathogenic strains . Clonal complexes ( CCs ) based on ST Linkages were built using the criteria of at least five shared alleles . Two CCs ( CC1 and CC143 ) and 5 singletons were identified , including the largest CC1 and the largest singleton CC17 ( S1 Fig and S1 Table ) . The CC1 and 5 singletons belonged to L . interrogans , whereas CC143 belonged to L . borgpetersenii . The CC1 contained 69 ST1 strains , 18 ST128 strains and 1 ST98 strain with ST1 being the most likely ancestral strain of this CC . The CC143 , including 9 ST143 strains and 2 ST209 strains , showed no predicted founder type . eBURST analysis has confirmed that there is no coexistence of different species within the same CCs . On the other hand , the relationships between the 10 STs representing 120 Chinese strains were depicted in a UPGMA dendrogram . Three main clades ( Clade1-3 ) were generated and the remaining isolates were dispersed as unrelated singletons ( Fig 2 ) . The UPGMA dendrogram revealed ST clustering patterns relatively similar with eBURST analysis ( Fig 2 and S1 Fig ) . The three clades in UPGMA dendrogram corresponded to the CC1 , CC143 and one Singleton CC17 in eBURST dendrogram , respectively . The rest of the strains were dispersed as unrelated singletons like the ones in eBURST dendrogram . The Clade2 corresponding to CC1 was found among four provinces ( Sichuan , Jiangxi , Hunan and Anhui ) and no relationship was observed between the isolates . Furthermore , the UPGMA dendrogram had shown a geographical relationship between the isolates and STs . For instance , The Clade3 corresponding to CC143 contained 11 ST143 and 2 ST209 strains isolated from Jiangxi province , The Clade1 corresponding to singleton CC17 contained 17 ST17 strains from Sichuan and Jiangxi provinces between 1969~2008 and one SLV of ST128 in Clade2 contained 18 strains isolated from Hunan province . Besides the 128 strains in this study , additional 116 serogroup Icterohaemorrhagiae isolates with diverse geographical regions or countries from MLST database and other studies were also added to perform MLST analysis . However , among the finally identified 22 STs from these 244 strains , only ten STs were found in China . The eBURST analysis revealed five CCs and 8 singletons ( S2 Fig ) . CC17 remained to be the most predominate CC which covered 125 strains corresponding to three different STs ( ST17 , ST199 and ST206 ) and followed by the CC1 including 89 strains corresponding to another three STs ( ST1 , ST128 and ST98 ) . The third largest CC143 included 11 ST143 strains and 2 ST209 strains isolated from China . ST1 and ST17 were defined as the predicted founders of CC1 and CC17 , respectively . The remaining three CCs comprised relatively dispersed STs with no predicted founding type . For instance , CC38 included two relatively distant STs: ST203 and ST38 . The geographic distribution and corresponding STs among the 244 international Leptospira isolates are listed in Table 3 . Generally close clustering of these strains from same geographical regions was observed . For example , 9 ( ST1 , ST92 , ST98 , ST128 , ST143 , ST209 ST201 , ST202 and ST203 ) of 10 STs were found exclusively in China between 1958~2008 . And all of the Brazil , Argentina and Belgium isolates belonged to ST17 , all nine Russia isolates and two Denmark isolates were clustered together into CC17 . The remaining isolates from Japan , Malaysia , Sri Lanka , and Indonesia were classified as relatively independent singletons . Therefore , ST17 was found as one of the most common STs worldwide , including in Asia ( China , Japan ) , Latin America ( Brazil and Argentina ) and Europe ( Denmark , Belgium and Russia ) between 1915~2009 . At the same time , the 244 strains were further analyzed by minimum spanning tree ( MST ) . Five CCs ( CC1 , CC17 , CC143 , CC38 and CC65-122 ) were generated and the remaining isolates were dispersed as unrelated singletons ( Fig 3 ) . The MST dendrogram showed relatively similar ST clustering patterns with eBURST analysis ( Fig 3 and S2 Fig ) . In some cases , isolates within same CCs were generally restricted to one or several countries . For instance , the CC1 and CC143 only contained these Chinese strains which had closely genetic relationship but were distant from all other isolates , whereas , the other three CCs included isolates from more than one country . For example , The CC17 included 120 clustered isolates from Asia ( China , Japan ) , Latin America ( Brazil and Argentina ) and Europe ( Denmark , Belgium and Russia ) . CC38 included 3 clustered isolates from China and Sri Lanka . The CC65-122 included 6 clustered isolates from Africa ( Zaire , Congo ) and Latin America ( Jamaica ) corresponding 4 different STs . The remaining isolates were dispersed as singletons . Based on the MST dendrogram , more than half of Chinese strains were clustered into three large CCs: CC1 ( 88/120 ) , CC143 ( 11/120 ) and CC17 ( 17/120 ) . The remaining 4 isolates from China were dispersed as 4 independently singletons . As seen in Table 3 and Fig 3 , the genetic diversity of Leptospira strains belonging to serogroup Icterohaemorrhagiae from China was generally different from that of isolates elsewhere . From the global population , no common CCs with potential founders were identified as a whole distribution , indicating high diversity of STs . Based on seven MLST housekeeping genes , Neighbor-joining trees were constructed with three distinct clusters corresponding to three different Leptospira species ( Fig 4 ) . The L . interrogans cluster containing 6 international strains and 109 Chinese strains . The L . borgpetersenii cluster containing 11 strains were further divided into two sub-groups that originated from Jiangxi province between 2005~2007 . In addition , the L . kirschneri cluster containing 2 strains originating from Congo . Phylogenetic analysis revealed relatively similar species clustering patterns with 16S rRNA gene sequencing . Among the 120 Chinese strains in this study , 31 isolates with previously confirmed serovar information were utilized to investigate the relationships between serovars and STs ( S4 Table ) . It was found that there were some isolates in same STs generally corresponding to two or more different serovars . 12 STs contained strains in a single serovar ( S4 Table ) . However , for the first time , we reported that ST1 strains distributed among 4 serovars: Lai , Naam , Liangshan and Honghe , and similarly ST17 corresponded to serovar Icterohaemorrhagiae , Lai , Copenhageni , Renshou and Smithi . In addition , interestingly , some serovars were also found to correspond to multiple STs . For example , serovar Copenhageni was found among three STs-ST17 , ST199 and ST201 . Serovar Honghe was also associated with ST1 , ST92 , ST98 and ST203 . Serovar Lai was associated with ST1 and ST17 and serovar Naam was associated with ST1 and ST23 . These observations have shown that the relationship between serovars and STs was highly complicated , suggesting serovar classification as a poor indicator of genetic relatedness .
Although the incidence of leptospirosis has significantly decreased in the past few years , leptospirosis is still considered as an important zoonosis in China . Since 2004 , leptospirosis was routinely included in the national epidemic surveillance system that included systematic case reporting and the monitoring efforts aimed at environmental and host animal populations such as pigs , dogs , cattle and rats . The southern provinces of Mainland China had the highest leptospirosis prevalence rates in recent years . A recent report indicated that serogroup Icterohaemorrhagiae has been historically the most prevalent serogroup associated with human and animals outbreaks in China [17] . MAT has been performed only in a limited number of reference laboratories , primary due to the requirement of long-term maintenance of large range of reference strains and serogroup or serovar-specific standard anti-rabbit sera . In addition , pathogenic Leptospira spp . include more than 230 serovars , the majority of them have no corresponding specific antisera and cannot be identified by MAT . On the contrary , MLST has a higher discriminatory power among Leptospira spp . and is widely used for bacterial genotyping [24] , including Leptospira [12 , 13 , 19 , 25] . 16sRNA sequencing used as a tool for phylogenetic analysis has led to a better understanding of evolution of Leptospira . These two techniques can be directly applied to biological ( serum , urine or blood of maintenance hosts and human ) and environmental samples . Furthermore , MLST is also supported by an updated website at http://leptospira . mlst . net/ , which helps to exchange of new information among laboratories or countries . This would allow for epidemiological studies in some laboratories where they are not able to culture Leptospira spp . So far , no detailed studies focusing on the major prevalence and the genetic characterization of leptospirosis disease are available . To investigate the genetic diversity of leptospirosis , a total of 120 Chinese strains and 8 international reference strains belonging to serogroup Icterohaemorrhagiae were analyzed using 16S rRNA gene sequencing and MLST analysis . These isolates were primarily obtained from leptospirosis patients , or a wide range of rodent sources from five major provinces known to have a high incidence of leptospirosis in China . All the 120 strains in this study were differentiated effectively as indicated by clustering patterns ( Fig 1 ) . Two different pathogenic species of L . interrogans , L . borgpetersenii were identified , which was in agreement with previous studies in China [17 , 26] . Among the 120 Chinese isolates , L . interrogans accounted for 90 . 83% ( 109/120 ) ; this has been the predominant pathogenic species of leptospirosis in China over the last fifty years ( 1958–2008 ) . These findings were in agreement with previous studies conducted in Guizhou province [27 , 28] . 9 . 17% ( 11/120 ) strains isolated in Jiangxi province between 2006~2007 in China were identified as L . borgpetersenii . One previous report found that Icterohaemorrhagiae was the serogroup in 51 L . interrogans and L . kirschneri strains isolated from a variety of sources and geographical areas in France [25] . In addition , 43 L . interrogans were uncovered in three outbreaks in Brazilian urban centers [29] . Serovar Copenhageni accounted for 87% of L . interrogans cases in another large urban outbreak in Brazil [30] . The predominant pathogen species isolated in Mayotte were L . borgpetersenii and L . kirschneri [31] . Thaipadungpanit et al . in 2007 had demonstrated that ST34 , corresponding to L . interrogans serovar Autumnalis , accounted for 76% of isolates in 101 L . interrogans isolates in Thailand [12] . Together , these data revealed that the major Leptospira species studied here from different counties were distinct and that the great genetic diversity in geographic epidemiology shown by these isolates reflected this observation . [13 , 19 , 23 , 25] . When compared with L . borgpetersenii , the two species of L . interrogans and L . kirschneri seem to have more closely related to one another and probably evolved from the L . noguchii clade . Close phylogenetic relationships between L . interrogans , L . kirschneri and L . noguchii were reported by Ahmed et al . based on MLST phylogenetic analysis [13] . Furthermore , the UPGMA dendrogram revealed relatively similar ST clustering patterns with eBURST analysis , 2 CCs ( CC1 and CC143 ) and 5 singletons were clustered in 120 Chinese strains ( Fig 2 and S1 Fig ) . CC1 consisted of 3 different STs ( ST1 , ST128 and ST98 ) from diverse sources over the past 50 years in China , with ST1 as the likely founder . Our results were similar with those of previous studies performed in Guizhou province [32] . The predominant serotype in China , ST1 , was widely distributed ( 37 . 68% ( 26/69 ) Sichuan province; 28 . 99% ( 20/69 ) Jiangxi province; 28 . 99% ( 20/69 ) Anhui province and 4 . 35% ( 3/69 ) Hunan province ) . The host ranges were 68 . 12% ( 47/69 ) in Apodemus agrarius , 13 . 04% ( 9/69 ) in human , 13 . 04% ( 9/69 ) in Rattus rattoides and 13 . 04% ( 9/69 ) in Rattus norvegicus during the 1958~2008 collection time span . In addition , ST17 , widely distributed in Sichuan and Jiangxi provinces , was the second most common serotype isolated during this time period . In general , the predominant serotypes , ST1 and ST17 , having distinct sources yet formed a tight group , indicating that there might be one original strain which subsequently diverged evolutionarily into the two STs above within southern China provinces . The remaining 4 singletons of ST92 , ST201 , ST202 and ST203 shared no more than 3 strains , and presumably dispersed independently . The genotyping results from this study showed that Apodemus agrarius could be a main source of leptospirosis transmission in China . MLST is also useful to explore the transmission of specific species between maintenance animal hosts and human . Therefore , it may be useful to implement control strategies for Apodemus agrarius to reduce the transmission from animals to humans . Interestingly , although serogroup Icterohaemorrhagiae strains were found in most Leptospira endemic regions in China and some STs such as ST1/ST17 were widely distributed in this study , there were some prominent serogroups consisting of more than one ST/species in specific regions . For example , CC143 , belonging to L . borgpetersenii , comprised of 11 strains isolated from Rattus rattoides and Rattus norvegicus . CC143 was locally confined to Jiangxi province between 2006~2007 in China . One SLV of ST128 comprised of 18 strains belonging to L . interrogans , and was isolated from Apodemus agrarius and Rattus rattoides hosts in Hunan province in 2007 . Therefore , some clustering of strains from specific geographical regions was observed in China . The strains isolated from Jiangxi and Hunan provinces were from the same monitoring sites , respectively , suggesting that the isolates may have an epidemiological link in that given locale . This also gives a clue that the MLST scheme is capable of dissecting the molecular geographic epidemiology of leptospirosis . No other obvious epidemiological relationship was found between STs and source specimens or isolated locations in these 120 Chinese strains . These results were not surprising because these isolates were epidemiologically unrelated and showed a great diversity in STs; no clustering was detected . More isolates and molecular typing data are needed in order to better understand the epidemiology of leptospirosis in China . Basis on the genotyping results in this study , certain Leptospira genotypes are prevalent in a particular geographical region and associated with special animal reservoirs . The diverse distributions of genotypes may provide a clue for species-specific vaccine preparation to increase the efficacy of a vaccination program in different epidemic regions . This information may also be useful for tracking of the source of leptospirosis outbreak and to establish a control program against leptospirosis in each region . In order to explore the global genetic diversity and evolutionary relationships in the serogroup Icterohaemorrhagiae strains worldwide , a total of 244 serogroup Icterohaemorrhagiae strains from 13 different countries were analyzed and 22 STs were found . The MST dendrogram revealed relatively similar ST clustering patterns with eBURST analysis; five CCs ( CC1 , CC17 , CC38 , CC143 and CC65-122 ) and 8 singletons were clustered in 244 international strains ( Fig 3 and S2 Fig ) . CC1 and CC143 were the dominant clones in China; these two CCs shared a close genetic relationship and were distant from all the other global isolates . The other 3 CCs , on the other hand , included isolates from more than one country . The predominant ST recovered in Asia , Latin America and Europe between 1915~2009 was ST17 . Furthermore , the remaining isolates were dispersed in 8 unrelated singletons . In addition , the eBURST and MLST analyses revealed that the genetically diverse species/strains of serogroup Icterohaemorrhagiae isolates from China was generally different from those isolated in other countries belonging to that particular serogroup . The remaining 9 STs were found in China exclusively , which may indicate that Leptospira may evolve according to different locations and the epidemiology of leptospirosis in China is relatively independent from other countries . This also indicated that MLST is a useful technique to explore the genetic diversity and molecular epidemiology of leptospirosis on a global and/or historical scale . What is more , Thaipadungpanit et al . had applied MLST typing scheme to 101 L . interrogans isolates and 12 STs were identified in Thailand in 2007 . Among the 12 STs found , ST34 , corresponding to L . interrogans serovar Autumnalis , accounted for 76% of isolates [12] . Moreover , Caimi et al . demonstrated that ST37 corresponded to two serogroups of Pomona and Canicola , and was the most frequent genotype in 18 isolates in Argentina . All the 3 serogroup Icterohaemorrhagiae strains isolated between 1993~2005 were identified as ST17 [21] . Among 11 serogroup Icterohaemorrhagiae strains in Russia , four STs ( ST17 , ST199 , ST23 and ST206 ) were found [22] . It was previously reported that Icterohaemorrhagiae was the most prevalent serogroup in Brazil [23 , 33] , and all the 90 serogroup Icterohaemorrhagiae strains isolated between 1986~2009 in the state of Sao Paulo were identified as ST17 [23] . In all , it was indicated that the predominant serogroups or STs were different in different geographical regions of the world . Whereas , ST17 was the most predominant ST in serogroup Icterohaemorrhagiae in Argentina , Russia and Brazil and ST1 was the most frequent ST in serogroup Icterohaemorrhagiae in China irrespective of the scattered spatial and geographic distribution . These predominant isolates are likely to have adaptive selective advantages in the environment or in maintenance hosts , allowing them to develop into pathogenic strains . Based on concatenated sequences of the 7-locus MLST scheme , 128 strains were differentiated effectively into three distinct clusters corresponding to three species , L . interrogans , L . kirschneri and L . borgpetersenii ( Fig 2 ) by Phylogenetic analysis . This is consistent with previous studies that MLST allowed differentiation of the major pathogenic species of Leptospira [13 , 19 , 23 , 25] . The Neighbor-joining tree revealed the phylogenetic relationship between these three different pathogenic species in this study and had shown that two pathogenic species of L . interrogans and L . kirschneri seem to be more closely related than L . borgpetersenii , which was also confirmed using 16S rRNA sequencing in this study . The close genetic relationship of L . interrogans and L . kirschneri was also confirmed by Boonsilp et al [11] . From the Phylogenetic analysis of MLST data , the Leptospira strains belonging to the same serovars were not clustered together . This was also confirmed in earlier findings [12 , 13 , 19 , 23] . This may be due to horizontal gene transfer . Therefore , the MLST method is an alternative suitable method to identify Leptospira up to genome species level . To explore the relationship between STs and serovars , 31 isolates that had both STs and serovar designations in this study were analyzed . We found that there were some isolates belonged to the same STs , but generally corresponded to different serovars . On the other hand some serovars usually were associated with more than one different ST . These observations have shown that serovars are not suitable indicators of genetic relatedness . The diversity of serovars is most likely to be due to horizontal gene transfer events , leading to differences in sequences . Here our focusing on serogroup Icterohaemorrhagiae strains of Leptospira using MLST analysis and 16sRNA gene sequencing as a tool for phylogenetic analysis has led to a better understanding of evolution of Leptospira . MLST provides evidence that the diversity of STs among the serogroup Icterohaemorrhagiae strains is very high in China . The result may be useful to develop a strategy and/or guidelines for the control of leptospirosis in China . However , phylogenetic analysis of more globally dispersed Leptospira strains is necessary; we nonetheless believe that our present study provides a blueprint for further phylogenetic research . More convenient molecular techniques have to be developed to identify and characterize Leptospira species and STs .
|
Leptospirosis , caused by pathogenic Leptospira spp , is a globally widespread zoonosis . In this study , our focusing on serogroup Icterohaemorrhagiae strains of Leptospira using MLST as a tool for phylogenetic analysis that has led to a better understanding of evolution of Leptospira . This totally consisted of 120 epidemic strains from five geographically diverse regions were isolated over the past 50 years in China and 8 strains from seven different countries . 17 STs were identified in these 128 strains by MLST analysis . Adding 116 serogroup Icterohaemorrhagiae in the Leptospira MLST database and studies previously described , 22 STs were identified in the 244 isolates . The genetic diversity of Leptospira belonging to serogroup Icterohaemorrhagiae from China was generally different from that of isolates elsewhere . Results of the 16S rRNA sequencing typing and MLST genotyping method were nearly consistent . Here , MLST revealed the high diversity of STs among the serogroup Icterohaemorrhagiae strains in China . Our present study provides a blueprint for further phylogenetic research . More convenient molecular techniques have to be developed to identify and characterize Leptospira species and STs .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Molecular Typing of Pathogenic Leptospira Serogroup Icterohaemorrhagiae Strains Circulating in China during the Past 50 Years
|
Simian immunodeficiency viruses of sooty mangabeys ( SIVsm ) are the source of multiple , successful cross-species transmissions , having given rise to HIV-2 in humans , SIVmac in rhesus macaques , and SIVstm in stump-tailed macaques . Cellular assays and phylogenetic comparisons indirectly support a role for TRIM5α , the product of the TRIM5 gene , in suppressing interspecies transmission and emergence of retroviruses in nature . Here , we investigate the in vivo role of TRIM5 directly , focusing on transmission of primate immunodeficiency viruses between outbred primate hosts . Specifically , we retrospectively analyzed experimental cross-species transmission of SIVsm in two cohorts of rhesus macaques and found a significant effect of TRIM5 genotype on viral replication levels . The effect was especially pronounced in a cohort of animals infected with SIVsmE543-3 , where TRIM5 genotype correlated with approximately 100-fold to 1 , 000-fold differences in viral replication levels . Surprisingly , transmission occurred even in individuals bearing restrictive TRIM5 genotypes , resulting in attenuation of replication rather than an outright block to infection . In cell-culture assays , the same TRIM5 alleles associated with viral suppression in vivo blocked infectivity of two SIVsm strains , but not the macaque-adapted strain SIVmac239 . Adaptations appeared in the viral capsid in animals with restrictive TRIM5 genotypes , and similar adaptations coincide with emergence of SIVmac in captive macaques in the 1970s . Thus , host TRIM5 can suppress viral replication in vivo , exerting selective pressure during the initial stages of cross-species transmission .
The Simian immunodeficiency viruses ( SIVs ) are widespread among African primates [1] . However , host and viral phylogenies are not completely congruent; such a pattern argues against co-divergence of virus and host lineages since the time of a common , infected primate ancestor and argues instead that the modern distribution of SIVs among extant primates resulted , at least in part , from cross-species transmission events followed by emergence of new virus/host combinations [2] . The most notable examples include cross-species transmission of SIV from apes to humans , which gave rise to HIV-1 and initiated the worldwide AIDS epidemic , and cross-species transmission of SIV from sooty mangabeys ( SIVsm ) to humans , which gave rise to the more limited HIV-2 epidemic [1] , [3] , [4] . In a striking parallel to the emergence of HIV-1 and HIV-2 , SIVsm also jumped into captive Asian macaques in the United States , resulting in emergence of SIVmac and outbreaks of AIDS-like disease at several U . S . National Primate Research Centers in the 1970s [3] , [5] , [6] . The exact time and means by which SIVsm was transmitted to macaques are unknown , but since isolation of the first SIV strains from captive macaques in the 1980s , experimental infection of rhesus macaques with SIV has become the primary animal model for preclinical research on AIDS vaccines and pathogenesis . Variation in susceptibility to infection and disease progression in nonhuman primate models often confounds such studies , and identifying the sources of variation will lead to more efficient use of AIDS models . At the same time , genetic variation in nonhuman primate hosts of SIV provides unique and powerful opportunities to study the impact of host genetics on cross-species transmission , adaptation , and emergence of viruses . In the present study , we establish that allelic variation in the rhesus macaque TRIM5 gene results in differences in susceptibility to infection and viral replication in the early stages of cross-species transmission of SIVsm and that emergence of pathogenic SIVmac in rhesus macaques required adaptations in the viral capsid protein ( CA ) to overcome suppression by two distinct types of TRIM5 allele . Viral inhibition by TRIM5α , a product of the TRIM5 gene , is initiated by an interaction between the protein's B30 . 2/SPRY domain and virion capsid-cores released into the target-cell cytoplasm after viral attachment and entry [7]–[13] . The target of TRIM5α involves the N-terminal domain ( NTD ) of the viral CA protein [14] . We previously discovered that rhesus macaque TRIM5 is highly polymorphic , including eight nonsynonymous polymorphisms tightly clustered in the B30 . 2/SPRY domain [15] . One of these , a six-nucleotide insertion/deletion , results in a TFP/Q length polymorphism . When tested against multiple lentiviruses , TFP339-341 and Q339 alleles ( TRIM5TFP and TRIM5Q ) give different patterns of restriction [16] . An unusual haplotype encoding a TRIM5-cyclophilin-A chimera ( TRIM5CypA ) is also found among rhesus macaques [17] , [18] . The chimeric TRIM5CypA lacks a B30 . 2/SPRY domain and in its place encodes a CypA domain derived from a retrotransposed CypA reading frame inserted in the 3′UTR [17]–[21] . TRIM5CypA restriction of lentiviral infection involves specific binding to a peptide loop between helices 4 and 5 of the viral CA ( also the binding site for cellular CypA ) [22] , [23] .
Common variants of rhesus TRIM5 can be grouped into three allelic classes: TRIM5CypA , TRIM5TFP , and TRIM5Q ( Figure 1 ) [16] , [24] . We established the existence of all six possible genotypes in a large colony of captive rhesus macaques , using archived genomic DNA samples from the Genetics Core of the New England Primate Research Center . In this colony , we observed frequencies of 46% ( TRIM5TFP/TFP ) , 36% ( TRIM5TFP/Q ) , 5% ( TRIM5TFP/CypA ) , 10% ( TRIM5Q/Q ) , 1% ( TRIM5Q/CypA ) , and 2% ( TRIM5CypA/CypA ) . These values indicate allele frequencies in one particular colony of rhesus macaques; because of differences in animal husbandry practices and potential founder effects , these values do not necessarily reflect the distribution of genotype frequencies within other captive colonies or in wild rhesus monkey populations . Nonetheless , the presence of allelic variation in the rhesus TRIM5 gene can be exploited to study the impact of TRIM5 expression in vivo . To ask whether TRIM5-mediated restriction plays a role in cross-species transmission and emergence of primate lentiviruses , we tested six representative alleles of rhesus macaque TRIM5 for restriction activity against four closely related viruses of old-world monkeys , SIVmac239 , SIVsmE543 , SIVsmE041 , and SIVstm/37 . 16 ( Table 1 ) . SIVmac239 is a molecular clone of a highly adapted , emergent virus of rhesus macaques [25] , generated in the 1980s by experimental passage of SIV-positive plasma through a series of five monkeys [26] . In all likelihood , SIVmac239 is descended from a cross-species transmission event that took place in the 1960s in captive colonies of rhesus macaques [5] , probably as the unintended consequence of experiments involving transfer of biological material from SIVsm-positive sooty mangabeys to rhesus macaques [27] . Regardless of origin , as a result of this long association with macaques , experimental infection with SIVmac239 reproducibly results in high levels of persistent viral replication [28] . In contrast , SIVsmE543-3 is a molecular clone derived by intentional inoculation of a rhesus macaque with plasma from an SIVsm-infected sooty mangabey , followed by passage through one additional rhesus macaque [29]; thus , opportunity for SIVsmE543-3 to adapt to macaques was limited to only two animals . As a result , SIVsmE543-3 replication in macaques is highly variable , with acute viral loads ranging from 103 to 108 viral RNA copies/ml plasma , and set-point values from <100 to 108 . Variation in SIVsmE543-3 infected animals is consistent with an influence of genetic variation in a host gene or genes [30] . SIVsmE041 is a biological isolate cultured directly from an SIV-positive sooty mangabey [31] and has therefore not experienced any prior adaption to rhesus macaques . SIVstm/37 . 16 is an SIV isolate from a different species , the stump-tailed macaque ( M . arctoides ) , and represents an independent cross-species transmission event involving transmission of SIVsm directly to M . arctoides animals [3] , [27] , [32] , [33] . The relevant properties of these four viruses are summarized in Table 1 . Infectivity of all four viruses was measured on cell lines stably expressing six common alleles of rhesus macaque TRIM5 , including three TRIM5TFP alleles , two TRIM5Q alleles , and the TRIM5CypA allele ( Figure 2 ) . SIVmac239 was resistant to all six . SIVsmE041 , SIVsmE543 , and SIVstm were also resistant to TRIM5Q but unlike SIVmac239 were sensitive to both TRIM5CypA alleles and TRIM5TFP alleles ( Figure 2B , C , D ) . Of the four SIV strains tested , only SIVmac239 is the product of decades of replication and spread in rhesus macaques . Thus , the comparison suggests that sensitivity to TRIM5TFP and TRIM5CypA alleles represents the ancestral phenotype and that emergence of SIVsm ( as SIVmac ) in rhesus macaques required acquisition of adaptive changes to overcome those particular types of alleles . In contrast , the SIVsm variants that first invaded rhesus macaques were probably inherently resistant to TRIM5Q alleles . To analyze the impact of TRIM5 variation on cross-species transmission directly , we acquired samples from two independent SIVsm/macaque cohorts . The smaller cohort consisted of four sooty mangabeys and four rhesus macaques that had been experimentally inoculated with SIVsmE041 [34] . Infection of sooty mangabeys ( natural hosts of SIVsm ) resulted in ready take of virus and persistent infection ( unpublished data ) , whereas infection of macaques resulted in a transient infection followed by a decrease in viral replication to a point near or below the detection limit ( Figure 3 ) . Using archived DNA , we determined that the four macaques included three TRIM5TFP/TFP homozygotes and one TRIM5TFP/CypA heterozygote . Thus , the alleles present in these four animals were all of the same types that restricted SIVsmE041 in tissue culture ( Figure 2B ) . In one animal , replication resurged during the first year , reaching ∼104 RNA copies/ml plasma ( Figure 3 ) . Capsid sequences recovered from this animal revealed the appearance of two fixed changes , R97S and V108A ( SIVmac239 numbering ) at the late time point ( Figure 3C ) . In contrast , amplification and sequencing from acute infection , from the SIVsmE041 inoculum , and from an infected sooty mangabey , revealed only the ancestral states at both positions ( R97 and V108 ) ( Figure 3B ) . The second and larger cohort consisted of historical samples from 44 SIVsmE543-3-infected rhesus macaques . Genotype frequencies in this cohort were 30% TRIM5TFP/TFP , 23% TRIM5TFP/Q , 26% TRIM5TFP/CypA , 12% TRIM5Q/Q , 9% TRIM5Q/CypA , and 0% TRIM5CypA/CypA . Animals with two restrictive alleles ( TRIM5TFP/TFP and TRIM5TFP/CypA ) had dramatically diminished viral replication compared to TRIM5Q/Q homozygotes , with mean ( geometric ) differences of 830-fold and 1 , 728-fold , respectively , by 8 wk post-infection ( Figure 4 ) . Animals with one restrictive allele ( TRIM5TFP/Q and TRIM5CypA/Q heterozygotes ) displayed intermediate levels of viral replication . Taken in conjunction with the clear differences in restriction of SIVsmE543-3 by TRIM5TFP , TRIM5CypA , and TRIM5Q in vitro ( Figure 2C ) , these results are consistent with allelic variation in TRIM5 having a significant impact on SIVsmE543 replication kinetics in rhesus macaques . We also obtained archived DNA from animal #E543 , the source of the original SIVsmE543-3 clone [29] , and determined that this animal had been a TRIM5Q/Q homozygote . The fact that E543 bore a non-restrictive genotype may well have facilitated isolation of the original SIVsmE543-3 clone . Several reports describe a correlation between specific alleles of class I MHC and enhanced control of SIVmac239/SIVmac251 infection in rhesus macaques [35]–[37] . Specifically , MHC class-I Mamu-B*08 and Mamu-B*17 alleles have been associated with lower levels of chronic phase viral replication in SIVmac239-infected animals [35] , [37] . Several observations argue against the dramatic differences in viral replication of SIVsmE543-3 in rhesus macaques being due to Mhc class I rather than TRIM5 . First , the TRIM5 and class I MHC loci are on different chromosomes , reducing the probability of a chance association between suppressive alleles of TRIM5 and a specific allele or alleles of class I MHC with activity against SIVsmE543-3 . More importantly , the effects of class I Mhc on SIVmac239 replication do not manifest during the acute stage of infection [38] , whereas the correlation with TRIM5 genotype is already apparent during acute infection ( Figure 4D ) . Finally , Goldstein et al . demonstrated that variation in susceptibility of cells taken from naïve animals ( prior to infection ) and tested ex vivo correlated with viral replication levels in vivo , when the same animals were subsequently infected with SIVsmE543; such results argue in favor of an inherent , genetic cause of variation in susceptibility and against an effect of virus-specific adaptive immune responses induced by infection [30] . Finally , we typed all 43 animals in the SIVsm543-3 cohort for the presence of the Mamu-B*08 and Mamu-B*17 alleles and found no association between these alleles and the observed differences in Figure 4 . Most importantly , none of the infected animals with TRIM5Q/Q or TRIM5TFP/CypA genotypes were Mamu-B*08 or Mamu-B*17 positive , so the statistically significant differences between those two groups cannot be attributed to Mamu-B*08 or Mamu-B*17 associated control of SIVsmE543 . Thus , the observed correlation between TRIM5 genotype and SIVsmE543-3 replication levels is not due to a spurious association between suppressive alleles of TRIM5 and class I MHC alleles previously associated with control of SIVmac239 . It is important to note , however , that this result does not rule out a general influence of class I Mhc on viral replication levels in rhesus macaques , only that MHC genotype does not explain the correlation depicted in Figure 4 . Among other things , allelic variation in MHC class I may well contribute to the significant variation observed within groups ( Figure 4C , D , E ) . Three macaques in the SIVsmE543-3 cohort also had patterns of resurgent viral replication consistent with escape from suppression ( Figure 5A ) . All three animals had two restrictive alleles and included one TRIM5TFP/TFP homozygote and two TRIM5TFP/CypA heterozygotes . We amplified gag sequences encoding the NTD of CA from all three animals , and from a TRIM5Q/Q animal , and compared these to the original SIVsmE543-3 clone ( Figure 5B ) . Strikingly , an R97S change was present in every clone from all three animals , typically due to an AGA->AGC substitution , although 3/15 clones in one animal were AGA->AGT . No changes were found in the corresponding region of virus from the non-restrictive TRIM5Q/Q animal . Thus , an identical R97S change appeared independently in four animals with suppressive TRIM5 genotypes , including three SIVsmE543-infected animals ( Figure 5B ) and one SIVsmE041-infected animal ( Figure 3 ) . Phylogenetic analyses are consistent with a minimum of two historical transmissions of SIVsm into macaques , one into stump-tailed macaques ( SIVstm ) , and the other into rhesus macaques ( SIVmac ) ; both transmissions are thought to have occurred in captive macaques sometime prior to the 1970s [3] . If TRIM5-mediated restriction influences cross-species transmission of primate lentiviruses , we predict the existence of adaptive changes in SIV isolates corresponding to such events . Indeed , alignment of the NTD of several lentiviruses in the SIVsm/SIVmac/HIV-2 lineage revealed multiple potential adaptations in SIVmac ( Figure 6 ) . Most striking is an inferred R97S change , identical to the change that appeared in the SIVsmE041 and SIVsmE543 experimental cohorts described above ( Figures 3 and 5B ) . Based on phylogeny of the SIVsm/SIVmac/HIV-2 lineage [3] , the R97S change was probably selected twice , once coinciding with emergence of SIVmac in rhesus macaques and once coinciding with emergence of SIVstm in stump-tailed macaques . Consistent with this interpretation , the underlying nucleotide substitutions are different in the two viruses ( AGA->AGC in SIVstm , and AGA->TCA in SIVmac ) . However , S97 is also found in a small percentage of HIV-2 and SIVsm isolates; thus , it is possible that one or both historical transmissions were initiated by an SIVsm with serine at position 97 , rather than de novo mutation as seen in the experimental cohorts . In either case , these combined observations strongly suggest that S97 is selectively advantageous in macaques . The second putative adaptation is a highly unusual LPA/QQ substitution at the tip of the CA 4–5 loop ( QQ89 , 90 in SIVmac239 ) . How this change was generated is unclear but must have involved multiple point mutations and a net loss of three nucleotides . While QQ89 , 90 is very common among SIVmac isolates , P90 ( the proline in LPA89–91 ) is extremely well conserved in SIVsm and HIV-2 isolates ( Figure 6 ) . The highly unusual nature of the LPA->QQ substitution , together with its location in a stretch of residues known to affect TRIM5-mediated restriction [14] , [39] , mark it as a potential adaptation that arose to circumvent restriction by rhesus macaque TRIM5 proteins . To ask whether the R97S and LPA/QQ changes in SIVmac strains arose as adaptations to overcome rhesus TRIM5 , we reverted these sites to the ancestral sequence in the context of the macaque-adapted strain SIVmac239 . We then tested SIVmac239S97R and SIVmac239QQ->LPA for gain-of-sensitivity to different rhesus TRIM5 alleles . Indeed , the S97R reversion resulted in increased sensitivity to all three TRIM5TFP alleles but had no effect on resistance to the TRIM5Q or TRIM5CypA alleles . In contrast , the QQ89 , 90-to-LPA89-91 reversion resulted in sensitivity to both TRIM5TFP and TRIM5CypA alleles . This closely resembles the pattern displayed by SIVsm isolates ( see Figure 2B and C ) , confirming that these represent bona fide adaptations to overcome restriction by rhesus TRIM5 alleles . Because QQ89 , 90 is in the 4–5 loop of capsid , the adaptation probably functions by altering the TRIM5CypA binding site . The structural basis for the interaction between the TRIM5 B30 . 2/SPRY domain and viral CA is not yet well defined , and it is not clear why QQ89 , 90 also affects resistance to TRIM5TFP alleles . However , this result is consistent with studies showing that site-directed mutations in the HIV-1 CypA-binding domain influence binding and restriction by TRIM5α proteins [39] . A better understanding of the influence of CA positions 89–91 on restriction awaits more detailed , structural understanding of the interaction between CA and TRIM5α . Neither reversion affected resistance of SIVmac239 to TRIM5Q , consistent with our observation that SIVsmE041 and SIVsmE543 , which retain ancestral states at these sites , were also resistant to TRIM5Q ( Figure 2 ) . We conclude that TRIM5Q alleles did not significantly hamper SIVsm colonization of rhesus macaques or its emergence as SIVmac in the 1970s . In fact , TRIM5Q/Q animals may have facilitated initial transmission of SIVsm among macaques , permitting higher levels of replication and increasing the probability of adaptation and spread . Unlike the R97S change , we did not see the LPA89–91/QQ89 , 90 change appear in either experimental cohort ( Figures 3 and 5B ) , possibly because the multiple mutations involved make it a low probability occurrence . Once virus with the QQ89 , 90 motif in CA appeared , however , it probably contributed to emergence of SIVmac by facilitating spread among animals bearing restrictive TRIM5CypA alleles . Consistent with this hypothesis , experimental introduction of the QQ89 , 90 sequence at the homologous positions in the sooty mangabey strain SIVsmE041 by site-directed mutagenesis rendered the virus resistant to TRIM5CypA ( Figure 6 ) . The mutation did not affect sensitivity to TRIM5TFP alleles or resistance to TRIM5Q alleles . Thus , reversion to the ancestral state in SIVmac239 ( changing QQ89 , 90 to LPA89–91 ) resulted in sensitivity to restriction by TRIM5CypA , and the reciprocal substitution to recreate the evolutionarily derived state in SIVsmE041 ( IPA to QQ ) resulted in resistance to TRIM5CypA . These results are consistent with the hypothesis that the QQ89 , 90 sequence is an adaptation to overcome rhesus TRIM5CypA that arose during emergence of the SIVmac lineage . In the case of SIVmac239QQ/LPA , the reversion also affected resistance to TRIM5TFP alleles , suggesting that the effects of the adaptation may be influenced by additional differences between the SIVsm and SIVmac capsids . Finally , note that two mutants , SIVsmE041IPA->QQ and SIVmac239S97R , have similar patterns of restriction ( compare Figure 6B and 6D ) , consistent with the fact that these two viruses are identical at the two sites in question ( i . e . , both viruses are QQ89 , 90 and R97 ) .
Host-encoded , dominant-acting blocks to retroviral infection , or restriction factors ( RF ) , were first defined genetically for murine and avian retroviruses more than 40 years ago [40] . Over the last decade , experiments seeking the causes of defined cellular blocks to HIV-1 infection uncovered three new RF genes , APOBEC3G ( now considered to be the prototype of a cluster of ∼8 APOBEC3 genes ) , TRIM5 , and Tetherin [11] , [41] , [42] . Consistent with the notion that these genes played a role in protecting host organisms from retroviral infections during the course of mammalian evolution , Tetherin , APOBEC3G , and TRIM5 display signs of long-term positive selection , including high levels of amino-acid diversification between closely related species [9] , [43]–[46] . It is generally assumed that RF genes such as TRIM5 , APOBEC3G , and Tetherin influence the distribution and spread of viruses between hosts , with viral epidemics resulting in selective sweeps of these loci , and successful viruses in turn adapting to the spectrum of restriction encoded by each host species [47] . Our results provide direct evidence that expression of one of these factors ( TRIM5 ) can indeed suppress viral replication during the early stages of cross-species transmission and that under such conditions , viral emergence and pathogenesis in the new host species requires adaptation to overcome restriction . Given overall similarities in evolutionary patterns and biological function , it seems very likely that the APOBEC3 enzymes and Tetherin/BST2 will also be found to influence patterns of cross-species transmission and viral emergence among extant species . In this study , we specifically demonstrated that the RF gene TRIM5 can suppress replication of a primate immunodeficiency virus in vivo at the time of cross-species transmission , and that TRIM5-mediated suppression of viral replication selects for acquisition of adaptive changes in the viral capsid protein . The data included retrospective genotyping of two independent , cross-species transmission cohorts , identification of at least two adaptations selected for resistance to TRIM5 during emergence of the SIVmac lineage , and functional assessment of allele-specific restriction in cell-culture . While the adaptations in SIVmac239 capsid identified here ( R97S and LPA89–91/QQ89 , 90 ) have a direct effect on restriction sensitivity , the modification of the otherwise highly conserved residues R97 and P90 may well have required additional adaptive adjustments elsewhere in the CA or the Gag polyprotein . Identifying such changes , if they exist , will require systematic testing of combinations of residues that vary between the SIVmac and SIVsm lineages for impact on relative infectivity in the presence and absence of TRIM5 expression . The transmission event ( s ) leading to spread and emergence of SIVsm in the U . S . macaque colonies have not been identified , although there is indirect evidence suggesting that transmission may have resulted from experimental procedures involving the transfusion of material from one species into another [27] . Regardless of the mechanism ( s ) of initial transmission , our data strongly suggest that adaptation to overcome restriction by a specific subset of rhesus macaque TRIM5 alleles played a key role in spread of SIVsm and its emergence as SIVmac in captive colonies of rhesus macaques . Based on cell-culture assays , analysis of experimental cohorts , and functional identification of adaptive changes in the SIVmac239 capsid , we propose a simple but likely scenario for the influence of variation in the rhesus macaque TRIM5 gene on emergence of SIVmac in the U . S . National Primate Research Center macaque colonies ( Figure 7 ) . Briefly , at the time of initial cross-species exposure , we believe that the SIVsm source was likely to be sensitive to restriction by a subset of rhesus macaque TRIM5 alleles , including those with a TFP sequence at residues 339–341 in the B30 . 2/SPRY domain ( TRIM5TFP ) and those that produce a TRIM5CypA chimera by alternative splicing ( TRIM5CypA ) . In contrast , we have yet to identify an SIV strain that is sensitive to alleles with a Q at position 339 ( TRIM5Q ) ; thus , it seems likely that animals bearing TRIM5Q alleles ( particularly TRIM5Q/Q homozygotes ) permitted initial spread and adaptation of SIVsm to rhesus macaques , potentially facilitating acquisition of adaptations to overcome other , as yet unidentified genetic barriers encountered in the newly invaded host species . The appearance of the R-to-S and the LPA-to-QQ changes in CA would have opened the way for the virus to spread to a larger percentage of the macaque population , ultimately leading to emergence of pathogenic SIVmac . None of the viruses tested in this study except HIV-1 were sensitive to the rhesus macaque TRIM5Q alleles . It is noteworthy that all known human alleles of TRIM5 have a Q at the homologous position , suggesting that human TRIM5 may not pose a critical barrier to transmission of SIVsm into human populations; in this regard , it would be interesting to assess the ability of human TRIM5 variants to restrict divergent primary isolates of SIVsm found in regions of endemic infection among African nonhuman primates and to look for correlations between TRIM5 and susceptibility to HIV-2 in human AIDS cohorts . An unexpected finding from analysis of the experimental transmission cohorts was the establishment of infection even in animals with restrictive genotypes . In many cases , we found that once animals with restrictive genotypes were infected , viral replication often continued to be suppressed to low levels for many weeks ( Figure 4 ) . In nature , lower viremia could reduce the probability that the infected individual will pass the virus on to additional individuals in the new host population , thus extending the role of TRIM5 in preventing emergence beyond the initial cross-species transmission event ( s ) . However , the retrospective analysis of SIVmac emergence does not necessarily address the impact of TRIM5 during natural exposure and transmission , and it is possible that the observed pattern is , at least in part , the consequence of experimental routes of transmission . For example , in many SIV/macaque studies , including the cohorts analyzed here , high titer stocks of virus ( TCID50 in the range of 1 to 1 , 000 ) are used to initiate infections by intravenous injection . Since the effect of restriction by TRIM5α is known to be saturable , it is possible that a large bolus of localized infection manages to initially swamp out restriction and permits virus to seed a large number of virus producing-cells and initiate a transient acute infection . In this regard , experimental intravenous infection of macaques may more closely resemble accidental human exposure to HIV-1 via blood-transfusion or contaminated , hypodermic syringes . It is possible that the effect of TRIM5 could be greater under conditions of natural exposure to virus , such as through sexual contact or fighting; however , we are currently unaware of any studies that address this issue . In recent years , many SIV/AIDS vaccine studies have incorporated low-dose mucosal challenges , to more closely mimic the conditions of sexual transmission . Under such conditions , it is possible that the effect of TRIM5 genotype on infection will be even more pronounced . Just as historical samples were used for this study , genotyping of appropriate archived samples from completed or ongoing SIV/vaccine studies may also prove useful for evaluating the interactions between TRIM5 genotype , viral dose , and route of transmission . Alternatively , there are published reports indicating that TRIM5 gene expression is interferon-regulated [48] , [49]; thus , it is also possible that the full impact of TRIM5 on viral replication in vivo is not manifest until after the first round of infection is well underway . Such a pattern is consistent with the observation that the impact of TRIM5 genotype in the SIVsmE543-3 infected animals appeared to persist during the weeks following acute infection ( Figure 4 ) . The TRIM5 gene is also known to encode multiple splice-isoforms [50] , including mRNA molecules lacking viral specificity encoded by a B30 . 2/SPRY or CypA domain; however , it is not known whether these are differentially expressed in vivo in response to infection or whether patterns of isoform expression differ between tissues . In vivo studies will be required to determine how TRIM5 is regulated , whether regulation changes in response to viral infection , and to identify the patterns of TRIM5 expression in cell types responsible for initial infection and spread in vivo . The impact of variation in rhesus macaque TRIM5 has practical implications for preclinical AIDS vaccine research . SIVsmE543 and the related SIVsmE660 serve as heterologous challenge strains for widely used SIVmac239-derived vaccine immunogens in rhesus macaques [38] , [51]–[54] . The pronounced effect of TRIM5 variation on SIVsmE543 infection is likely to confound comparison of vaccinated and control groups , particularly if they are not balanced for restrictive and permissive alleles; this may be especially true for studies with typically small numbers of animals ( n = 4–6 ) in each group . This may also be true of SIV strains not examined here , and possibly extends to other SIV/AIDS model organisms including Chinese-origin rhesus macaques and other commonly studied Macaca species , such as M . nemestrina and M . fascicularis . Moreover , vaccines expressing Gag , the target of TRIM5 , may be affected by TRIM5 genotype . This is potentially the case for live-attenuated SIV , where vaccine efficacy is influenced by the degree to which the vaccine strain replicates in the inoculated animal [55] . Finally , as discussed above , the potential effects of viral dose or route of transmission on restriction by TRIM5 remain to be determined . TRIM5 variation and susceptibility to infection has been explored in candidate gene studies in HIV/AIDS cohorts , although reported associations were weak [56]–[59] . Recently , a modest but significant association between TRIM5 genotype and infection in a cohort of SIVmac251-infected macaques was described [24] . While the SIV study did not correlate individual TRIM5 polymorphisms with viral replication levels , we note that all haplotypes associated with lower viremia in that report were of the TRIM5TFP class ( TRIM5CypA was excluded from their analysis ) [24] . The magnitude of the effect on SIVmac251 ( 1 . 3 log ) was smaller than we observed with SIVsmE543 or SIVsmE041 ( ∼2 . 0 to 3 . 0 log ) . This is consistent with the fact that SIVmac251 , like its derivative SIVmac239 , is the product of years of adaptation in rhesus macaques , while SIVsmE543 and SIVsmE041 have had little or no opportunity , respectively , to adapt to this species [25] , [29] , [31] . Similarly , lack of a strong association between TRIM5 variation and infection in HIV/AIDS cohorts may be due to limited variation in human TRIM5 and the fact that HIV-1 has been spreading in human populations for decades [60] , [61] . Our results , particularly when taken in light of the results from HIV/AIDS cohorts and the SIVmac251 cohort , support the conclusion that TRIM5 primarily governs the transmission of viruses between genetically distinct populations or species and further suggest that the impact of restriction can diminish as a virus spreads and adapts to a new host population .
All analyses were performed using archived material , and none of the analyses described involved new animal experiments . Animals at the New England Primate Research Center ( NEPRC ) were maintained in accordance with standards of the Association for Assessment and Accreditation of Laboratory Animal Care and the Harvard Medical School Animal Care and Use Committee . Animal experiments were approved by the Harvard Medical Area Standing Committee on Animals and conducted according to the principles described in the Guide for the Care and Use of Laboratory Animals . All animals at the NIH were housed in accordance with American Association for Accreditation of Laboratory Animal Care standards , and the NIH investigators adhered to the Guide for the Care and Use of Laboratory Animals prepared by the Committee on Care and Use of Laboratory Animals of the Institute of Laboratory Resources , National Resource Council , and the NIAID Animal Care and Use Committee-approved protocols . The SIVmac239-based retroviral vector V1EGFP ( gift from Hung Fan , University of California , Irvine , CA ) was modified to contain a functional gag-pol ORF . The NarI and SphI sites were used to replace SIVmac239 sequence with that of SIVsmE543 ( pGEM-E543 was a gift from Vanessa Hirsch , NIAID , Bethesda , MD ) . For this purpose the SphI site was introduced by PCR ( forward: 5′-CCTAGCAGGTTGGCGCCTGAACAGG-3′ , reverse: 5′-GTTATAGCATGCCTCTAGAGGGCGG-3′ ) . A 5′ half of SIVsmE041 was synthesized by GENEART ( Regensburg , Germany ) with the sequence based on a consensus of sequences obtained by A . K . and S . O . The fragment was engineered to contain NarI and SphI sites for subcloning into V1EGFP . SIVstm/37 . 16 ( gift from Frank Novembre , Emory University , Atlanta , GA ) was used for subcloning the SIVstm gag into V1EGFP using the NarI and SbfI sites . Primary Blood Mononucleated Cells ( PBMC ) were isolated from heparin- or EDTA-treated rhesus macaque blood samples by density centrifugation with Lymphocyte Separation Medium . PBMC were activated with RPMI/10% FBS containing 50 IU/ml phytohaemagglutinin or 5 mg/ml Concanavalin A for 2 d; afterwards cells were maintained in R10 supplemented with 50 IU/ml interleukin-2 . Crandell-Rees Feline Kidney ( CRFK ) cells as well as Human Embryonic Kidney 293T/17 ( HEK293T/17 ) cells were obtained from American Type Culture Collection ( Manassas , VA ) and grown in DMEM/10% FBS . Generation of stable CRFK cell lines was done as previously described in Newman et al . , PNAS 103 ( 50 ) , 2006 [15] . For the experiments performed here , new stable CRFK cell lines were made to express N-terminally HA-tagged rhesus TRIM5 alleles . For this purpose an HA tag was added by PCR ( forward: 5′-GCGGCCGCATGGCTTCTGGAATC-3′; reverse: 5′-CCACCGGTGGCTCAAGCGTAGTCTGGGACGTCGTATGGGTAGCCGCCAGAGCTTGGTGAGCACAGAG-3′ ) and the NotI and AgeI sites were used for subcloning into pQCXIN ( BD Biosciences , Franklin Lakes , NJ ) . Stable cell lines were maintained in DMEM/10% FBS supplemented with 0 . 5 mg/ml G418 . All cultured cells were maintained at 37°C with 5% CO2 . All single-cycle SIV viruses were produced in HEK293T/17 cells by cotransfection of the appropriate V1EGFP-SIV plasmid and pVSV-G ( Clontech Laboratories , Mountain View , CA ) , using the GenJet transfection system ( SignaGen; Ijamsville , MD ) . The single-cycle HIV-1 virus stock was made by cotransfection of pNL43DenvGFP and pVSV-G . Single-cycle HIV-2 was made by cotransfection of pHIV-2 ( ROD ) GFP and pVSV-G . Culture supernatants containing the single-cycle , GFP/EGFP expressing , VSV-G-pseudotyped virions were titered on untransfected CRFK cells; supernatant volumes resulting in approximately 30% GFP/EGFP+ CRFK cells were used for infectivity assays on the stably transfected cell lines expressing the various rhesus TRIM5 alleles . Stable CRFK cells were seeded at a concentration of 5×104 cells per well in 12-well-plates and infected with the appropriate amount of VSV-G pseudotyped , single-cycle , GFP/EGFP expressing viruses . All infections were done in triplicate . After 3 d , expression of GFP/EGFP was analyzed by fluorescence-activated cell sorting ( FACS ) performed on a FACSCalibur™ flow cytometer ( BD , Franklin Lakes , NJ ) , and data were analyzed using FlowJo ( Tree Star , Inc . , Ashland , OR ) . The Q89Q90 to LPA and S97R mutants of SIVmac239 capsid were made by site-directed mutagenesis on a SIVmac239 5′ plasmid ( Q89Q90 to LPA: forward: 5′-GCAGATTGGGACTTGCAGCACCCAATACCAGGCCCCTTACCAGCGGGACAACTTAGGGAGCCGTCAGG-3′; reverse: 5′CCTGACGGCTCCCTAAGTTCTCCCGCTGGTAAGGGGCCTGGTATTGGGTGCTGCAAGTCCCAATCTGC-3′ ) ( S97R: forward: 5′-CCCACAACCAGCTCCACAACAAGGACAACTTAGGGAGCCGAGGGGATCAGATATTGCAGGAAC-3′; reverse: 5′-GTTCCTGCAATATCTGATCCCCTCGGCTCCCTAAGTTCTCCTTCTTCTGGAGCTGGTTGTGGG-3′ ) . The IPA-to-Q89Q90 mutants of SIVsmE041 were made by site-directed mutagenesis on a SIVsmE041 5′ plasmid ( IPA-to-QQ: forward: 5′-CCACAGCCAGGTCCACAACAAGGACAACTTAGAGACCCGAGAGG-3′; reverse: 5′-CCTCTCGGGTCTCTAAGTTGTCCTTGTTGTGGACCTGGCTGTGG-3′ ) . For all mutants , NarI and SphI sites were used for subcloning into V1EGFP . TRIM5 genotypes of rhesus macaques were determined by isolation of genomic DNA from PBMC using the QIAamp DNA Blood Mini kit ( QIAGEN , Valencia , CA ) and direct sequencing of a PCR fragment ( forward: 5′-CAGTGCTGACTCCTTTGCTTG-3′; reverse: 5′-GCTTCCCTGATGTGATAC-3′ ) of the C-terminal B30 . 2/SPRY domain of TRIM5 . The following PCR protocol was implemented: 1 min at 94°C initial denaturation , 15 s at 94°C denaturation , 30 s at 55°C annealing , 1 min at 68°C extension , 10 min at 68°C final extension; steps 2 through 4 were repeated for 30 cycles . PCR fragments were sequenced by Retrogen ( San Diego , CA ) and data were analyzed with the Codoncode software ( Codoncode Corporation , Dedham , MA ) . Viral RNA was extracted from blood plasma with the High Pure Viral RNA kit ( Roche Diagnostics Corporation , Indianapolis , IN ) and partial capsid sequence was determined by direct sequencing of an RT-PCR fragment ( forward: 5′-GAAGCTTGCCACCATGGGCGCGAGAAACTC-3′; reverse: 5′-CCTCTCTGTTGGACTGCTGC-3′ ) . Stable CRFK cells expressing the various HA-tagged TRIM5 alleles were seeded at a density of 5×104 cells per well in a 6-well-plate . After 48 h , cells were lysed in M-PER reagent ( Pierce Biotechnology , Rockford , IL ) , and total protein concentration of each lysate was determined by measurement of A280 with a NanoDrop spectrophotometer ( Thermo Fisher Scientific , Waltham , MA ) . Equal amounts of total protein were separated by SDS/PAGE and HA-tagged TRIM5 proteins were detected with rat monoclonal Anti-HA-Peroxidase High Affinity antibody ( Roche Diagnostics Corporation , Indianapolis , IN ) . b-actin was detected with mouse monoclonal beta Actin-HRP antibody ( Abcam Inc . , Cambridge , MA ) . Forty-four rhesus macaques of Indian origin were infected with SIVsmE543-3 either intravenously ( n = 35 ) or intrarectally ( n = 9 ) with a TCID 50% ranging from 1 to 1 , 000 . There was no statistically significant difference in viral RNA levels relative to route of inoculation ( unpublished data ) . None of the animals were treated or vaccinated prior to infection . The majority of the animals ( n = 41 ) were inoculated with a cell-free virus stock generated by infection of pigtailed macaque PBMC with virus produced by transfection of CEMx174 cells with the SIVsmE543-3 molecular clone . The quantification of plasma viral loads was described previously in Hirsch et al . [29] . Briefly , viral RNA loads were determined by real-time reverse transcriptase-PCR ( RT-PCR ) . For this purpose , viral RNA from plasma was serially diluted and used as a template in a RT-PCR reaction , together with known amounts of pSG83 as an internal control template . Results were normalized to the volume of plasma extracted and expressed as SIV RNA copies per milliliter of plasma .
|
The human immunodeficiency viruses HIV-1 and HIV-2 originated from cross-species transmission of simian immunodeficiency viruses ( SIVs ) from chimpanzees ( SIVcpz ) and sooty mangabeys ( SIVsm ) , respectively . A related virus , SIVmac , causes AIDS-like pathogenesis in rhesus macaques; like HIV-2 , SIVmac is the product of a cross-species jump of SIVsm from sooty mangabeys . The primate TRIM5 gene encodes a factor with potent antiviral activity when tested in the laboratory , and TRIM5 proteins are thought to play a role in restricting the movement of viruses between species in nature . In this study , we show that genetic variation in the TRIM5 gene of rhesus macaques heavily influences the outcome of cross-species transmission of SIVsm and that emergence of SIVmac in rhesus macaques in the 1970s required adaptations to circumvent the genetic barrier imposed by the rhesus macaque TRIM5 gene . Our results confirm the hypothesis that TRIM5 can influence the process of cross-species transmission and emergence of viruses related to HIV-1 and HIV-2 and serve as a striking illustration of how host genes can influence virus evolution .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"virology/virus",
"evolution",
"and",
"symbiosis",
"virology/immunodeficiency",
"viruses",
"virology/animal",
"models",
"of",
"infection",
"virology",
"virology/mechanisms",
"of",
"resistance",
"and",
"susceptibility,",
"including",
"host",
"genetics"
] |
2010
|
TRIM5 Suppresses Cross-Species Transmission of a Primate Immunodeficiency Virus and Selects for Emergence of Resistant Variants in the New Species
|
Pollen tubes extend through pistil tissues and are guided to ovules where they release sperm for fertilization . Although pollen tubes can germinate and elongate in a synthetic medium , their trajectory is random and their growth rates are slower compared to growth in pistil tissues . Furthermore , interaction with the pistil renders pollen tubes competent to respond to guidance cues secreted by specialized cells within the ovule . The molecular basis for this potentiation of the pollen tube by the pistil remains uncharacterized . Using microarray analysis in Arabidopsis , we show that pollen tubes that have grown through stigma and style tissues of a pistil have a distinct gene expression profile and express a substantially larger fraction of the Arabidopsis genome than pollen grains or pollen tubes grown in vitro . Genes involved in signal transduction , transcription , and pollen tube growth are overrepresented in the subset of the Arabidopsis genome that is enriched in pistil-interacted pollen tubes , suggesting the possibility of a regulatory network that orchestrates gene expression as pollen tubes migrate through the pistil . Reverse genetic analysis of genes induced during pollen tube growth identified seven that had not previously been implicated in pollen tube growth . Two genes are required for pollen tube navigation through the pistil , and five genes are required for optimal pollen tube elongation in vitro . Our studies form the foundation for functional genomic analysis of the interactions between the pollen tube and the pistil , which is an excellent system for elucidation of novel modes of cell–cell interaction .
Cell-cell interactions can regulate the fate , morphology , and migration patterns of cells during development of multicellular organisms . Cell surface molecules mediate these interactions by initiating intracellular signal transduction cascades that cause changes in nuclear gene expression patterns [1] . Since the pollen tube of flowering plants interacts with several distinct cell types during its migration to an ovule , it represents an attractive model system for studying changes in global gene expression patterns in response to cell-cell interactions . Flowering plants alternate between haploid gametophytic and diploid sporophytic phases of their life cycle . Male and female gametophytes develop through a series of mitotic divisions of haploid spores , which are produced when diploid sporophytic cells within the anther ( male ) and ovule ( female ) undergo meiosis [2] . Male spores divide asymmetrically to produce a vegetative cell that engulfs a smaller generative cell . The generative cell then divides to form two sperm cells within the cytoplasm of the pollen grain , which constitutes the mature male gametophyte [3] , [4] . Upon binding a compatible stigma , the pollen grain germinates a tube that penetrates the stigma and grows rapidly through a protein and carbohydrate-rich extracellular matrix secreted by specialized cells of the pistil [5] . Pollen tubes extend by an actin-myosin-based tip-growth mechanism that transports vesicles loaded with new cell wall material to the extending apex [6]–[9] . In response to guidance cues from female cells , individual pollen tubes target and enter an ovule micropyle [10] , contact the female gametophyte [11] , arrest growth [12] , [13] , and burst [14] , releasing two sperm for fertilization of female gametes [15] . Pollen is released from anthers at anthesis and has therefore been amenable to global gene expression profiling . Transcriptome analysis showed that pollen expresses a unique subset of the Arabidopsis genome relative to sporophytic tissues [16]–[19] and revealed changes in the patterns of gene expression as the male gametophyte develops from a spore to a tricellular pollen grain [18] . Determination of the transcriptome of purified sperm cells showed that male gametes have a distinct gene expression program that contributes to the transcriptome of the pollen grain [20] . Recently , genome-wide expression profiling of pollen tubes grown in vitro identified a set of genes that are expressed in the pollen tube but not in pollen [21] . This important study suggests that there is de novo mRNA synthesis in the growing pollen tube and raises the interesting possibility that a novel set of genes may be expressed in response to growth through the pistil . Studies in maize and petunia suggest that pistils induce gene expression changes in pollen tubes . For example , exposure of petunia pollen to kaempferol , a pollen germination-inducing molecule produced by the stigma [22] , resulted in significant gene expression changes during the first 0 . 5 hours after pollen germination . Eight novel cDNAs whose expression increased in response to kaempferol were identified in petunia pollen tubes [23] . It is also clear that pollen tube physiology changes as a consequence of growth through pistil tissue , but the molecular bases for these changes are largely unknown . Pollen tubes extend at faster rates in a pistil and achieve substantially greater terminal lengths compared to pollen tubes grown in vitro [24] . Furthermore , pollen tubes germinated in vitro target the ovule micropyle at very low efficiencies; however , if pollen tubes are first grown through pistil tissues , guidance to ovules is significantly enhanced [25] , [26] . Therefore , it is likely that the transcriptome of pollen tubes grown through the pistil differs considerably from that of in vitro-grown pollen tubes . Defining these differences could lead to the discovery of genes that are activated by potentiation of the pollen tube by the pistil and are required for pollen tube guidance , and to the identification of gene regulatory networks that mediate the pollen tube response to the pistil environment . In this study , we defined the transcriptome of pollen tubes that have grown through pistil tissues using a semi-in vivo pollen tube ( SIV PT ) growth system we developed for Arabidopsis [26] . Importantly , the SIV PT transcriptome was significantly different from those of pollen grains or pollen tubes grown in vitro . In addition , we showed that a significant number of genes are shared between the SIV PT transcriptome and sporophytic tissues , which are not expressed in pollen or pollen tubes grown in vitro . We also defined a set of genes that are enriched in semi-in vivo grown pollen tubes relative to pollen , pollen tubes grown in vitro , and a collection of sporophytic tissues . The distribution of functional categories in this set of genes compared to pollen grains revealed a significant enrichment for the Toll/Interleukin-1 Receptor homology-Nucleotide Binding Site-Leucine Rich Repeat ( TIR-NBS-LRR ) -type receptor family of proteins [27] . These genes have been implicated in pathogen-derived-effector-protein recognition and could play a direct signaling role during pollen tube potentiation by pistils . To determine whether genes whose expression increases during pollen tube growth in vitro and/or semi in vivo are required for pollen tube function , we performed reverse-genetic analysis of selected genes . We identified five mutants that disrupt pollen tube growth in vitro and two mutants that specifically disrupt pollen tube growth in the pistil . Our studies lay the foundation for functional genomic analysis of pollen tube-pistil interactions .
To identify gene expression changes during pollen tube growth in vitro or through a pistil , we performed comparative microarray analysis with RNA isolated from dry , un-germinated pollen ( dry pollen , Figure 1A ) , pollen grown in vitro for 0 . 5 hours ( 0 . 5 h PT , Figure 1B ) , or for 4 hours ( 4 h PT , Figure 1C ) and pollen germinated and grown through the stigma and style ( SIV PT; Figure 1D-1F ) . Pollen tubes grown by the semi-in vivo method exit as a bundle from the cut end of a style and fan out on the solid pollen growth medium ( Figure 1E , 1F ) . Pollen tube bundles from ∼800 cut pistils were excised and combined for RNA isolation ( Figure 1E , 1F ) . Using RNA isolated from the four different pollen conditions ( dry pollen , 0 . 5 h PT , 4 h PT and SIV PT; also see Materials and Methods ) , we synthesized probes and hybridized to Affymetrix Arabidopsis ATH1 genome arrays . We generated probes from four biological replicates of dry pollen , 0 . 5 h PT , 4 h PT and three replicates of SIV PT . The raw expression data from these 15 experiments are provided ( Table S1 , Table S2 , Table S3 , Table S4 ) . To detect genes that are preferentially expressed in pollen tubes compared to other cell types , we obtained publicly available microarray data for seven sporophytic tissues: 7-day-old roots , 17-day-old roots , 8-day-old seedlings , 21-day-old seedlings , 17-day-old rosette leaves ( three replicates of each , [28] ) , unpollinated ovary and unpollinated stigma ( four and three replicates respectively , [29] ) . Data from a previously published pollen microarray was also included as a reference ( three replicates , [28] ) . By analyzing these 25 publicly available data sets along with our 15 arrays ( Robust Microarray Analysis tools , RMA , see Materials and Methods ) , we obtained normalized expression values for each gene that allowed us to make comparisons among these experiments ( Table S5 ) . The ranges of Pearson coefficients of array-array intensity were high for pairwise comparisons among the dry pollen , 0 . 5 h PT , 4 h PT and SIV PT replicates , suggesting that there is high reproducibility among the biological replicates and that one pollen type could be distinguished from the other ( Table S6 ) . Similar results were obtained with hierarchical clustering of pollen arrays ( Figure S1 ) . Pairwise comparisons of Pearson correlation coefficients showed that previously published pollen data [28] was most similar to our dry pollen ( 0 . 935–0 . 944 ) and 0 . 5 h PT ( 0 . 935–0 . 949 ) samples . Using our RMA-normalized data set ( Table S5 ) , we identified genes that are expressed during pollen tube growth . RMA analysis does not provide a ‘present’ or ‘absent’ score , so we set an expression value of 100 or higher as a stringent threshold for expression to obtain a conservative estimate of gene expression in each cell or tissue analyzed . In reverse transcription followed by real-time quantitative PCR ( qRT-PCR ) experiments , we could consistently confirm expression for genes above this threshold value ( see below ) . Using these criteria , we found that 6 , 304 , 6 , 308 , and 6 , 356 genes were present in dry pollen , 0 . 5 h PT , and 4 h PT , respectively ( Figure 2A , 2B ) . The number of genes expressed in SIV PT was greater ( 7 , 044 ) than any other pollen tube growth condition tested , suggesting a substantial change in the transcriptome following interaction with female reproductive tissues ( Figure 2B ) . We determined the extent of overlap among the transcriptomes of the four pollen conditions we tested . There is significant overlap between dry pollen and 0 . 5 h PT ( Figure 2A; Table S6 , Table S7 , and Figure S1 ) . Because of this extensive similarity , we combined dry pollen and 0 . 5 h PT into one group of 6 , 677 pollen genes [dry pollen and 0 . 5 h PT] ( sector S1 , Figure 2B ) . This combined set , when compared with 4 h PT ( sector S2 , Figure 2B ) and SIV PT ( sector S3 , Figure 2B ) identified 5 , 312 genes ( sector S4 , Figure 2B ) shared by all pollen samples , representing a core set of pollen genes . Previously characterized pollen and pollen tube-expressed genes known to be critical for pollen tube growth such as the ROP1 GTPase ( At3g51300 , [30] ) , AtGEF12 ( At1g79860 , [31] ) , RabA4d ( At3g12160 , [32] ) , ACA9 ( At3g21180 , [33] ) , CNGC18 ( At5g14870 , [34] ) , the VANGUARD pectinesterase ( At3g621790 , [35] ) , and AtMGD2 and AtMGD3 ( At5g20410 and At2g11810 , [36] ) were expressed in the SIV PT , 4 h PT , 0 . 5 h PT and dry pollen transcriptomes ( Table S5 ) . Gene Ontology ( GO ) term enrichment analysis of the three sectors ( S1–S3 ) also revealed that certain GO term categories are overrepresented in all pollen conditions relative to the whole genome ( first 15 GO categories , Figure 2C ) . The two pollen tube transcriptomes shared 273 genes not found in [dry pollen , 0 . 5 h PT] ( S1 , Figure 2B ) . GO term category overrepresentation analysis also highlighted the overlap between the 4 h PT and SIV PT transcriptomes: genes encoding kinases , antiporters , nucleoside triphosphatases , calcium ion binding proteins , and nucleic acid binding proteins are overrepresented in 4 h PT and SIV PT but not in dry pollen and 0 . 5 h PT ( Figure 2C ) . Interestingly , the number of genes detected only in SIV PT ( 1 , 254 ) was significantly higher than the number of genes detected only in 4 h PT ( 75 , Figure 2B ) . These data suggest that growth through the pistil elicits a significant change in the pollen tube transcriptome . We next explored the overlap in expression between sporophytic tissues ( expressed in any of seven sporophytic samples analyzed ) , SIV PT , and all other pollen conditions ( Figure 2D ) . Notably , SIV PT and sporophytic tissues share a set of 871 genes that are not expressed in the three other pollen samples analyzed . This analysis also identified 2 , 040 genes that were expressed in pollen but not sporophytic samples . Among these 2 , 040 genes , 1 , 097 are shared by all four pollen conditions ( Figure 2E ) . Our analysis also identified a set of 507 pollen tube-enriched genes , including the 100 genes that are common to SIV PT and 4 h PT ( Figure 2E ) . Interestingly , SIV PT has the largest number of unique genes ( 383 ) compared to any other pollen condition ( Figure 2E and Table S8; referred henceforth as SIV PT-enriched genes ) , further confirming that the SIV PT transcriptome is distinct from dry pollen or in vitro grown pollen tubes despite the overlap it shares with these transcriptomes . We determined if any GO terms were significantly overrepresented among the 383 SIV PT-enriched genes ( Figure 2D , Table S8 ) compared to pollen-expressed genes ( ATGE_73A-C; [28] ) . Twenty-one GO terms , including those related to signaling , cell extension and transcription , were significantly overrepresented in the SIV PT-enriched genes ( P value cut off <0 . 05 , Table 1 and Table S9 ) . The most overrepresented terms in the three GO categories were transmembrane receptor activity ( molecular function , n = 4 , P = 0 . 001 ) , defense response ( biological process , n = 7 , P = 0 . 003 ) , and intrinsic to membrane ( cellular component , n = 4 , P = 0 . 003 ) . There were four genes common to each of these three GO categories and all of them belong to the TIR-NBS-LRR receptor subfamily that is part of a ‘R’ gene superfamily implicated in pathogen recognition [27] . In addition , a set of protein kinases ( molecular function , n = 4 , P = 0 . 024 ) was enriched in SIV PT ( Table 1 , Table S9 ) . These signaling genes may facilitate pollen tube perception and response to pistil guidance cues . A set of genes annotated as polygalacturonases , sucrose transporters , and antiporters are overrepresented in SIV PT compared to pollen ( Table 1 , Table S9 ) . These categories have been implicated in pollen tube extension [37]–[42] . Several GO terms related to transcription were also overrepresented in SIV PT-enriched genes ( Table 1 ) ; they may respond to growth through the pistil and function as key regulators of expression of other genes required for pollen tube growth and guidance ( Table 2 ) . We used a t-test ( Materials and Methods ) on the RMA-normalized data ( Table S5 ) to define statistically significant changes in gene expression during pollen tube growth ( Table 2 ) . To minimize false positives , we established two stringent cut-off values: only those genes that had a B value ( false discovery rate ) of 3 or higher and a fold change of at least 3 were considered to have undergone a significant change in expression ( Table S5 ) . We analyzed changes that occurred between dry pollen and 0 . 5 h PT to assess the impact of pollen hydration on gene expression . This analysis defined a very small number of genes that increase ( 15 ) or decrease ( 8 ) during the hydration process ( Table 2 , Table S10 ) . To define the changes in transcript levels that occur after hydration and during pollen tube growth in vitro , we compared 0 . 5 h PT and 4 h PT; this time period accounts for nearly all of pollen tube extension observed in vitro ( Figure 1 ) . One hundred thirty-seven genes had significant increases in expression value in this comparison , while no genes were observed to have a significant decrease ( Table 2 , Table S11 ) . We also compared dry pollen with 4 h PT and identified 186 genes that increase and 11 genes that decrease during the entire process of hydration and growth in vitro ( Table 2 , Table S12 ) . These results are also consistent with a recent report that identified modest , but significant , changes in the transcriptome of in vitro-grown pollen tubes [21] . The number of genes whose expression was significantly different between SIV PT and dry pollen ( 1 , 578 ) or between SIV PT and 4 h PT ( 1 , 135 ) was dramatically greater than any other comparison among the pollen transcriptomes ( Table 2 , Table S13 , Table S14 ) . We compared SIV PT with 4 h PT and identified a large number of genes ( 900 ) with significantly higher expression values in SIV PT compared with 4 h PT . There were also a significant number of genes whose expression went down ( 235 ) in this comparison ( Table 2 , Table S13 , Table S14 ) . The large number of genes ( 1 , 135 ) that are altered when pollen tubes grow through pistil tissues are candidate factors that underlie the physiological and molecular changes in pollen tubes during a successful fertilization event [22]–[26] . Among the altered genes , we identified a set of genes that can best distinguish SIV PT from dry pollen and 4 h PT ( Table S15 ) using the non-hierarchical k-means clustering method ( [43]; also see Materials and Methods ) . These genes could be used as markers for pollen tubes that have interacted with the pistil . In SIV PT , the up-regulated genes ( compared to both dry pollen and 4 h PT ) included the overrepresented molecular function GO categories of transporter , antiporter , symporter activity and calcium ion binding . These functions are known to be critical for pollen tube growth [39]–[42] . Interestingly , a different set of transporter genes is down-regulated in SIV PT compared to 4 h PT; similarly , a separate set of antiporter genes is also down regulated in SIV PT compared to pollen ( Figure 2F ) . These results suggest that transporter and antiporter gene expression is highly dynamic during pollen tube growth in a pistil . There is a significant down regulation of a distinct set of pectinesterases in SIV PT compared to 4 h PT ( Figure 2F ) . Pectinesterases alter the mechanical strength and rigidity of the pollen tube wall during the process of pollen tube elongation [35] , [44] , [45]; our data suggest the possibility of functional specialization within this large gene family and that , as with transporters , expression of pectinesterases is dynamic in pollen tubes . We used qRT-PCR to verify pollen tube gene expression data obtained from microarray experiments ( Materials and Methods ) . We selected 15 pollen-enriched genes that had high levels of expression in pollen compared to sporophytic tissues and were expressed at significantly higher levels in 4 h PT compared to either dry pollen , 0 . 5 h PT , or sporophytic tissues ( Table S11 , Table S12 ) . qRT-PCR corroborated the microarray analysis , showing that all 15 of these genes were pollen-enriched and expressed at significantly higher levels in 4 h PT compared to either dry pollen or 0 . 5 h PT ( Figure S2A , Figure 3 , and Figure S3A , first 15 genes ) . Based on these results , we conclude that our microarray results accurately reflect gene expression patterns during in vitro pollen tube growth . We also tested a set of genes shown by microarray to be expressed at varying levels in pollen , pollen tubes and sporophytic tissues ( n = 6 ) ( Table S5 ) . qRT-PCR confirmed the expression of these genes in these cells and tissues ( Figure S2A , Figure 3 , and Figure S3A , bottom six genes ) . The relative expression of At3g60080 , At2g01290 , and At1g22410 were identical in qRT-PCR and microarray experiments ( Figure 3 , Figure S3A ) . Expression of the remaining three genes ( At1g69840 , At1g09070 , At3g23820 ) was confirmed in pollen and sporophytic tissues by qRT-PCR; however , there were discrepancies in the relative expression of these genes when qRT-PCR data were compared with microarray experiments ( Figure 3 , Figure S3A ) . These differences could be attributed to the variability in plant growth and RNA preparation from 8-day- and 21-day- old seedlings between different laboratories ( our data and that of [28] ) . We next determined whether the gene expression differences between SIV PT and 4 h PT in microarray experiments ( Table S14 ) could be detected by qRT-PCR by testing 10 genes with expression values that were significantly higher , and six genes that were significantly lower , in SIV PT compared to 4 h PT ( Figure S2B , Table 3 , Table S14 ) . All 10 genes expressed at higher levels in SIV PT compared to 4 h PT in microarray experiments were also higher in SIV PT by qRT-PCR experiments ( Table 3 ) . The reduction in gene expression detected by microarray for six genes was also confirmed by qRT-PCR ( Table 2 ) . Based on these results , we conclude that a high degree of confidence can be placed on the changes in gene expression identified by the microarray experiments reported in this study . To determine whether pollen tube growth in an intact pistil elicits similar changes in gene expression as those observed in microarray analysis of SIV PT , we used qRT-PCR , to monitor gene expression changes between dry pollen , unpollinated pistils , pistils pollinated for one minute , and pistils pollinated for two hours ( an in planta tissue type that most closely resembles SIV PT ) . We evaluated expression of a set of genes that were found to be induced in SIV PT compared to dry pollen by microarray experiments ( Table S13 ) . Pollen tubes comprise only a small fraction of pollinated pistil tissue; therefore , we undertook two strategies to allow detection of induction of gene expression in in vivo-grown pollen tubes . First , among the genes induced in SIV PT , we chose genes with relatively low expression in the stigma and ovary [29] ( Table S5 ) , as high expression in these tissues would preclude detection of induction in pollen tubes . Second , we isolated RNA from the stigma and style portion of the pollinated pistil because pollen tubes are concentrated here during the first two hours of growth ( Figure 4A ) . qRT-PCR shows that the mRNA abundance of all 10 genes was higher in pistils that had been pollinated for two hours compared to pistils that had been pollinated for one minute , unpollinated pistils , or dry pollen ( Figure 4B ) . These results suggest that increases in mRNA abundance detected in SIV PT also occur during pollen tube growth in a pistil . We used reverse genetic analysis of 33 genes that were significantly higher in SIV PT versus 4 h PT ( n = 10 , range = 5 to 22 fold change , Table S14 ) or 4 h PT versus dry pollen ( n = 23 , range = 3 to 41 fold change , Table S10 ) to determine whether candidate genes identified by microarray analysis were critical for pollen tube growth and guidance . T-DNA insertion mutants from the Syngenta Arabidopsis Insertion Lines ( SAIL , [46]; Table S16 ) were analyzed using a series of sensitive pollen function assays . A subset of the SAIL collection [46] was generated in the quartet ( qrt ) mutant background with a T-DNA carrying a β-glucuronidase ( GUS ) reporter gene expressed from the pollen-specific LAT52 promoter [46]–[48] and a Basta ( herbicide ) resistance gene . The qrt mutation causes the four products of male meiosis to be released as a tetrad of pollen grains but does not interfere with pollen tube growth [48] . These unique features of the SAIL collection offer significant advantages for analysis of pollen mutant phenotypes over other mutant collections . First , mutant pollen grains can be easily identified within tetrads produced by a heterozygous mutant plant ( 2 GUS+ mutant: 2 GUS- wild type ) . Second , GUS expression in mutant pollen tubes allows direct comparison between mutant and wild-type pollen tube growth in vitro or in a pistil . These attributes have been exploited previously in forward genetic analysis of pollen tube growth and guidance [5] , [49] . The pollen function assays we employed require cosegregation between Basta resistant ( Basta ) in seedlings , GUS expression in pollen , and a single locus T-DNA insertion in the gene of interest . We identified 50 SAIL lines with potential insertions in 33 genes chosen for analysis ( Table S16 ) . Using PCR , we verified T-DNA insertion sites in Basta progeny from 39 of the 50 lines ( Table S16 ) . Twenty-seven of these 39 insertion lines showed 2 GUS+ ( mutant ) : 2 GUS- ( wild type ) segregation in pollen tetrads and were heterozygous for the insertion by a PCR assay ( Table S16 and Materials and Methods ) , indicating cosegregation between the gene of interest and the T-DNA insertion . We discarded the other 12 lines because GUS expression in pollen tetrads was consistent with multiple T-DNA insertion sites ( Table S16 ) . To further confirm that the remaining 27 lines had a single-locus insertion of the T-DNA , we analyzed the segregation of Basta among the progeny of a self-fertilized heterozygous plant ( 2∶2 GUS+: GUS- tetrads , heterozygous by PCR assay ) . Plants heterozygous for a single insertion site are expected to generate 75% Basta progeny ( 3∶1 segregation of dominant marker ) . However , if the insertion disrupts a gene required for male/female gametophyte function , or seed development , the fraction of Basta progeny will be significantly reduced [50] . We found that the percentage of Basta progeny was ∼75% or significantly lower in all 27 lines , confirming that they had a single T-DNA insertion site and indicating that several ( 16/27 ) may disrupt the male and/or female gametophyte or seed development ( Table 4 ) . Mutations that completely disrupt pollen function are not transmitted to progeny through pollen , while milder defects reduce , but do not eliminate transmission [49] . To focus on transmission of the T-DNA through pollen , we pollinated male sterile 1 ( ms1 ) pistils with heterozygous pollen from 27 single-locus T-DNA insertion lines and determined the percentage of Basta plants in the progeny ( Table 4 ) . Any significant deviation from 50% in this assay indicates that mutant pollen is less likely to fertilize ovules than wild-type pollen . We found that progeny from two of the insertion lines , one in a 4 h PT-induced gene ( At1g60420 ) and another in a SIV PT-induced gene ( At3g18000 ) yielded significantly fewer than the expected 50% Basta plants ( Table 4 ) indicating that these genes are critical for pollen function in the pistil . At1g60420 encodes an uncharacterized protein with thioredoxin and C1-like domains . C1 domains have been shown to bind diacylglycerol and phorbol esters and are implicated in lipid signaling in mammals [51] . At3g18000 ( XIPOTL ) encodes one of three Arabidopsis S-adenosyl-L-methionine: phosphoethanolamine N-methyltransferase ( PEAMT ) required for synthesis of phosphatidylcholine , a major membrane lipid and the precursor of phosphatidic acid , an important lipid signaling molecule [52] . Since the proteins encoded by these two genes may be involved in generation of lipid signaling molecules , the in vivo transmission defects of insertions in these genes point to a potential role for lipid signaling in pollen tube growth through the pistil . To analyze the growth behavior of At1g60420-1 and At3g18000-1 pollen tubes in vivo and determine the specific stage of pollen tube growth disrupted by these insertions , we pollinated ms1 pistils with heterozygous pollen and stained for GUS activity 24 hours later [5] , [49] . When ms1 pistils were pollinated with heterozygous control pollen , GUS+ pollen tubes germinated , penetrated the stigmatic papillae , grew through the style , entered the ovary through the transmitting tract , and migrated toward an ovule . After entering the micropyle , GUS+ pollen tubes burst , releasing an aggregate of GUS activity in the micropylar end of the ovule serving as a convenient marker for successful ovule targeting by a pollen tube ( Figure 5A–5C ) . In this assay , ∼50% of ovules were targeted by GUS+ pollen tubes from the heterozygous control line ( Figure 5J; [5] ) . When ms1 pistils were pollinated with heterozygous At2g31550-1 or At5g22910-1 pollen ( insertions that did not affect mutant allele transmission through pollen , Table 4 ) the germination and growth of the GUS+ tubes in stigma , style and transmitting tract was normal ( data not shown ) and nearly 50% of the ovules were targeted by GUS+ pollen tubes ( Figure 5J ) . However , when ms1 pistils were pollinated with heterozygous At1g60420-1 or At3g18000-1 pollen , GUS+ pollen tubes were only half as efficient in targeting ovules as the GUS- tubes ( Figure 5D , 5G , 5J ) . These results are consistent with the reduction in mutant allele transmission in At1g60420-1 and At3g18000-1 insertion lines ( Table 4 ) . In addition to a significant reduction in the ability to target ovules , At1g60420-1 and At3g18000-1 GUS+ pollen tubes exhibited an increased frequency of abnormal pollen tube behaviors . Unlike in control crosses ( ms1 ovules with GUS+ pollen tubes from control heterozygotes , 0% , n = 182 ) , a noticeable fraction of ms1 ovules had At1g60420-1 or At3g18000-1 GUS+ pollen tubes that approached , but did not enter , the ovule micropyle ( Figure 5F , 5I; At1g60420-1 , 8 . 33% , n = 204; At3g18000-1 , 5 . 49% , n = 164 ) . For a small number of ovules , pollen tubes grew towards the chalazal end , instead of the micropylar end , of the ovule ( not shown; At1g60420-1 , 1 . 22% , n = 204; At3g18000-1 , 1 . 96% , n = 164; control , 0% , n = 182 ) . Finally , ovules that attracted multiple GUS+ At1g60420-1 or At3g18000-1 pollen tubes were observed ( Figure 5E , 5H; At1g60420-1 , 1 . 83% , n = 204; At3g18000-1 , 3 . 43% , n = 164; control , 0% , n = 182 ) . We wanted to determine whether At1g60420-1 or At3g18000-1 , insertions that disrupt pollen tube growth in the pistil , had inherent defects in pollen tube extension . We also wanted to examine if other single-locus insertion lines had subtle defects in the ability of pollen grains to form and extend a polar tube that may have been masked by growth in the pistil . In vitro pollen germination and tube growth provides a sensitive and direct assay for pollen function that is independent of pistil tissue . We assayed in vitro pollen tube germination and growth for the 27 single-locus insertion lines . We used heterozygous pollen so that we could analyze mutant pollen ( GUS+ , blue ) alongside wild-type pollen ( GUS- , white ) after staining for GUS activity . This side-by-side comparison between mutant and wild type is critical because it provides an internal control for the inter-experiment variability of pollen tube growth in vitro [53] . GUS staining in pollen tubes was dark enough to clearly distinguish mutant from wild type in 12 lines ( Table 5 , Figure S4 ) . We analyzed at least three replicates of all in vitro pollen germination and tube length experiments using statistical methods that account for variation between and within experiments and set a stringent criterion for statistical significance at P<0 . 001 . In vitro pollen germination rates and tube lengths were similar for GUS+ and GUS- pollen from heterozygous control plants ( Table 5 , Figure S4 ) . Although At1g60420-1 and At3g18000-1 were transmitted through the pollen with significantly reduced frequencies ( Table 4 ) and were less likely to target ovules ( Figure 5 ) , these insertions did not affect tube growth in vitro ( Table 5 , Figure S4 ) . These results indicate that the in vivo transmission defect in these insertion lines cannot be due to an inherent defect in pollen tube extension and is likely caused by loss of functions specifically required to navigate the pistil environment . The lengths of GUS+ pollen tubes were significantly shorter ( P<0 . 001 ) than GUS- pollen tubes for insertions in At2g31550 ( GDSL-motif lipase/hydrolase family protein ) , At5g23530 ( carboxyesterase 18 ) , At4g08670 ( similar to lipid transfer proteins ) , At5g67250 ( an SCF-type F-box and leucine rich repeat-containing E3 ubiquitin ligase ) , and At5g55020 ( MYB120 , Table 5 , Figure S4 ) . Of the 12 lines we analyzed , only the insertion in MYB120 ( At5g55020-1 ) also caused a significant defect ( P<0 . 001 ) in pollen tube germination ( Table 5 , Figure S4 ) . These results indicate that insertions in five genes resulted in pollen tube growth defects that were only detectable in vitro .
We characterized the global gene expression profiles of in vitro- and semi in vivo-grown Arabidopsis pollen tubes and present the first molecular and genetic analysis of a set of genes expressed by the pollen tube as it grows through pistil tissue . One approach to identify pollen tube genes that respond to the pistil would be to isolate intact , in vivo-grown pollen tubes . This is possible in species like lily that have a hollow style [54] , [55]; however , genomic resources are not currently available for these plants . In Arabidopsis , pollen tubes grow deep within a solid style , making it extremely difficult to obtain sufficient quantities of pure in vivo-grown pollen tubes for microarray analysis . We overcame this challenge by collecting a large number of pollen tubes that had grown through pistil tissue using the semi-in vivo method [26] . This procedure offered several advantages over alternative methods . First , harvested pollen tubes were directly used for RNA isolation without any further manipulations ( such as cell sorting or protoplast preparation ) . Second , it allowed the wild-type pollen tube transcriptome to be assessed directly and eliminated the need for using mutant or transgenic marker lines that could have inappropriately altered the dynamics of wild-type pollen tube gene expression . Third , our method enriches for the actively extending pollen tube tip , which includes the vegetative nucleus , two sperm cells , and majority of the pollen tube cytoplasm . Finally , because SIV PT is comprised solely of pollen tubes , we were able to detect even those genes that i ) are expressed at low levels in pollen tubes , ii ) exhibited pollen tube-specific expression , and iii ) undergo only modest changes in expression during pollen tube growth through the pistil . Characterization of the SIV PT transcriptome provides the first global view of pistil-dependent gene expression changes in pollen tubes . The SIV PT transcriptome is about 10% ( ∼700 genes ) larger than the pollen or in vitro-grown pollen tube transcriptome ( Figure 2 ) . The pollen tube gene expression profile undergoes a dramatic change upon interaction with the pistil; expression levels of nearly 1 , 500 and 1 , 100 genes are significantly altered in SIV PT compared to dry pollen and 4 h PT , respectively ( Table 2 ) . Finally , a distinct set of transcripts accumulate preferentially in SIV PT relative to pollen , 4 h PT and sporophytic tissues ( Table S13 ) ; defining these genes offers an opportunity to further investigate the molecular basis of the pollen tube response to the pistil environment . Additional analysis will be necessary to examine changes in the pollen tube transcriptome elicited by other pistil tissues such as the transmitting tract and ovules . All of the genes that were annotated as transmembrane receptors and overrepresented in the SIV PT-enriched gene list were TIR-NBS-LRR-type receptor proteins ( Table 1 ) , a subgroup of the resistance ( R ) gene family that mediate molecular recognition of pathogen-derived effector proteins [27] , [56] . The precise biological functions of many members of this large gene family , including the four in the SIV PT-enriched gene list , have not been determined . However , it is clear that some TIR-NBS-LRR-type genes have functions unrelated to plant defense . For example , an Arabidopsis TIR-NBS-LRR-type receptor mutant ( At5g17880 ) has a constitutive shade-avoidance response [57] . TIR-NBS-LRR receptors are highly variable and show signatures of rapid evolution [27] , features common in reproductive proteins that contribute to species-specific interactions between mating partners [58] , [59] . Intriguingly , another family of variable proteins , initially identified as defensins , were recently shown to function as pollen tube attractants in Torenia [10] . It will be interesting to explore the function of TIR-NBS-LRR receptors in pollen tube growth and guidance using the genetic approaches described here . Our microarray analysis identified a large number of mRNAs that increase in abundance as pollen tubes grow in vitro or through the pistil ( Figure 2B and 2C , Table 2 ) , adding support to the view that pollen tubes transcribe mRNA during pollen tube growth [21] . Several categories of genes involved in transcription were enriched in SIV PT compared to pollen ( Table 1 ) . A variety of transcription factors including MYB65 ( see below ) , other MYB-family proteins ( At5g38620 , At2g13960 , At2g20400 ) , MADS box-containing proteins ( At5g38620 , PHERES 2/AGL38 , AGL73 ) , and homeobox-containing proteins ( At3g19510 , At2g32370 ) were among the overrepresented genes , suggesting the existence of a network of gene regulatory mechanisms to mediate pollen tube growth in the pistil . Genetic analysis of pollen tube-expressed transcription factors ( see below ) offers the potential to identify key regulators of pollen tube gene expression . The SIV PT sample includes the two sperm cells; so , some SIV PT-expressed genes may be transcribed in the sperm nucleus . The transcriptome of sperm cells purified from pollen grains , comprising 5 , 829 Arabidopsis genes , was recently characterized [20] . We examined the overlap between SIV PT-enriched genes and genes called ‘present’ in sperm and found that 161 of the 383 SIV PT-enriched genes ( 43% ) are expressed in sperm ( Table 1 , Table S7 ) . Genes overrepresented in SIV PT , but not detected in sperm , include those potentially important for signaling ( transmembrane receptor activity , Table 1 ) , transcription ( histone acetyltransferase activity , Table 1 ) and pollen tube growth ( polygalacturonase , sucrose transport and antiporter activity , Table 1 ) . Genes proposed to be involved in DNA repair , chromosome segregation , and cell cycle regulation ( Table 1 ) were overrepresented in SIV PT-enriched genes; all of these genes are present in sperm ( Table 1 , Table S8 ) [20] . The pollen tube nuclear DNA ( vegetative nucleus ) does not replicate during pollen tube growth; however , sperm complete a round of DNA synthesis during pollen tube growth in the pistil [60] . This group of genes , identified in the SIV PT transcriptome , is therefore likely expressed in sperm as the pollen tube is growing through the pistil and function in sperm DNA synthesis . One of the goals of this study was to assess the extent to which microarray analysis identifies genes that are critical for pollen function . We identified single-insertion-locus T-DNA lines and employed four highly sensitive assays to determine loss-of-function phenotypes in pollen . Insertions in two genes ( At3g18000 , At1g60420 ) affected mutant allele transmission through pollen and disrupted pollen tube growth and guidance in vivo . Insertions in five additional genes ( At2g31550 , At4g08670 , At5g23530 , At5g55020 , At5g67250 ) caused pollen tube growth defects in vitro . A previous forward genetic screen yielded ∼30 mutants that disrupt pollen function from a population of ∼10 , 000 T-DNA insertion lines ( 0 . 3% , [49] ) . In this study , by starting with a population of 50 T-DNA insertions in genes induced during pollen tube growth , we identified seven mutations that disrupt pollen tube growth in vitro or in the pistil ( 14% ) ; this amounts to a ∼45 fold enrichment in identification of functionally significant genes over the forward genetic screen . We ascribed loss-of-function mutant phenotypes to seven Arabidopsis genes not previously implicated in pollen tube growth ( Figure S3B ) . Five of these genes ( At2g31550 , At4g08670 , At5g23530 , At5g55020 , At1g60420 ) were not characterized genetically before this study . In these mutants , we confirmed that the T-DNA disrupted the gene of interest using gene-specific PCR and showed that this PCR product cosegregated with two reporter genes ( Basta and GUS expression ) carried on the T-DNA . All pollen assays directly compare the function of pollen with the T-DNA insert ( GUS+ and carrying Basta gene ) with wild-type pollen ( GUS- and not carrying Basta gene ) as they were performed in pollen tetrads from heterozygous plants . The mutant phenotypes we identified are linked to the T-DNA insertion . Therefore , we can rule out the possibility that unlinked mutations , not tagged by the T-DNA in the gene of interest , are responsible for the observed phenotypes . Our data suggest that loss-of-function of the indicated genes caused the pollen phenotypes recorded here . In this study , we systematically addressed whether mutations that affect pollen tube growth in vitro also disrupt pollen tube growth in vivo . It is reasonable to predict that a mutation affecting a pollen tube structural component or a factor required for tip growth would disrupt growth in either the pistil or in a defined growth medium [34] , [35] , [61] . However , a mutation that specifically disrupts the ability of the pollen tube to re-orient growth in response to pollen tube guidance cues would not be expected to cause a defect in the ability of pollen tubes to extend in vitro . We found two insertions ( At3g18000-1 [XIPOTL]; At1g60420-1 [thioredoxin and C1-domain containing] ) that caused significant reductions in ovule targeting ( Table 4 , Figure 5 ) , but did not affect pollen tube growth in vitro ( Table 5 , Figure S4 ) . A third type of mutation would cause mutant phenotypes in vitro , but would not result in defective growth in the pistil environment . These mutations may define genes that play a role in the growth process , but whose mutant phenotypes in the pistil are masked by factors in the pistil environment that enhance growth [24] . For these mutations , in vitro pollen tube growth may be viewed as a sensitized environment capable of revealing subtle mutant phenotypes . We identified five insertions that disrupted pollen tube growth in vitro that did not obviously affect the ability of pollen to sire progeny in vivo . For example , At5g55020-1 ( MYB120 , discussed below ) significantly reduced pollen germination and tube length in vitro ( Table 5 , Figure S4 ) , but did not affect transmission of the mutant allele through pollen ( Table 4 ) . Our microarray data show that the transcriptome of pollen tubes grown in vitro is dramatically different from that of pollen tubes grown through pistil tissue . Our genetic experiments also confirm this difference by showing that the consequences of loss-of-function in a pollen tube gene are different in these distinct environments and that the combination of assays probing growth in vitro and in vivo is essential to comprehensively understand pollen tube growth . An insertion in MYB120 ( At5g55020-1 ) caused defective pollen germination and tube growth in vitro ( Table 5 , Figure S4 ) . Phylogentic analysis of 125 MYB-related transcription factors place MYB120 in subgroup 18 , which comprises seven closely related genes [62] , [63] . Analysis of 125 Arabidopsis MYBs in our data set showed that three of the four most abundant MYBs in the SIV PT are from subgroup 18 ( including MYB120 ) . Furthermore , four members of subgroup 18 are expressed at much higher levels in pollen than in other tissues we analyzed and three members ( including MYB120 and MYB65 ) of the subgroup have their peak expression in SIV PT ( Figure S5 ) . MYB65 is a SIV PT-enriched gene ( Table S8 ) and was identified among the transcription factors overrepresented in SIV PT-enriched genes compared to pollen-expressed genes ( Table S9 ) . Perhaps functional redundancy within this MYB subgroup explains why an insertion in MYB120 ( At5g55020-1 ) affected pollen tube growth in vitro did not cause a defect in the pistil ( Table 4 ) . Analyzing single and multiple mutations in members of this subgroup , using the assays described here , can be used to test this hypothesis and to determine whether this group of transcription factors is an important regulator of gene expression in actively extending pollen tubes . We uncovered a role in pollen tube growth for two genes ( At5g67250 , At3g18000 ) already shown to be critical for sporophytic growth and development . Our microarray analysis shows that both of these genes have broad expression patterns in sporophytic tissues and are significantly higher in SIV PT than 4 h PT ( Figure S3B ) . These expression patterns underscore an important aspect of the SIV PT transcriptome; 871 genes ( Figure 2D ) are shared between SIV PT and the sporophytic tissues we analyzed that are not expressed in pollen or pollen tubes grown in vitro . At5g67250 and At3g18000 illustrate how functional analysis using pollen can provide new insights into the function of this part of the Arabidopsis genome . Previous RNAi analysis of At5g67250 ( an SCF-type F-box and leucine rich repeat-containing E3 ubiquitin ligase , VFB-4 ) showed that reduction of expression was associated with defects in lateral root formation and rosette leaf expansion [64] . SCF-type E3 ubiquitin ligases determine substrate specificity for ubiquitination and proteolysis , thereby regulating an array of biological processes including cell cycle progression [65] , [66] and auxin signaling [67] , [68] . Here we have shown that VFB-4 is required for pollen tube growth in vitro ( Table 5 , Figure S4 ) , suggesting that regulated proteolysis is important for pollen tube extension . A limitation of microarray analysis is that it only documents changes in mRNA abundance and does not identify genes whose mRNA levels remain unaltered , but encode proteins that undergo post-translational modification in response to growth in the pistil . Post-translational regulation of protein function is likely an important mediator of pollen-pistil interactions . For example , LePRK2 , a pollen-specific receptor kinase required for pollen tube growth in tomato [69] has been shown to be dephosphorylated by a stigma extract [70] . Methods developed for large-scale SIV PT isolation ( this study ) and pollen proteomic analysis [71]–[73] , could be combined to identify the set of pollen tube proteins that are modified in response to growth in the pistil . At3g18000 ( XIPOTL ) was identified in a genetic screen for root architecture defects [52] , and encodes a PEAMT required for production of phosphatidylcholine ( see Results ) . At3g18000-1 disrupted ovule targeting in the pistil ( Table 4 , Figure 5 ) , suggesting that XIPOTL may be required for navigating the pistil environment and that lipid signaling and/or a particular plasma membrane composition is required for pollen tube growth and guidance .
Arabidopsis plants were grown in chambers at 21°C under illumination ( 100 µmol m−2 s−1 with a 16-hour photoperiod ) . Wild-type pollen and pollen tubes ( Col-0 accession ) were used for microarray experiments . The SAIL lines ( Col-0 accession ) and male sterile 1 mutant , ms1 ( CS75 , Landsberg ecotype ) were obtained from the Arabidopsis Biological Resource Center ( Columbus , OH ) . ms1 does not produce pollen , but has a normal pistil; this mutant therefore yields pistils that do not require emasculation . Dry pollen grains were collected by the vacuum method [74] into microfuge tubes containing 250 µl liquid pollen growth medium [75] and incubated for 0 . 5 or 4 hours in a 24°C growth chamber . Pollen tubes were centrifuged at 4 , 000 rpm for 5 minutes , the supernatant was removed and the microfuge tubes were frozen in liquid nitrogen and stored in −80°C until RNA isolation . Aliquots of pollen tube suspensions were observed under an Axiovert 100 microscope ( Carl Zeiss , Oberkochen , Germany ) to determine % pollen germination and pollen tube length using Metamorph software version 7 . 1 . 4 . 0 ( Molecular Devices Inc . , Downingtown , PA ) . Pollen grains with emerging tubes equal to or longer than their diameters were considered germinated . After 4 hours of growth , 58 . 3±6 . 0% of the pollen grains germinated and formed tubes ( average length of 383 . 8±32 . 1 µm; Figure 1C ) . Under our growth conditions , the remainder of the grains in the 4 hour sample did not germinate and the germination rates did not increase even with longer incubation times ( Figure 1C ) . Pollen tubes grown through the stigma and style were collected by the semi in vivo procedure essentially as described [26] . Pollinated ms1 pistils were placed vertically on solid pollen growth medium for one hour ( establish growth into the pistil ) before they were laid horizontally; tubes emerged from cut pistils after three hours of growth and were harvested as bundles after three hours of growth on the media surface . Bundles were excised at the point of emergence from the cut pistil and collected into liquid nitrogen-frozen microfuge tubes . Eight hundred pollen tube bundles ( obtained from 800 cut pistil explants ) were used for each of the three replicate RNA isolations . Pollen tube bundles were confirmed to be free of pistil tissue contamination by microscopy . For 8-day-old seedling samples , both shoots and roots of seedlings grown on 0 . 5X Murashige and Skoog ( MS ) media [MS salts ( Carolina Biological Supply Company , Burlington , North Carolina ) , 10% sucrose , pH 5 . 7 , 7% Bacto Agar] were included . However , for 21-day-old seedling samples , only aerial parts of the plants grown on soil were included . For in vivo confirmation of gene expression experiments , flower stage 14 [76] ms1 pistils were hand pollinated with wild-type ( Col-0 ) pollen . Pollinated pistils , either 1 minute or 2 hours after pollination , were cut at the junction of the style and ovary and the stigma and style portion ( cut pistils ) were used for RNA isolation . For unpollinated pistils , flower stage 14 [76] ms1 cut pistils devoid of any pollen were used . Fifteen cut pistils of each kind were used for each replicate ( 2 ) RNA isolation . Total RNA was extracted from dry pollen , 0 . 5 h PT , 4 h PT and SIV PT using the Qiagen RNeasy kit ( http://www . qiagen . com ) . The yield and RNA purity were determined by Nano-Drop ( Thermo Scientific , Wilmington , DE , USA ) and gel electrophoresis . RNA integrity was checked using an Agilent 2100 Bioanalyzer ( Agilent Technologies , Boblingen , Germany ) . Hybridization and post hybridization processing were performed as per the manufacturer's instructions by the Arizona Cancer Center Microarray facility ( http://www . azcc . arizona . edu/laboratory/l_microarray . htm ) . Total RNA ( 5 µg , dry pollen , 0 . 5 h PT and 4 h PT ) and 2 µg ( SIV PT ) was processed as per the Affymetrix GeneChip Expression Analysis protocol ( Part#701071 , Rev 5 , Affymetrix , Santa Clara , CA ) . Briefly , after first and second strand cDNA synthesis with total RNA , the cDNAs were used to generate cRNA labeled with biotin in an in vitro transcription reaction . For each pollen condition , labeled cRNA was fragmented and 15 µg of fragmented cRNA ( 25–200 nt as per Agilent 2100 Bioanalyzer RNA 6000 Nano Chip Series II Assay , Agilent Technologies , Waldbronn , Germany ) was hybridized to the GeneChip Arabidopsis ATH1 genome arrays ( http://www . affymetrix . com ) for 20 hours at 45°C . Standard washing and staining procedures were performed using the GeneChip Fluidics Station 450 ( Affymetrix , Santa Clara , CA ) . The arrays were then scanned using the GeneChip Scanner 3000 with 7 G upgrade ( Affymetrix , Santa Clara , CA ) . Signal intensities from each of the 15 arrays were converted to raw expression data ( with Present , “P” , Absent , “A” and Marginal , “M” scores ) using GeneChip Operating Software ( GCOS ) ( Affymetrix , Santa Clara , CA ) and are provided as supplementary files ( Table S1 , Table S2 , Table S3 , Table S4 ) . Raw data ( . CEL and CHP files ) from all 15 microarrays reported in this study have been deposited in Gene Expression Omnibus [77] public repository and can be accessed from ( http://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE17343 ) using the Series accession number GSE17343 . In addition to the 15 arrays from this study , we obtained 25 publicly available array data ( AtGenExpress , http://www . ebi . ac . uk [28] and stigma and ovary microarray data from Gene Expression Omnibus at http://www . ncbi . nlm . nih . gov/geo/[29] ) . The following cel files were downloaded: 7 day old roots–ATGE_3A-C , 17 day old roots–ATGE_9A-C , rosette leaves–ATGE_17A-C , pollen–ATGE_73A-C , 8 day old seedlings–ATGE_96A-C , 21 day old seedlings–ATGE_100A-C . Probe data for ovary–GSM67078 . cel , GSM67079 . cel , GSM67080 . cel , SM67081 . cel and stigma–GSM67084 . cel , GSM67086 . cel , GSM67087 . cel . Using RMA ( Robust Microarray Analysis tool in the affy library ) [78] , we normalized the data from all of these 40 arrays ( Table S5 ) . Using the affy and limma BioConductor libraries ( http://www . bioconductor . org ) and the R programming project ( http://www . R-project . org ) , we calculated the statistical significance in expression level changes of the following comparisons: 0 . 5 h PT vs . dry pollen; 4 h PT vs . dry pollen; 4 h PT vs . 0 . 5 h PT; SIV PT vs . dry pollen; SIV PT vs . 4 h PT . After estimating the variance of mean signal intensities for each probe set , the significance of this value in the two conditions was evaluated by performing a t-test [79] . The probabilities obtained were corrected for multiple hypothesis testing by reshuffling the data to obtain an estimate of the false discovery rate ( B values ) and applying this estimate to lower the probability of the t-value ( adjusted P values ) . The complete results from the statistical analysis for each of the five comparisons and for every probe set in an array are also reported in Table S5 . To investigate the relationships among pollen samples , agglomerative hierarchical clustering of the fifteen microarrays representing four pollen conditions , was performed as described [43] , [80] . To find genes that had the best discriminative ability , based on its expression profile , we employed non-hierarchical k-means clustering method [43] . For this , we compared dry pollen and 4 h PT hierarchical clusters to the SIV PT hierarchical cluster and in each comparison , for every gene , we calculated discriminative weight , a parameter that measures the ability of a gene's expression values to distinguish two clusters . The discriminative weight of the gene for a pair of clusters is defined by , where dB is the distance between centers of the clusters , dwi is the average Euclidean distance among all sample pairs within cluster i , where ti is the total number of sample pairs in the cluster i . GO term enrichment analysis reported in Figure 2 was performed essentially as described in [81] , [82] . Briefly , the hypergeometric distribution test was applied on the gene sets in sectors 1–4 ( Figure 2B ) using the GOHyperGAll function [81] which yielded raw and Bonferroni corrected p-values ( adjusted p value ) . GO terms from the “Molecular Function” category that had an adjusted P value of <0 . 05 were considered highly enriched and are shown in Figures 2C and 2E . Arabidopsis gene-to-GO mappings were downloaded from the GO site ( 10/12/2007 release; http://geneontology . org ) . A complete list of GO-terms ( for all three broad GO categories ) associated with sectors 1–4 are provided in Table S7 . Within each sector , only unique genes belonging to each GO term category were considered . For the total number of genes for each GO term category reported in Figure 2C , unique numbers of genes from each sector were added , without eliminating gene overlap between sectors . The Fisher exact test was performed to determine if any GO term was significantly overrepresented in SIV PT-enriched genes ( Table 1 ) , given that this gene list was smaller compared to those used in Figure 2C [83] . From the SIV PT-enriched gene list ( 383 , Table S8 ) , we excluded genes that were also expressed in pollen samples ( ATGE_73A-C ) ; the remaining 357 probes were mapped to genes . Only single probes were chosen if multiple probes mapped to the same gene or a gene family . This criteria resulted in a final list of 349 SIV-enriched genes that was then compared to pollen-expressed genes ( ATGE_73A-C , [28] to obtain P-values for overrepresented GO terms in SIV PT-enriched genes . The GO terms with P<0 . 05 were considered significantly overrepresented in SIV PT-enriched genes and reported in Table 1 . Complete results of this analysis and the genes associated with GO-terms listed in Table 1 are provided in Table S9 . For each of the RT-PCR experiments , new RNA samples were isolated from the indicated cells/tissues , cDNA was synthesized and used as template for PCR ( Figure S2 ) and quantitative real-time PCR ( Figure 3 and Table 3 ) . Total RNA was isolated using the Qiagen RNeasy kit followed by treatment with Fermentas DNase I ( http://www . fermentas . com ) prior to first-strand cDNA synthesis using Invitrogen ThermoScript RT-PCR kit ( http://www . invitrogen . com ) . PCR ( with PowerTaq DNA Polymerase PCR system , Altila Biosystems , Palo Alto , CA ) was performed as follows: 3 minutes ( min ) at 94°C , 38 cycles of 30 seconds ( sec ) at 94°C , 1 min at 60°C and 1 min at 72°C , followed by 5 min at 72°C . Real-time RT-PCR was performed using the Roche FastStart DNA Master SYBR Green I master mix ( http://www . roche . com ) in a LightCycler system ( Roche , http://www . roche . com ) . The PCR primers used in RT-PCR and qRT-PCR experiments are listed in Table S17 . The PCR cycle conditions used for real time PCR were as follows: a 95°C for 5 min followed by 45 cycles of 95°C for 10 sec , 60°C for 15 sec , and 72°C for 15 sec . For each gene analyzed by RT-PCR and qRT-PCR , four reactions were carried out , including two technical replicates and two biological replicates ( using RNA from independently harvested tissues ) . In each qRT-PCR run , ACTIN2 ( Threshold Cycle ( CT ) value of 18–19 ) , was used to normalize for mRNA levels . We considered a gene to be expressed only if it had a CT value <36 . When expression was not detected in a qRT-PCR reaction , a CT value of 45 ( since 45 cycles were used in a real time PCR reaction ) was used to calculate the fold change . Manually self-pollinated ms1 pistils were harvested either one minute or 2 hours after pollination . The pollinated pistils were stained with aniline blue to visualize in vivo pollen tube growth as described previously [84] . Stained pistils were observed on a Zeiss Axiovert 100 microscope with a Zeiss 365 G filter ( Carl Zeiss , Oberkochen , Germany ) . By this staining procedure , the majority of the pollen tubes reached the style tissue in 2 hours ( Figure 4A ) . SAIL lines were chosen ( http://signal . salk . edu/[85] ) with insertions between 300 bp upstream of the 5′ UTR and 300 bp downstream of the 3′UTR ( exons were prioritized over introns ) ; and for which there was a TAIL PCR sequence that corroborated the T-DNA insert site to a single locus in the Arabidopsis genome . Determination of Basta was performed as reported [49] . Basta plants were transferred to soil and T-DNA insertion sites were confirmed using left border ( LB3 , LB2 , and/or LB1 [86] ) and gene-specific ‘right’ primers ( Table S16 , designed using http://signal . salk . edu/tdnaprimers . 2 . html ) in a PCR reaction . The PCR program used for this reaction was: 94°C for 5 min followed by 36 cycles of 94°C for 15 sec , 60°C for 30 sec , 72°C for 2 min , and a final elongation step of 72°C for 4 min . Pollen tetrads from one stage 14 flower [76] from each Basta plant were stained and assayed for segregation of the LAT52:GUS transgene as described [49] . Transmission of the T-DNA following self-fertilization or through crosses to ms1 were tested as described [49] . Pollen grains were incubated in liquid pollen growth medium [53] on upside-down slides [87] for 6 hours and stained for GUS activity as described [49] . Images were captured ( Zeiss Axiovert 200 M , Carl Zeiss , Oberkochen , Germany ) and used to determine pollen tube germination rates and pollen tube length for GUS- and GUS+ pollen using ImageJ software ( http://rsbweb . nih . gov/ij/docs/faqs . html ) . These experiments were analyzed as a randomized complete block design with dates of observation as blocks and genotype of pollen ( GUS+ , insertion; GUS- , wild type ) as treatments . Pollen germination and pollen tube length values were subjected to mixed-model analysis of variance with block considered a random effect and treatment a fixed effect . Untransformed least-squares means and P values [88] from this analysis are reported ( Table 5 ) . Analysis was done using PROC MIXED in SAS/STAT Version 9 . 1 of the SAS System for Windows . ( Copyright © 2002-2003 SAS Institute Inc ) . Pollen tube growth in the pistil was examined after crossing pollen from heterozygous insertion plants to three or more ms1 pistils . Pollinated pistils were harvested 24 hours after pollination , prepared for GUS staining and microscopy observations as described previously [49] . Stained pistils were imaged ( differential interference contrast ) using a Zeiss Axiovert 200 M microscope ( Carl Zeiss , Oberkochen , Germany ) .
|
For successful reproduction in flowering plants , a single-celled pollen tube must rapidly extend through female pistil tissue , locate female gametes , and deliver sperm . Pollen tubes undergo a dramatic transformation while growing in the pistil; they grow faster compared to tubes grown in vitro and become competent to perceive and respond to navigation cues secreted by the pistil . The genes expressed by pollen tubes in response to growth in the pistil have not been characterized . We used a surgical procedure to obtain large quantities of uncontaminated pollen tubes that grew through the pistil and defined their transcriptome by microarray analysis . Importantly , we identify a set of genes that are specifically expressed in pollen tubes in response to their growth in the pistil and are not expressed during other stages of pollen or plant development . We analyzed mutants in 33 pollen tube–expressed genes using a sensitive series of pollen function assays and demonstrate that seven of these genes are critical for pollen tube growth; two specifically disrupt growth in the pistil . By identifying pollen tube genes induced by the pistil and describing a mutant analysis scheme to understand their function , we lay the foundation for functional genomic analysis of pollen–pistil interactions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"biology/plant",
"genetics",
"and",
"gene",
"expression",
"developmental",
"biology/plant",
"growth",
"and",
"development",
"plant",
"biology/plant",
"genetics",
"and",
"gene",
"expression",
"genetics",
"and",
"genomics/plant",
"genetics",
"and",
"gene",
"expression"
] |
2009
|
Penetration of the Stigma and Style Elicits a Novel Transcriptome in Pollen Tubes, Pointing to Genes Critical for Growth in a Pistil
|
Recent large cancer studies have measured somatic alterations in an unprecedented number of tumours . These large datasets allow the identification of cancer-related sets of genetic alterations by identifying relevant combinatorial patterns . Among such patterns , mutual exclusivity has been employed by several recent methods that have shown its effectiveness in characterizing gene sets associated to cancer . Mutual exclusivity arises because of the complementarity , at the functional level , of alterations in genes which are part of a group ( e . g . , a pathway ) performing a given function . The availability of quantitative target profiles , from genetic perturbations or from clinical phenotypes , provides additional information that can be leveraged to improve the identification of cancer related gene sets by discovering groups with complementary functional associations with such targets . In this work we study the problem of finding groups of mutually exclusive alterations associated with a quantitative ( functional ) target . We propose a combinatorial formulation for the problem , and prove that the associated computational problem is computationally hard . We design two algorithms to solve the problem and implement them in our tool UNCOVER . We provide analytic evidence of the effectiveness of UNCOVER in finding high-quality solutions and show experimentally that UNCOVER finds sets of alterations significantly associated with functional targets in a variety of scenarios . In particular , we show that our algorithms find sets which are better than the ones obtained by the state-of-the-art method , even when sets are evaluated using the statistical score employed by the latter . In addition , our algorithms are much faster than the state-of-the-art , allowing the analysis of large datasets of thousands of target profiles from cancer cell lines . We show that on two such datasets , one from project Achilles and one from the Genomics of Drug Sensitivity in Cancer project , UNCOVER identifies several significant gene sets with complementary functional associations with targets . Software available at: https://github . com/VandinLab/UNCOVER .
Several recent methods have used mutual exclusivity signals to identify sets of genes important for cancer [24] . RME [25] identifies mutually exclusive sets using a score derived from information theory . Dendrix [26] defines a combinatorial gene set score and uses a Markov Chain Monte Carlo ( MCMC ) approach for identifying mutually exclusive gene sets altered in a large fraction of the patients . Multi-Dendrix [27] extends the score of Dendrix to multiple sets and uses an integer linear program ( ILP ) based algorithm to simultaneously find multiple sets with mutually exclusive alterations . CoMET [18] uses a generalization of Fisher exact test to higher dimensional contingency tables to define a score to characterize mutually exclusive gene sets altered in relatively low fractions of the samples . WExT [18] generalizes the test from CoMET to incorporate individual gene weights ( probabilities ) for each alteration in each sample . WeSME [28] introduces a test that incorporates the alteration rates of patients and genes and uses a fast permutation approach to assess the statistical significance of the sets . TiMEx [29] assumes a generative model for alterations and defines a test to assess the null hypothesis that mutual exclusivity of a gene set is due to the interplay between waiting times to alterations and the time at which the tumor is sequenced . MEMo [17] and the method from [30] employ mutual exclusivity to find gene sets , but use an interaction network to limit the candidate gene sets . The method by [31] and PathTIMEx [32] introduce an additional dimension to the characterization of inter-tumor heterogeneity , by reconstructing the order in which mutually exclusive gene sets are mutated . None of these methods take quantitative targets into account in the discovery of significant gene sets and sets showing high mutual exclusivity may not be associated with target profiles ( Fig 1 ) . [33] recently developed the repeated evaluation of variables conditional entropy and redundancy ( REVEALER ) method , to identify mutually exclusive sets of alterations associated with functional phenotypes . REVEALER uses as objective function ( to score a set of alterations ) a re-scaled mutual information metric called information coefficient ( IC ) . REVEALER employs a greedy strategy , computing at each iteration the conditional mutual information ( CIC ) of the target profile and each feature , conditioned on the current solution . REVEALER can be used to find sets of mutually exclusive alterations starting either from a user-defined seed for the solution or from scratch , and [33] shows that REVEALER finds sets of meaningful cancer-related alterations . In this paper we study the problem of finding sets of alterations with complementary functional associations using alteration data and a quantitative ( functional ) target measure from a collection of cancer samples . Our contributions in this regard are fivefold . First , we provide a rigorous combinatorial formulation for the problem of finding groups of mutually exclusive alterations associated with a quantitative target and prove that the associated computational problem is NP-hard . Second , we develop two efficient algorithms , a greedy algorithm and an ILP-based algorithm to identify the set of k genes with the highest association with a target; our algorithms are implemented in our method fUNctional Complementarity of alteratiOns discoVERy ( UNCOVER ) . Third , we show that our algorithms identify highly significant sets of genes in various scenarios; in particular , we compare UNCOVER with REVEALER on the same datasets used in [33] , showing that UNCOVER identifies solutions of higher quality than REVEALER while being on average two order of magnitudes faster than REVEALER . Interestingly , the solutions obtained by UNCOVER are better than the ones obtained by REVEALER even when evaluated using the objective function ( IC score ) optimized by REVEALER . Fourth , we show that the efficiency of UNCOVER enables the analysis of large datasets , and we analyze a large dataset from Project Achilles , with thousands of genetic dependencies measurements and tens of thousands of alterations , and a large dataset from the Genomics of Drug Sensitivity in Cancer ( GDSC ) project , with hundreds of drug sensitivity measurements and tens of thousands of alterations . On such datasets UNCOVER identifies several statistically significant associations between target values and mutually exclusive alterations in genes sets , with an enrichment in well-known cancer genes and in known cancer pathways .
The workflow of our algorithm UNCOVER is presented in Fig 2 . UNCOVER takes in input information regarding 1 . the alterations measured in a number of samples ( e . g . , patients or cell lines ) , and 2 . the value of the target measure for each patient . UNCOVER then identifies the set of mutually exclusive alterations with the highest association to the target , and employs a permutation test to assess the significance of the association . Details regarding the computational problem and the algorithms used by UNCOVER are described in the following sections . The implementation of UNCOVER is available at https://github . com/VandinLab/UNCOVER .
We ran UNCOVER on the GDSC dataset for k = 3 and compared the results obtained when the target values are not considered in the analysis , obtained running UNCOVER ILP with k = 3 while setting the target values to 1 for all the samples considered in the analysis of a target ( S1 Table ) . The latter analysis corresponds to the extraction of sets with high mutual exclusivity ( e . g . , by [34] ) . As expected , the solutions obtained in the two cases are very different: the solution obtained without considering the target values has one alteration in common with the solution obtained by UNCOVER using either positive or negative values of the target for only 11 targets of the 265 in the GDSC dataset , and for no target the solutions share more than 1 alteration . An example of the solutions obtained target using UNCOVER and without considering the target values are shown in Fig 3 . We observe that while the solutions obtained considering the target values display an association with the target profile ( positive or negative ) , the solution obtained when the target values are not considered , while covering a large set of samples , does not display any positive or negative association with the target profile . To asses the association between target values and alterations more consistently we calculated the point biserial coefficient [40] for all 265 solutions . The coefficient varies between −1 and +1 with 0 implying no correlation . The average value obtained when ignoring the target is −0 . 02 with standard deviation 0 . 05 , while the the average value obtained by UNCOVER is 0 . 20 with standard deviation 0 . 05 . These results show that a mutual exclusivity analysis that disregards the values of the target does not identify sets of mutually exclusive alterations associated with target values . In addition , the genes in solution identified by considering the drug target have a much more significant enrichment in known cancer genes , as reported in [11] , than the genes in solution identified disregarding the values of the target ( p = 3 × 10−12 vs p = 10−2 ) . We run the greedy algorithm and the ILP from UNCOVER on the same four datasets considered by the REVEALER publication [33] . We used the same values of k used in [33] , that is k = 3 for all the datasets , except from the KRAS dataset where k = 4 was used . For each dataset we recorded the solution reported by the greedy algorithm , the solution reported by the ILP , the value of the objective functions for such solutions and the running time to obtain such solutions . For ILP solutions , we also performed the permutation test ( see Materials and methods ) to compute a p-value using 1000 permutations . The results are reported in Table 1 , in which we also show the results from REVEALER ( without initial seeds ) . Fig 4 shows alteration matrices and the association with the target for the solutions identified by UNCOVER . We can see that the greedy algorithm identifies the same solution of the ILP based algorithm in three out of four cases , and that the runtime of the ILP and the runtime of greedy algorithm are comparable and very low ( < 40 seconds ) in all cases . In contrast , the running time of REVEALER is much higher ( > 1000 seconds in most cases ) . ( We included all preprocessing in the reported UNCOVER runtimes in Table 1 to ensure a fair comparison with REVEALER; not including preprocessing our running times are all under 10 seconds ) . Comparing the alteration matrices of the solutions by UNCOVER and the ones of solutions by REVEALER ( S1 Fig ) we note that alterations in solutions by UNCOVER tend to have higher mutual exclusivity and to be more concentrated in high weight samples than alterations in solutions by REVEALER . As expected , the value of the objective function we use is much lower for solutions from REVEALER than for solutions from our algorithm . We then compared the solutions obtained by our algorithms with the solutions from REVEALER in terms of the information coefficient ( IC ) , that is the target association score used in [33] as a quality of the solution . Surprisingly , in two out of four datasets UNCOVER , which does not consider the IC score , identifies solutions with IC score higher ( by at least 5% ) than the solutions reported by REVEALER . For the other two cases , in one dataset the IC score is very similar ( 0 . 50 vs 0 . 49 ) while in the other case the IC score by REVEALER is higher ( 0 . 7 vs 0 . 67 ) but the solution reported by REVEALER differs from the solution reported by UNCOVER by 1 gene only . Interestingly , the latter is the only case where the solution from the ILP has a p-value > 0 . 1 ( p < 0 . 03 in all other cases ) , and therefore the solutions ( by our methods and by REVEALER ) for such dataset may be , at least in part , due to random fluctuations of the data . In terms of biological significance , in most cases the solutions by UNCOVER and by REVEALER are very similar , with cancer relevant genes identified by both methods . For NFE2L2 activation , both methods identify KEAP1 , a repressor of NFE2L2 activation [41] . For MEK-inhibitor , both methods find BRAF , KRAS , and NRAS , three well known oncogenic activators of the MAPK signaling pathway , which contains MEK as well . For KRAS essentiality , both methods report mutations in KRAS in the solution . For β-catenin activation , both methods identify CTNNB1 mutations and APC mutations , that is known to be associated to β-catenin activation [42] . These results show that UNCOVER identifies relevant biological solutions that are better than the ones identified by REVEALER when evaluated using our objective function and also when evaluated according to the objective function of REVEALER with a running time that is on average two orders of magnitude smaller than required by REVEALER . Since UNCOVER and REVEALER consider two different objective functions , it is unclear whether the improvement in running time comes from differences in implementation choices or from a inherently different computational complexity . However , since UNCOVER’s objective function is easier to compute than REVEALER’s objective function , we believe that the use of our objective function plays an important role in the efficiency of UNCOVER . We also compared the solutions obtained by UNCOVER and by REVEALER on the GDSC dataset ( S2 Table ) . For both algorithms we obtained the solutions for k = 3 . For UNCOVER , we considered the solution returned by the ILP . For REVEALER , we could only obtain solutions for 246 targets , since for the other targets REVEALER terminated with an error message . Due to the high running time of REVEALER , we only obtained sets of alterations associated with positive values of the target ( Table 2 ) . For 33 targets the solution by UNCOVER and the solution by REVEALER share 1 alteration , while for 33 targets the solution by UNCOVER and the solution by REVEALER share 2 alterations; for no target UNCOVER and REVEALER report the same solution . This shows that the two methods identify completely different solution in most ( > 73% ) of the cases . We compared the solutions obtained by UNCOVER and by REVEALER using the IC score considered by REVEALER but not from UNCOVER: surprisingly , in more than 50% of the cases ( 113 out of 208 ) the IC score of the solution from UNCOVER is higher than the IC of the solution from REVEALER . On the other hand , for all targets the solution by REVEALER is worst than the solution by UNCOVER when the UNCOVER objective function is considered . We also compared UNCOVER and REVEALER evaluating the association between target values and alterations in the solutions using a measure of association that is not considered by the two algorithms . In particular , we considered the point biserial correlation coefficient [40] . In more than 95% of the cases ( 199 out of 208 ) the point biserial correlation coefficient between the solution from UNCOVER and the target is higher than the point biserial correlation coefficient between the solution from REVEALER and the target , that is , the solution from UNCOVER has an higher association with the target than the solution from REVEALER . On average , the solution from UNCOVER has a point biserial correlation coefficient that is 37% higher than the point biserial correlation coefficient of the solution from REVEALER . Moreover , the average effect size of solutions from UNCOVER is more than 80% higher than the average effct size of solutions from REVEALER ( Table 2 ) . In addition , the genes in solutions from UNCOVER have a much higher enrichment ( p = 3 × 10−13; 7-fold enrichment ) for known cancer genes than solutions from REVEALER ( p = 2 × 10−4; 3-fold enrichment ) . Analogously , more KEGG pathways display a significant enrichment in genes from UNCOVER solutions than from REVEALER solutions ( 22 vs 11 ) . We also compared the running time of the two methods: UNCOVER required 3 hours to complete the analysis , while REVEALER required 9 days . Overall , these results show that UNCOVER obtains better results than REVEALER not only in terms of the UNCOVER objective function but also in terms of the score from REVEALER as well as in terms of a independent measure of association , while being 70 times faster than REVEALER . For each combination we generated 10 simulated datasets as described in Materials and methods . Each dataset contains a planted set of 5 alterations associated with the target . We used both the greedy algorithm and the ILP from UNCOVER with k = 5 to attempt to find the 5 correct alteration , and evaluated our algorithms both in terms of fraction of the correct ( i . e . , planted ) solution reported and running time . As shown in Fig 5 , the greedy algorithm is faster than the ILP for all datasets , and the difference in running time increases as the number m of samples increases , with the runtime of the greedy algorithm being almost two orders of magnitude smaller than the runtime of the ILP for m = 1000 samples . In addition , for a fixed number of samples and alterations , the running time of the greedy algorithm is constant , that is it does not depend on the properties of the planted solution , while the running time of the ILP varies greatly depending on these parameters . For m = 10 , 000 samples the running time of the ILP becomes extremely high , so we restricted to consider only two sets of parameters ( p − n = 0 . 95 and p − n = 0 . 2 ) . In this case the ILP took between 44 minutes and 7 hours to complete , while the greedy algorithm terminates in 5 minutes . In terms of the quality of the solutions found , as expected the ILP outperforms the greedy ( Fig 6 ) but the difference among the two tends to disappear when the number of samples is higher . In addition , since the ILP finds the optimal solution , we can see that for a limited number of samples we may not reliably identify the planted solution with 200 samples unless the planted solution appears almost only in positive targets and in almost all of them ( p − n = 0 . 95 ) , while for m = 1000 we can reliably identify the planted solution using both the ILP and the greedy algorithm even when the association with the target is weaker ( p − n = 0 . 6 ) . When m = 10 , 000 , both the ILP and the greedy algorithm perform well in terms of the quality of the solution: the ILP finds the correct alterations on every experiment and the greedy identifies the whole planted solution in all experiments but one for p − n = 0 . 2 , for which it still reports a solution containing 4 out of 5 genes in the planted solution . These results show that for a large number of samples the greedy algorithm reliably identifies sets of alterations associated with the target , as predicted by our theoretical analysis , and is much faster than the ILP . For smaller sample size the ILP identifies better solutions than the greedy and has a reasonable running time . The efficiency of UNCOVER renders the analysis of a large number of targets , such as the ones available through the Achilles project , possible . After preprocessing the dataset included 5690 functional phenotypes as targets , and for each of these the CCLE provides alteration information for 205 samples and 31137 alterations . In total we have therefore run 11380 instances ( i . e . , 5690 targets screened for positive and for negative associations ) looking for both positive and negative association with target values . Since the number of samples ( 205 ) is relatively small , we have run only the ILP from UNCOVER on the whole Achilles dataset and looked for solutions with k = 3 genes . The runtime of UNCOVER to find both positive and negative associations , including preprocessing , is 24 hours . Based on the runtime required on the instances reported in [33] ( see the Section Comparison with REVEALER ) , running REVEALER on this dataset would have required about 5 months of compute time . To identify statistically significant associations with targets in the Achilles project dataset we used a nested permutation test . We first run the permutation test with 10 permutations on all instances ( i . e . , on all targets for both positive association and negative association ) . We then considered all the instances with the lowest p-value ( 1/11 ) and performed a permutation test with 100 permutations only for such instances . We the iterated such procedure once more , selecting all the instances with lowest p-value ( 1/101 ) and performing a permutation test with 1000 permutations only for such instances . For positive association we found 60 solutions with p-value < 0 . 001 , and for negative association we found 102 solutions with p-value < 0 . 001 . The solutions with p-value < 0 . 001 ( with 1000 permutations ) are reported in S3 Table . See S2 Fig for some corresponding alteration matrices . The genes in the solutions by UNCOVER with p-value 1/1001 are enriched ( p = 2 × 10−12 by Fisher exact test; 8 fold enrichment ) for well-known cancer genes . We also tested whether genes in solutions by UNCOVER ( with p-value 1/1001 ) are enriched for interactions , by comparing the number of interactions in iRefIndex [43] among genes in such solution with the number of interactions in random sets of genes of the same cardinality . Genes in solutions by UNCOVER are significantly enriched in interactions ( p = 7 × 10−3 by permutation test; 2 fold enrichment ) . In addition , the genes in solutions by UNCOVER are also enriched in genes in well-known pathways: 12 KEGG pathways [44] have a significant ( corrected p ≤ 0 . 05 ) overlap with genes in solutions by UNCOVER and four of these ( endometrial cancer , glioma , hepatocellular carcinoma , EGFR tyrosine kinase inhibitor resistance ) are cancer related pathways . In addition , the targets ( i . e . , genes ) with solutions of p-value 1/1001 are enriched ( p = 5 × 10−4 by permutation test; 6 fold enrichment ) for interactions in iRefIndex and for well-known cancer genes ( p = 2 × 10−12 by Fisher exact test; 8 fold enrichment ) as reported in [11] . These results show that UNCOVER enables the identification of groups of well known cancer genes with significant associations to important targets in large datasets of functional target profiles . For example , for target ( i . e . , silenced gene ) TSG101 , related to cell growth , UNCOVER identifies the gene set shown in Fig 7 as associated to reduced cell viability . ERBB2 is a well known cancer gene and CDH4 is frequently mutated in several cancer types , and both are associated to cell growth . We use UNCOVER to analyze the GDSC project data , identifying sets of alterations associated with drug sensitivity . After preprocessing , the dataset included 64144 alterations and 265 targets , and for each of these the number of cell lines with available data varied between 240 and 705 . In total we have therefore run 530 instances ( i . e . , 265 targets screened for positive and for negative associations ) looking for both positive and negative association with target values . We used the UNCOVER ILP for all instances to obtain solutions with k = 3 genes . For each solution , we use 100 permutations to compute its p-value . For positive association we found 51 solutions with p-value < 0 . 01 , and for negative association we found 41 solutions with p-value < 0 . 01 . We used the following procedure to focus on the most significant solutions: we run UNCOVER with k = 4 and computed the p-values for the solutions using 100 permutations; we then identified targets whose solution for k = 3 have p-value < 0 . 01 and are contained in the solution for the same target with k = 4 and have p-value p < 0 . 01 for k = 4 . In total , this procedure identifies 23 solutions for positive association and 22 solutions for negative associations . These solutions are reported in S4 Table . The genes in the solutions identified as above are enriched ( p = 9 × 10−10 by Fisher exact test; 20 fold enrichment ) for well-known cancer genes , as reported in [11] . We also tested whether these genes in solutions are enriched for interactions , by comparing the number of interactions in iRefIndex [43] among genes in such solution with the number of interactions in random sets of genes of the same cardinality . Genes in solutions by UNCOVER are significantly enriched in interactions ( p = 2 × 10−2 by permutation test; 6 fold enrichment ) . In addition , these genes are also enriched in genes in well-known pathways: 21 KEGG pathways [44] have a significant ( corrected p ≤ 0 . 05 ) overlap with genes in solutions by UNCOVER and 19 of these are cancer related pathways ( e . g . , ErbB signaling pathway ) or related to drug resistance ( e . g . , EGFR tyrosine kinase inhibitor resistance ) . For Palbociclib , UNCOVER identifies RB1 mutations , GRB7 amplifications , and RB1 deletions with significant association with reduced sensitivity to drug . RB1 is a well known cancer gene . The alterations are shown in Fig 3a . While RB1 mutations and RB1 deletions are significantly associated when considered in isolation ( the association of single alterations with drug sensitivity and the drug targets have been obtained from https://www . cancerrxgene . org/ ) , GRB7 amplification is not associated with the target values when considered in isolation . GRB7 encodes a growth factor receptor-binding protein that interacts with epidermal growth factor receptor ( EGFR ) . Both RB1 and EGFR are related to the cell cycle pathway , that is the pathway target of the compound , and the drug targets ( CDK4 , CDK6 ) as well EGFR are members of the PI3K-AKT pathway . For Sunitinib , UNCOVER identifies mutations in SETD2 , ARHGAP19 , and RB1 , with significant association with reduced sensitivity to drug . The alterations are shown in Fig 8a . RB1 is a well known cancer gene and SETD2 has tumor suppressor functionality . None of these alterations have significant association with drug sensitivity when considered in isolations . RB1 and SETD2 are involved in protein localization to chromatin , and ARHGAP19 is part of Rho mediated remodeling . For PLX-4720-2 , UNCOVER identifies mutations in BRAF , CD244 , and ARSB with significant association to increased sensitivity to drug . The alterations are shown in Fig 8b . BRAF is a well-known cancer gene; it is the target of the compound and BRAF mutations have significant association to increased sensitivity to the compound , while the other two alterations do not . BRAF and CD244 are part of natural killer cell mediated cytotoxicity pathway , while ARSB is involved in the regulation of cell adhesion , cell migration and invasion in colonic epithelium [45] , and is also part of metabolism related pathways . For VX-11e , UNCOVER identifies mutations in BRAF , KRAS , and NRAS , with significant association to increased sensitivity to drug . The alterations are shown in Fig 8c . Only BRAF mutations have significant association with the target when considered in isolation . The pathway target for the compound is the ERK MAPK signaling pathway , to which all three alterations are related . All three genes have well identified roles in cancer . These results show that UNCOVER enables the identification of groups of relevant genes , many related to cancer , with significant associations to important targets in large datasets of drug sensitivity profiles . In this work we study the problem of identifying sets of mutually exclusive alterations associated with a quantitative target profile . We provide a combinatorial formulation for the problem , proving that the corresponding computational problem is NP-hard . We design two efficient algorithms , a greedy algorithm and an ILP-based algorithm , for the identification of sets of mutually exclusive alterations associated with a target profile . We provide a formal analysis for our greedy algorithm , proving that it returns solutions with rigorous guarantees in the worst-case as well under a reasonable generative model for the data . We implemented our algorithms in our method UNCOVER , and showed that it finds sets of alterations with a significant association with target profiles in a variety of scenarios . By comparing the results of UNCOVER with the results of REVEALER on four target profiles used in the REVEALER publication [33] and on a large dataset from the GDSC project , we show that UNCOVER identifies better solutions than REVEALER , even when evaluated using REVEALER objective function . Moreover , UNCOVER is much faster than REVEALER , allowing the analysis of large datasets such as the dataset from project Achilles and from the GDSC project , in which UNCOVER identifies a number of associations between functional target profiles and gene set alterations . Our tool UNCOVER ( as well as REVEALER ) relies on the assumption that the mutual exclusivity among alterations is due to functional complementarity . Another explanation for mutual exclusivity is the fact that each cancer may comprise different subtypes , with different subtypes being characterized by different alterations [27] . UNCOVER can be used to identify sets of mutually exclusive alterations associated with a specific subtype whenever the subtype information is available , by assigning high weight to samples of the subtype of interest and low weight to samples of the other subtypes . In addition , while we consider a penalty based on mutual exclusivity , other types of penalties may be used to identify sets of alterations associated with a target profile . The study of the theoretical properties of the problem and the analysis of the results with different penalties are interesting directions of future research .
|
Sequencing technologies allow the measurement of somatic alterations in a large number of cancer samples . Several methods have been designed to analyze these alterations , but the characterization of the functional effects of such alterations is still challenging . A recent promising approach for such characterization is to combine alteration data with quantitative profiles obtained , e . g . , from genetic perturbations . The analysis of these data is complicated by the extreme heterogeneity of alterations in cancer , with different cancer samples exhibiting vastly different alterations . This heterogeneity is due , in part , to the complementarity of alterations in cancer pathways , with alterations in different genes resulting in the same alteration at the functional level . We develop UNCOVER , an efficient method to identify sets of alterations displaying complementary functional association with a quantitative profile . UNCOVER is much more efficient than the state-of-the-art , allowing the identification of complementary cancer related alterations from genome-scale measurements of somatic mutations and genetic perturbations .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"and",
"discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"employment",
"engineering",
"and",
"technology",
"applied",
"mathematics",
"sociology",
"social",
"sciences",
"simulation",
"and",
"modeling",
"algorithms",
"mutation",
"mathematics",
"pharmacology",
"discrete",
"mathematics",
"combinatorics",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"labor",
"economics",
"communications",
"mass",
"media",
"economics",
"drug",
"discovery",
"encyclopedias",
"permutation",
"drug",
"research",
"and",
"development",
"software",
"engineering",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"preprocessing"
] |
2019
|
Efficient algorithms to discover alterations with complementary functional association in cancer
|
Nitric oxide ( NO ) , a key antimicrobial molecule , was previously shown to exert a dual role in paracoccidioidomycosis , an endemic fungal infection in Latin America . In the intravenous and peritoneal models of infection , NO production was associated with efficient fungal clearance but also with non-organized granulomatous lesions . Because paracoccidioidomycosis is a pulmonary infection , we aimed to characterize the role of NO in a pulmonary model of infection . C57Bl/6 wild type ( WT ) and iNOS−/− mice were i . t . infected with 1×106 Paracoccidioides brasiliensis yeasts and studied at several post-infection periods . Unexpectedly , at week 2 of infection , iNOS−/− mice showed decreased pulmonary fungal burdens associated with an M2-like macrophage profile , which expressed high levels of TGF-β impaired ability of ingesting fungal cells . This early decreased fungal loads were concomitant with increased DTH reactions , enhanced TNF-α synthesis and intense migration of activated macrophages , CD4+ and CD8+ T cells into the lungs . By week 10 , iNOS−/− mice showed increased fungal burdens circumscribed , however , by compact granulomas containing elevated numbers of activated CD4+ T cells . Importantly , the enhanced immunological reactivity of iNOS−/− mice resulted in decreased mortality rates . In both mouse strains , depletion of TNF-α led to non-organized lesions and excessive influx of inflammatory cells into the lungs , but only the iNOS−/− mice showed increased mortality rates . In addition , depletion of CD8+ cells abolished the increased migration of inflammatory cells and decreased the number of TNF-α and IFN-γ CD4+ and CD8+ T cells into the lungs of iNOS−/− mice . Our study demonstrated that NO plays a deleterious role in pulmonary paracoccidioidomycosis due to its suppressive action on TNF-α production , T cell immunity and organization of lesions resulting in precocious mortality of mice . It was also revealed that uncontrolled fungal growth can be overcome by an efficient immune response .
Phagocytes are important effector cells of innate and adaptative immunity and use several mediators and mechanisms to control pathogen growth . The production of nitric oxide ( NO ) has been shown to be an important microbicidal mechanism of macrophages in the protective immune responses against different pathogens [1] , [2] , [3] , [4] . NO is generated from the amino acid L-arginine by the catalytic action of the inducible isoform of NO synthase ( iNOS or NOS2 ) [5] . The synergistic interaction of NO with hydrogen peroxide ( H2O2 ) or superoxide ( O2− ) anion can generate extremely potent oxidizing compounds resulting in cell damage and microbicidal activity [6] , [7] . Besides its action on pathogens viability , there are also evidences that NO has an inhibitory effect in the innate and adaptive immunity of hosts . For example , it reduces the antigen-presenting ability of pulmonary dendritic cells , inhibits MHC class II antigen expression , controls the production of cytokines and expression of costimulatory and adhesion molecules [8] , [9] . The iNOS gene expression is regulated by an ever-increasing number of agonists , especially proinflammatory cytokines such as IFN-γ and TNF-α and bacterial products such as lipopolysaccharides from Gram-negative bacteria [10] , [11] . On the other hand , type 2 cytokines , especially IL-4 , IL-10 and TGF-β , were shown to inhibit NO production [12] , [13] , [14] . Similarly with other microorganisms , the production of NO has been associated with the protective immunity against several fungal pathogens [15] , [16] . Recent evidences , however , suggested that certain fungal species such as Cryptococcus neoformans and Aspergillus fumigatus developed ingenious mechanisms to evade nitric oxide-dependent death [17] , [18] . Paracoccidioidomycosis ( PCM ) , a fungal disease caused by the inhalation of P . brasiliensis spores , presents a wide spectrum of immunopathological manifestations [19] . Patients with benign PCM usually develop adequate cellular immune responses and their antigen-stimulated leukocytes preferentially secrete type 1 cytokines; in contrast , patients with the severe form of the disease show impaired cell mediated immunity and type 2-skewed immune response [20] , [21] , [22] . Recent investigations , however , indicate that other regulatory mechanisms , not involving Th1/Th2 cells , play an important role in the immunopathogenesis of PCM [21] . Although the mechanisms involved in resistance to P . brasiliensis infection are not completely understood , it appears that alveolar macrophages have a fundamental role , acting as the first line of host defense . The enhanced fungicidal ability of cytokine-activated macrophages was shown to be mainly mediated by NO [23] , [2] . Despite this protective activity , in some studies NO production was associated with suppression of lymphoproliferation and MHC class II expression [24] , [25] . Interestingly , in P . brasiliensis infection we could detect an inverse correlation between TNF-α synthesis and NO production . Peritoneal and alveolar macrophages from resistant A/J mice in vitro infected with P . brasiliensis yeasts secreted high TNF-α levels , low NO amounts and displayed low fungicidal ability associated with enhanced TGF-β secretion . In contrast , macrophages from susceptible B10 . A mice secreted high NO levels , presented efficient fungal killing but produced low levels of TNF-α [26] , [27] . NO deficiency was also associated with organized granulomas of i . p . infected mice [28] , while exacerbated inflammatory reactions and cytokines production was described in i . v . infected mice [29] . Because P . brasiliensis infection is acquired by the respiratory route and the role of NO was never investigated in the pulmonary model of PCM , we aimed to further understand the immunoregulatory function of this mediator using i . t . infected iNOS-deficient ( iNOS−/− ) and normal ( WT ) C57BL/6 mice . We could characterize the temporal effects of NO synthesis in the control of fungal growth . At week 2 of infection , absence of NO results in lower fungal loads but at week 10 , increased numbers of yeasts were detected in the lungs of iNOS−/− mice . Unexpectedly , the deficient mouse strain showed increased survival times and this behavior was associated with high levels of TNF-α production increased and persistent delayed type hypersensitivity reactions and enhanced migration of activated T cells and macrophages into the lungs of infected mice . In addition , the increased fungal loads lately developed by iNOS-deficient mice appeared to be contained by better organized granulomatous lesions . Furthermore , in vivo depletion experiments showed that the protective effect of iNOS deficiency was mainly mediated TNF-α and the expansion of IFN-γ and TNF-α CD4+ and CD8+ T cells .
Animal experiments were performed in strict accordance with the Brazilian Federal Law 11 , 794 establishing procedures for the scientific use of animals , and the State Law establishing the Animal Protection Code of the State of São Paulo . All efforts were made to minimize suffering , and all animal procedures were approved by the Ethics Committee on Animal Experiments of the Institute of Biomedical Sciences of University of São Paulo ( Proc . 76/04/CEEA ) . Breeding pairs of homozygous iNOS-deficient ( iNOS−/− ) and wild type ( WT ) control C57BL/6 mice ( intermediate susceptibility to P . brasiliensis ) were bred at the University of São Paulo animal facilities under specific-pathogen-free ( SPF ) conditions in closed-top cages . Clean food and water were given ad libitum . Mice were 8 to 11 weeks of age at the time of infection , and procedures involving animals and their care were approved by the Ethics Committee on Animal Experiments of our Institution . The P . brasiliensis 18 isolate , which is highly virulent , was used throughout the study . To ensure the maintenance of its virulence , the isolate was used after three serial animal passages [30] . P . brasiliensis 18 yeast cells were then maintained by weekly subcultivation in semisolid Fava Netto culture medium [31] at 35°C and used on the seventh day of culture . The fungal cells were washed in phosphate-buffered saline ( PBS; pH 7 . 2 ) , counted in a hemocytometer and the concentration was adjusted to 20×106 fungal cells ml−1 . The viability of fungal suspensions , determined by Janus Green B vital dye ( Merck , Darmstadt , Germany ) [32] , was always higher than 80% . Mice were anesthetized and submitted to i . t . P . brasiliensis infection as previously described [33] . Briefly after intraperitoneal anesthesia , the animals were infected with 1×106 P . brasiliensis 18 yeast cells , contained in 50 µL of PBS , by surgical i . t . inoculation , which allowed dispensing of the fungal cells directly into the lungs . The skins of the animals were then sutured , and the mice were allowed to recover under a heat lamp . Mice were studied at several time points after infection . The number of viable microorganisms in infected organs ( lung , liver and spleen ) from experimental and control mice were determined by counting the number of CFU . Animals ( n = 6–8 ) from each group were sacrificed , and the enumeration of viable organisms was done as previously described [34] . Briefly , aliquots ( 100 µL ) of the cellular suspensions and serial dilutions were plated on brain heart infusion agar ( Difco , Detroit , USA ) supplemented with 4% ( vol/vol ) horse serum ( Instituto Butantan , São Paulo , Brazil ) and 5% P . brasiliensis 192 culture filtrate , the latter constituting a source of growth-promoting factor . The plates were incubated at 35°C , and colonies were counted daily until no increase in counts was observed . The number ( log10 ) of viable P . brasiliensis colonies per gram of tissue were expressed as means ± standard errors ( SEs ) . Mice ( n = 6–8 ) were infected i . t . with P . brasiliensis , their right lung and liver of anti-TNF-α and IgG treated mice , were removed aseptically and individually disrupted in 5 . 0 mL of PBS . Supernatants were separated from cell debris by centrifugation at 2 , 000× g for 15 min , passed through 0 . 22 µm pore-size filters ( Millipore , Bedford , Mass , USA ) , and stored at −70°C . The levels of IL-2 , IL-12 , IFN-γ , TNF-α , IL-4 , IL-5 , IL-10 and TGF-β were measured by capture enzyme-linked immunosorbent assay ( ELISA ) with antibodies pairs purchased from Pharmingen ( Pharmingen , San Diego , CA , USA ) . The ELISA procedure was performed according to the manufacture's protocol . The concentrations of cytokines were determined with reference to a standard curve for several twofold dilutions of murine recombinant cytokines . As an addition control , lung homogenates were added to recombinant cytokines used to obtain standard curves; no interference was detected , indicating the absence of inhibitory substances ( e . g . , soluble cytokine receptors ) . Two and ten weeks after i . t . infection , lungs of mice were lavaged by repeated injections of 0 . 5 ml of sterile PBS ( final volume 2 . 0 ml ) after cannulation of the trachea with polyethylene tubing which was attached on a tuberculin syringe . An aliquot of the recovered bronchoalveolar lavage fluid ( BALF ) was assayed by CFU to determine the presence of viable yeasts . Then , the remaining BALFs obtained from individual mice were spun at 1200 rpm , the supernatants removed and alveolar macrophages cultivated to characterize the presence of viable fungal cells . Cell pellets were resuspended in RPMI containing 10% fetal calf serum , 2 mM L-glutamine , 100 U/ml penicillin and 100 µg/ml streptomycin , adjusted at 4×105/ml of culture medium and 0 . 5 ml dispensed in 24-well tissue culture plates for a 2 h adhesion step . Non-adherent cells were discarded and some cultures treated with 0 . 5 ml of culture medium supplemented with 100 U/ml of IFN-γ ( Pharmingen , San Diego , CA , USA ) and cultivated for 48 h . Plates were then centrifuged , supernatants stored , cells disrupted by five washes with 0 . 5 ml of distilled water and suspensions collected in individual tubes . Pellets were resuspended in culture medium , and aliquots ( 100 µL ) and their serial dilution were assayed for the presence of viable yeasts . In addition , NO and H2O2 levels were determined in the supernatants of alveolar macrophages cultures . Peritoneal macrophages from WT and iNOS −/− C57BL/6 mice were induced by i . p . injection of brewer thioglycollate medium ( Difco , Detroit , MI , USA ) . Macrophages were isolated by adherence ( 2 h at 37°C in 5% CO2 ) to plastic-bottom tissue-culture plates ( 1×106 cells/well in 24 well plates ) , cultivated overnight with fresh complete medium in the presence or absence of recombinant IFN-γ ( 40 ng/ml , BD-Pharmingen San Diego , CA , USA ) , TNF-α ( 20 ng/ml , BD-Pharmingen San Diego , CA , USA ) or 1-methyl-DL-tryptophan ( 1 MT , 1 mM in culture medium , Sigma Aldrich , St . Louis , MO , USA ) , a specific inhibitor of 2 , 3 indoleamine dioxygenase . Macrophage cultures were infected with P . brasiliensis yeasts in a macrophage∶yeast ratio of 12 . 5∶1 . After 48 h of culture , plates were centrifuged and supernatants removed . The wells were washed with distilled water to lyse macrophages , the suspensions collected and assayed for the presence of viable yeasts . All assays were done with five wells per condition in over three independent experiments . For phagocytic assays , macrophages from WT and iNOS −/− mice were infected with heat-inactivated , FITC labeled P . brasiliensis yeasts at a macrophage∶yeast ratio of 1∶1 for 2 h at 37°C in 5% CO2 to allow fungi adhesion and ingestion as previously described [35] . Some macrophage cultures were treated with IFN-γ ( 40 ng/ml , BD-Pharmingen ) , or TNF-α ( 20 ng/ml BD-Pharmingen ) overnight , before infection . Macrophages were gently washed twice with PBS and cells detached from plastic with fresh cold medium and a rubber cell scraper on ice . The cells were transferred to tubes , centrifuged ( 400×g . 10 min . , 4°C ) , and the pellets were labeled with anti-F4/80 ( APC ) antibodies ( eBioscience , San Diego , CA , USA ) . The cells were washed twice in PBS , the pellets were suspended in 200 µL of PBS 1% FCS and were immediately read on FACScalibur ( Becton Dickinson , Franklin Lakes , NJ , USA ) and data analyzed using the FlowJo software program ( Tree Star , Ashland , OR , USA ) . For the distinction between internalized and surface-bound yeasts ( FITC- labeled P . brasiliensis particles ) , trypan blue ( TB , 250 µg/mL , Sigma Aldrich , St . Louis , MO , USA ) was used for quenching the green surface-bound fluorescence on macrophages . TB quenching technique was performed as described by Busetto et al . [36] with minor modifications . Phagocytic assays were performed as above described and adherent/ingested cells measured using the FL1 and FL4 channels of a FACscalibur cytometer . Cell suspensions were then treated in an ice bath with 0 . 1 ml of a TB solution prepared in 0 . 1 M citrate buffer , pH 4 . 0 , lowering samples pH to nearly 4 . 0 , thereby optimizing the TB quenching effect . After 1 min of incubation in ice bath , the samples were again analyzed . APC-labeled macrophages were gated , and FL1 and FL3 channels used to discriminate ingested ( green fluorescent , FL1 ) from adherent ( red fluorescent , FL3 ) yeasts . Total RNA was extracted from cultures of normal or P . brasiliensis-infected macrophages using the TRIzol reagent ( Invitrogen , Carlsbad , CA , USA ) according to the manufacturer's instructions . The RNA concentrations were determined by spectrophotometer readings at an absorbance of 260 nm . First-strand cDNAs were synthesized from 2 µg RNA using the High Capacity RNA-to-cDNA kit ( Applied Biosystems , Foster City , CA , USA ) according to the manufacturer's instructions . Real-time polymerase chain reaction ( RT-PCR ) was performed using the TaqMan real-time PCR assay ( Applied Biosystems ) for the following molecules: ARG1 ( Mm00475988_m1 ) , NOS2 ( Mm00440502_m1 ) , IL-12 p40 ( Mm00434174_m1 ) , TGF-β ( Mm00441727_m1 ) , TNF-α ( Mm99999068_m1 ) , IDO ( Mm00492586_m1 ) . Cycling conditions were as follows: 10 min at 95°C , followed by 45 cycles of 20 s at 95°C , 20 s at 58°C , and 20 s at 72°C . Analysis was performed with the ABI PRISM 7000 sequence detection system ( Applied Biosystems ) . GAPDH was used as an internal control . All values were normalized to GAPDH , and the relative gene expression was calculated using the Pfaffl method [37] Nitric oxide production was quantified by the accumulation of nitrite in the supernatants by a standard Griess reaction . Briefly , 50 µL of supernatants was removed from 24-well plates and incubated with an equal volume of Griess reagent ( 1% sulfanilamide , 0 , 1% naphthylene diamine dihydrochloride , 2 , 5% H3PO4 ) at room temperature for 10 min . The absorbance at 550 nm was determined with a microplate reader . The conversion of absorbance to micromolar NO was deduced from a standard curve by using a known concentration of NaNO2 diluted in RPMI medium [38] . All determinations were performed in triplicates and expressed as micromolar NO . DTH reactions were evaluated employing a footpad test previously described [34] . Briefly , mice were inoculated with 25 µl of a soluble P . brasiliensis antigen [31] and footpad thickness was measured with a caliper ( Mitutoyo , Tokyo , Japan ) immediately before and 24 h after antigen inoculation . The increase in thickness was calculated and expressed in millimeters . Non-infected mice submitted to the footpad test were used as controls . Lungs from each mouse were excised , washed in PBS , minced , and digested enzymatically for 1 hour in 15 mL/lung of digestion buffer [RPMI , 5% fetal calf serum , 1 mg/mL collagenase and 30 µg/mL DNase ( Sigma Aldrich , USA ) ] . After erythrocyte lysis using NH4Cl buffer , cells were washed , resuspended in complete media , and centrifuged for 30 minutes at 2 , 000× g in presence of 20% Percoll ( Sigma ) to separate leukocytes from cell debris and epithelial cells . Total lung leukocyte numbers were assessed in the presence of trypan blue using a hemocytometer; viability was always higher than 85% . The absolute number of a leukocyte subset was equal to the percentage of that cell subset multiplied by the total number of leukocytes recovered from the digested lung/100 . For surface staining alone , leukocytes were washed and resuspended at a concentration of 1×106 cells/mL in staining buffer ( PBS 1× , 2% serum calf bovine and 0 , 5% NaN3 ) . Fc receptors were blocked by the addition of unlabeled anti-CD16/32 ( Fc block; BD Pharmingen , San Diego , CA , USA ) . The leukocytes were then stained for 20 min at 4°C with the optimal dilution of each antibody . Anti-CD4 , CD8 , CD69 , CD25 , CD40 CD80 , CD86 , CD11b and CD11c FITC or PE-conjugated antibodies were from BD Pharmingen . Cells were washed twice with staining buffer resuspended in 100 µl , and an equal volume of 2% formalin was added to fix the cells . The stained cells were analyzed immediately on a FACScalibur equipment using the Cell-Quest software ( Becton & Dickinson , Sparks , MD , USA ) gating on macrophages or lymphocytes as judged from forward and side light scatter . Ten thousands cells were counted and the data expressed as the percentage or the absolute number of positive cells which was calculated trough the percentage obtained by FACS and the number of cells determined in Neubauer chambers . The intracellular detection of FoxP3 , the X-linked forkhead/winged helix transcription factor , in leukocytes obtained from the lung lesions was performed in fixed and permeabilized cells using Cytofix/Cytoperm ( BD Biosciences , San Diego , CA , USA ) . Initially , the cells were labeled with antibodies for cell surface molecules such as FITC-conjugated anti-CD4 and PE-conjugated anti-CD25 . Next , the cells were fixed , permeabilized , and stained with Cy-conjugated anti-FoxP3 , for 1 , 5 h at 4°C . Cells were then washed twice with staining buffer , resuspended in 100 µl , and an equal volume of 2% formalin was added to fix the cells . A minimum of 20 , 000 events was acquired on FACScalibur flow cytometer ( BD Pharmingen ) using the Cell-Quest software ( BD Pharmingen ) . The graphs represent the number of Foxp3+ cells in the gate of CD4+ CD25+ T cells . For intracellular cytokine ( IL-4 , IFN-γ and TNF-α ) staining , cells were stimulated for 6 h in complete medium in the presence of 50 ng/ml phorbol 12-myristate 13-acetate , 500 ng/ml ionomycin ( both from Sigma-Aldrich ) and monensin ( 3 mM , eBioscience ) . After surface staining for CD4 ( Pacific Blue anti-CD4 ) and CD8 ( Alexa Fluor 488 anti CD8 ) , cells were fixed , permeabilized , and stained by PerCP- Cy5 . 5 anti-IFN-γ , Pe-Cy7 anti-IL-4 and PE anti-TNF-α antibodies ( eBioscience , San Diego , CA , USA ) . The cell surface expression of leukocyte markers as well as intracellular expression of IL-4 , IFN-γ and TNF-α in lung infiltrating leucocytes , were analyzed in a FACScalibur flow cytometer ( BD Pharmingen ) using the FlowJo software ( Tree Star , Ashland , OR , USA ) . In selected experiments , lungs from infected iNOS−/− and WT mice were removed and digested enzymatically as above described . DCs were purified by magnetic cell sorting with microbeads ( Miltenyi , Bergisch Gladbach , Germany ) conjugated to hamster anti-mouse CD11c monoclonal antibodies . Positively selected DCs contained more than 90% CD11c+ cells . Cell-surface markers of pulmonary DCs were characterized by flow cytometry using monoclonal antibodies anti-CD11c PE-Cy7 , anti-CD11b PerCP-Cy5 . 5 , anti-CD8a-Alexa Fluor 488 , and anti- B220-PE . P . brasiliensis-infected ( iNOS−/− and WT ) mice were given i . p . injections of 0 . 25 mg/0 . 5 mL of an anti-TNF-α MAb ( MP6 XT 22 . 11 ) , a rat IgG1 monoclonal antibody , against murine TNF-α [39] , 4 h before the infection , and at days 6 and 12 postinfection . Normal rat IgG was given i . p . to mice as a control for antibody administration . Treated and untreated mice ( n = 6–7 ) were studied at weeks 2 and 8 weeks after infection and mortality rates were also evaluated . H-35 ( rat IgG1 anti-CD8α ) hybridoma was grown in BALB/c nu/nu mice . Monoclonal antibodies ( MAbs ) were purified from ascites as previously described [40] and assayed for purity by sodium dodecyl sulfate-polyacrylamide gel electrophoresis . Groups ( n = 6–7 ) of WT and iNOS−/− mice were given 200 µg of anti-CD8α mAbs or normal rat IgG ( controls ) by the i . p . route , 48 and 24 h before infection and 150 µg of the mAbs or rat IgG at days 6 and 12 postinfection . The severity of infection and lung infiltrating leukocytes were characterized at week 2 after infection . Groups of iNOS−/− mice and their normal counterparts were killed at the second , eighth and tenth week postinfection . Lungs were collected , fixed in 10% formalin and embedded in paraffin . Histolopathogical studies were also performed with anti-TNF-treated and untreated mice at week 8 after infection . Five-micrometer sections were stained by the hematoxilin-eosin ( H&E ) for an analysis of the lesions . Pathological changes were analyzed based on the size , morphology and cell composition of granulomatous lesions , presence of fungi and intensity of the inflammatory infiltrates . Mortality studies were done with groups ( n = 6–8 ) of iNOS−/− and WT control mice inoculated i . t . with 1×106 yeast cells or PBS . Deaths were registered daily for a 350-day period , and the median survival time postinfection was calculated . Mortality of untreated and anti-TNF-α -treated mice ( n = 6–7 ) of both mouse strains were also studied . Experiments were repeated three times . All values are means ± SEM , unless otherwise indicated . Depending on the number of experimental groups , data were analyzed by Student's t test or two-way analysis of variance and the Bonferroni posttests to compare groups . Differences between survival times were determined with the LogRank test using GraphPad Prism software ( GraphPad Software , San Diego , CA , USA ) . P value<0 . 05 was considered significant .
The evolution of the disease of i . t . infected iNOS−/− mice and their WT controls was monitored by CFU counts in the lung and liver at different postinfection periods ( 48 h , 2 , 6 and 10 weeks ) ( Figure 1A and 1B ) . At the first 48 h of infection , an equivalent number of viable yeasts cells was recovered from lungs of both mouse strains . Interestingly , at the 2nd week , iNOS deficiency resulted in decreased CFU counts in the lungs ( 4 . 55±0 . 86 log10 CFU/g of tissue ) , compared with normal controls ( 5 . 23±0 . 29 log10 CFU/g of tissue ) . No differences were noted in the dissemination to liver ( Figure 1B ) and spleen ( data not shown ) . Although at the 6th week both groups of mice showed equivalent pulmonary fungal loads , at week 10 postinfection the iNOS−/− mice presented augmented fungal burden in the lungs ( 5 . 99±0 . 79 log10 CFU/g of tissue ) relative to WT mice ( 4 . 64±0 . 80 log10 CFU/g of tissue ) . Again , no differences in fungal dissemination were observed . These data clearly showed the opposite temporal effect of NO: early in the infection , the absence of nitric oxide synthesis resulted in a protective effect , but at the chronic phase led to a more severe disease ( Figure 1A ) . Because the differences in pulmonary fungal burdens of iNOS−/− and WT mice were detected at the 2nd and the 10th weeks postinfection , these periods were chosen to next determine the levels of type 1 ( IFN-γ , TNF-α , IL-2 and IL-12 ) and type 2 ( IL-4 , IL-5 and IL-10 ) cytokines . As shown in Figure 1C , type 1 and type 2 cytokines were present in the lungs of both studied groups , but at the early phase of infection only TNF-α appeared in significantly higher levels in NO-deficient mice . Later in the infection , however , no differences in pulmonary cytokines were detected ( Figure 1D ) . These data suggested that TNF-α could be involved in the early immunoprotection conferred by NO deficiency . We have also evaluated the presence and the activation profile of leukocytes in the lungs from both mouse strains . A higher number of mononuclear phagocytes expressing activation molecules was detected in iNOS−/− mice when compared with WT mice . As shown in Figure 2A , at the 2nd week , the number of double positive CD11b+CD86+ and CD11b+CD40+ cells was higher in iNOS−/− mice . Further phenotypic characterization of macrophages ( CD11bhighCD80high ) and dendritic cells ( CD11chighCD86high ) demonstrated the increased presence of macrophages in the deficient mouse strain , while the number of dendritic cells was equivalent in both studied groups ( Figure 2A ) . By week 10 , no significant differences in the number of pulmonary macrophages were detected ( Figure 2B ) . In addition , to determine the lymphocyte influx and the activation profile of CD4+ and CD8+ T cells in the lungs of P . brasiliensis infected mice , we evaluated the expression of CD69 and CD25 by freshly isolated T cells . The marker CD69 is a very early activation antigen [41] as well as CD25 , the α-chain of the interleukin-2 receptor [42] , which is rapidly upregulated on activated T cells . Compared with the control group , at the 2nd week , NO-deficient mice presented an increased recruitment of activated CD4+CD69+ , CD4+CD25+ and CD8+CD69+ T lymphocytes to the lungs ( Figure 2C ) . Although in lower intensity , at the chronic phase , the recruitment/expansion of CD4+CD69+ and CD4+CD25+ T cells to the lungs of iNOS−/− mice remained higher than in WT mice . The same was not verified for CD8+CD69+ T cells that appeared in equivalent numbers in both mice strains ( Figure 2D ) . Because Treg cells control the expansion of effector T cells , and the number and function of these cells were shown to be influenced by NO production [43] , [44] we characterized the presence of CD4+CD25+FoxP3+ T cells in the pulmonary cell infiltrates ( Figure 2C and 2D ) . Although at the week 2 , no differences in the numbers of Treg cells were seen , by week 10 increased numbers of CD4+CD25+FoxP3+ cells were detected in iNOS−/− mice . This finding paralleled the increased pulmonary CFU counts , the diminished number of T cells and the impaired macrophage activation detected in the lungs of iNOS−/− mice at this late period of the infection . After characterizing the main features of the infection , we aimed to clarify the mechanisms involved in the early immunoprotection conferred by iNOS deficiency . Thus , the behavior of alveolar macrophages was assessed at two opposed periods of infection . In agreement with lung CFU data , at the 2nd week a lower number of yeasts was recovered from the bronchoalveolar lavage fluid of iNOS−/− mice ( Figure 3A ) . The microbicidal activity of alveolar macrophages was further determined after 48 h cultivation in the presence or absence of IFN-γ . Again , lower CFU counts were recovered from deficient macrophages ( Figure 3B ) . As expected , only WT macrophages showed decreased fungal counts associated with increased NO production after IFN-γ treatment ( Figure 3B and 3C , respectively ) . At week 10 , iNOS-deficient macrophages showed higher fungal loads than WT cells , which increased their fungicidal ability after IFN-γ treatment ( Figure 3D and 3E ) . Studies with alveolar macrophages raised two important questions . Were the decreased fungal loads of iNOS−/− macrophages due to their increased fungicidal activity or decreased phagocytic ability ? Why are iNOS−/− macrophages refractory to IFN-γ activation ? Therefore , inflammatory peritoneal macrophages were used to better understand the behavior of iNOS-deficient phagocytes . Peritoneal macrophages were obtained , activated or not by IFN-γ , TNF-α or both cytokines and infected by P . brasiliensis . Some activated and non-activated cells were also treated with 1MT , a specific inhibitor o 2 , 3 indoleamine dioxygenase , an enzyme that catalyzes the degradation of tryptophan along the kynurenine pathway . In fungal infections , this enzyme was shown to exert an efficient fungicidal activity but also an important suppressive effect on the immune response [45] . Recapitulating the results obtained with alveolar macrophages , lower numbers of viable yeasts were recovered from iNOS−/− macrophages ( Fig . 4A ) . Only WT cells increased their fungicidal ability when activated by IFN-γ , although TNF-α was not able to modify the fungicidal activity of cells from both mouse strains . Similar result was observed with 1MT-treated macrophages . The IDO inhibitor was not able to modify the behavior of activated and non-activated macrophages from WT and iNOS−/− mice ( Fig . 4A ) . We have then used FITC-labeled P . brasiliensis yeasts to discriminate adherent and ingested cells . As shown in Fig . 4B , iNOS−/− macrophages have reduced capacity to adhere/ingest fungal cell . When the green fluorescence of adhered yeasts was quenched by trypan blue treatment , a decreased ingestion activity was observed with WT and iNOS-deficient macrophages , although more evident with the latter cells . Thus , iNOS−/− macrophages have a reduced ability to adhere and ingest fungal cells , and this behavior appears to explain the reduced CFU counts detected in vitro and in vivo with iNOS−/− cells . To better characterize the differentiation of iNOS−/− and WT macrophages , the expression of iNOS , ARG1 , TGF-β , IL-12 , TNF-α and IDO mRNA was measured in uninfected and P . brasiliensis infected cells ( Fig . 5 ) . iNOS-deficient macrophages showed reduced expression of IL-12 associated with increased production of ARG1 and TGF-β mRNA , demonstrating a prevalent anti-inflammatory behavior and some characteristics of M2 , healing , or alternatively activated macrophages . An opposite result was observed with WT macrophages which expressed high levels of iNOS and IL-12 mRNA . Interestingly , no differences in IDO expression were detected between WT and iNOS−/− cells , but the latter showed increased expression of TNF-α . The delayed type hypersensitivity ( DTH ) reactions developed by iNOS−/− and WT mice were evaluated at weeks 2 and 10 after infection using a soluble P . brasiliensis antigen . As depicted in Figure 6A , both mouse strains developed increased footpad reactions , but these were significantly higher in iNOS−/− than in WT mice . To better typify the effect of iNOS deficiency in the severity of pulmonary PCM , histopathological analysis of lung sections from iNOS−/− and WT mice at both periods of infection was also performed . By week 2 , iNOS−/− and WT mice showed diffuse inflammatory reactions composed of monocytes and lymphocytes surrounding interlobular spaces localized around the bronchi , bronchioles and blood vessels . However , an increased presence of inflammatory cells accompanied the lower CFU counts observed in the lungs of iNOS−/− mice ( data not shown ) . Surprisingly , at week 10 remarkable histopathological differences were detected; iNOS−/− mice presented a large number of well-organized granulomas ( Figure 6B , lower panel ) containing an elevated number of yeast cells surrounded by epithelioid and multinuclear giant cells , and a well-defined lymphocytic mantle . Plasma cells and eosinophils were scarce . The fungi , detected in great numbers , were large and have multiple buds . Compared with iNOS−/− mice , WT mice presented more extensive , non-organized lesions , and decreased fungal loads irregularly distributed in the lung parenchyma ( Figure 6B , upper panel ) . Importantly , despite the higher number of yeasts recovered late in the infection , iNOS−/− mice showed decreased mortality rates ( Figure 6C ) . Thus , the increased T cell immunity as evidenced by increased DTH reactions , and the more organized lesions ( possibly mediated by the increased TNF-α secretion ) appear to have compensated the poor control of fungal multiplication resulting from iNOS deficiency Supported by the remarkable increase of TNF-α levels observed early in the infection in the lungs of iNOS−/− mice , and to further understand the mechanisms of immunoprotection used by this mouse strain to control P . brasiliensis infection , iNOS−/− and WT groups were in vivo depleted of TNF-α and the severity of infection analyzed by CFU counts and pulmonary inflammation . At week 1 after infection , a significant difference in fungal burdens were observed between IgG-treated WT and iNOS−/− mice . The anti-TNF-α treatment , however , abolished this difference ( Figure 7A ) . Anti-TNF-α treatment was shown to be effective , since reduced levels of this cytokine were detected in the treated groups of both mouse strains ( Figure 7B ) . In addition , in both mouse strains at weeks 2 and 8 after infection TNF-α depletion did not result in significant increases in pulmonary fungal burdens ( Figure 7C , D ) . Yet , no significant differences were noticed in the dissemination to liver and spleen . Cytokines measurements in lung homogenates at week 2 after infection demonstrated that , although anti-TNF-α treatment remained until day 12 postinfection , TNF-α rapidly reached the pre-treatment levels in the liver and lungs of both mouse strains ( Figure 7E and 7F ) . On the other hand , a higher concentration of TGF-β was detected in the lungs , while decreased levels of hepatic IFN-γ were seen in anti-TNF-α treated iNOS−/− mice ( Figure 7E , F ) . Unexpectedly , when lung infiltrating leukocytes were characterized at week 2 postinfection , an impressive increase in the numbers of lymphocytes ( Figure 8A ) and mononuclear phagocytes ( Figure 8B ) were seen in the lungs of anti-TNF-α-treated mice . Thus , a higher influx of activated CD4+CD69+ , CD8+CD69+ T cells into the lungs of TNF-depleted WT and iNOS−/− mice was observed , although CD4+CD25+ T appeared in higher numbers only in the former strain ( Figure 8A ) . In addition , compared with IgG-treated controls , iNOS−/− and WT-depleted groups showed elevated numbers of CD11b+CD80+ and CD11b+CD40+ mononuclear phagocytes ( Figure 8B ) . To further characterize the role of TNF-α in our model , TNF-α-depleted and IgG-treated controls were studied regarding mortality and histopathology of lungs . As depicted in Figure 8C , the effect of TNF-α neutralization was much more prominent in the iNOS-deficient strain . Indeed , 100% of TNF-α depleted iNOS−/− mice died within 70 days of infection while only 25% of WT mice died in the same period . This result demonstrated that , at least partially , the relative protection of NO-deficient mice was due to the enhanced TNF-α production . In vivo neutralization of TNF-α did not appear to cause clear alterations in the already poor-organized and confluent lesions of WT mice ( Figure 8D , left panels ) . As described for untreated-infected mice , IgG-treated iNOS−/− mice presented better-defined granulomas , and a high influx of inflammatory cells , which appears to restrain fungal spreading ( Figure 8D , upper right panel ) . In contrast , anti-TNF-treated iNOS−/− mice lose this organized pattern of lesions and non-organized , confluent inflammatory reactions containing fungal cells were scattered through the pulmonary tissue ( Figure 8D , lower right panel ) . The lower pulmonary fungal growth detected at week 2 postinfection of iNOS−/− mice paralleled the enhanced presence of inflammatory cells ( Figures 1A and 2C ) . Importantly , among T cells , only the CD8+ T cell subset lost their activation profile at week 10 of infection when iNOS−/− mice were unable to control of fungal growth . In order to study the involvement of CD8+ T cells in the early immunoprotection of NO-deficient mice , CD8α+ cells were in vivo depleted by monoclonal antibodies , and the severity of infection characterized at week 2 postinfection . As depicted in Figure 9A , anti-CD8 treatment abolished the differences in CFU counts previously observed in the lungs of mice . Moreover , this treatment also abolished the higher presence of activated T cells and macrophages observed in the lungs of iNOS-deficient mice at this early period of infection . Compared with IgG-treated controls , CD8-depleted iNOS−/− mice presented significantly diminished influx of CD4+CD69+ and CD8+CD69+ T cells to the lungs ( Figure 9B ) . The same occurred with CD11b+CD80+ , CD11b+CD40+ and CD11b+CD86+ mononuclear phagocytes ( Figure 9C ) . Subsequently , we asked whether this difference in cell influx caused by anti-CD8α treatment was due to the depletion of the CD8α+ subset of dendritic cells . Thus , IgG-treated and anti-CD8-depleted WT and iNOS−/− groups were i . t . infected and at the second week post-infection the lungs were removed , digested and pulmonary dendritic cells obtained with anti-CD11c magnetic beads . The phenotype of these cells was then assessed by flow cytometry . The Figure 9D demonstrates that no differences in the numbers of conventional ( CD11chighCD11blow ) , lymphoid ( CD11c+CD8α+ ) and plasmacytoid ( CD11c+B220+ ) DCs were detected between anti-CD8-treated and untreated mice of both mouse strains . In summary , this result suggested that the effect of anti-CD8 treatment was mainly mediated by CD8α+ T lymphocytes and not lymphoid DCs . Next , we characterized the phenotype of IFN-γ , TNF-α and IL-4-secreting lymphocytes in the lungs of anti-CD8-treated and IgG-treated iNOS−/− and WT mice by intracellular cytokine staining . Comparing IgG-treated controls , enhanced numbers of TNF-α + CD4+ and CD8+ T cells were seen in the lungs of iNOS−/− than in the WT group ( Figure 10 ) . Anti-CD8 treatment led to increased numbers of TNF-α+ , IFN-γ+ and IL-4+ CD4+ T cells in WT mice . However , reduced numbers of TNF-α+ and IFN-γ+ CD8+ T cells appeared in the lungs of this strain . Importantly , in iNOS−/− mice the depletion of CD8+ cells resulted in a remarkable reduction of TNF-α+ and IFN-γ+ CD8+ T cells , besides a decreased presence of IFN-γ+ CD4+ T cells ( Figure 10 ) . Therefore , at the 2nd week of P . brasiliensis infection , the reduction of pro-inflammatory CD8+ T cells observed in WT mice appeared to be compensated by the increased presence of TNF-α- and IFN-γ-secreting CD4+ T cells whereas in iNOS-deficient mice a prevalent reduction of pro-inflammatory CD4+ and CD8+ T cells was seen .
In the present work , we investigated the temporal significance of nitric oxide synthesis in the evolution of pulmonary paracoccidioidomycosis and the immunopathological mechanisms associated with iNOS deficiency . Early in infection , the protective effects of iNOS deficiency was associated with decreased fungal burdens , enhanced secretion of TNF-α augmented DTH reactions , and increased migration of activated T cells and macrophages to the lungs , which subsequently organize as well-compact granulomas . On the other hand , at later periods , increased fungal loads were concomitant with sustained T cell immunity allied with increased presence of regulatory T cells at the site of infection . Unexpectedly , the mortality rates of WT mice were higher than those of iNOS−/− mice . In vivo depletion of TNF-α and CD8+ T lymphocytes demonstrated a division of labor carried out by these two components in the immunoprotection developed by iNOS−/− mice . While upregulated TNF-α secretion avoided precocious mortality and organized pulmonary lesions , the increased expansion of CD8+ T cells controlled fungal growth and secretion of pro-inflammatory cytokines . Both , TNF-α and CD8+ T cells were involved in the enhanced recruitment of inflammatory cells to the lungs . This protective effect was persistent , but excessive inflammatory reactions were possibly controlled by the increased expansion of Treg cells at late stages of immunity developed by NOS−/− mice . Altogether , these mechanisms appear to confer sustained protection to iNOS−/− mice , which , despite the elevated fungal loads , presented increased survival time and better disease outcome . In our model , we could verify that iNOS deficiency seems to be compensated by the deviation of the immune response to a more pronounced Th1 pattern . The lower fungal loads were concomitant with the high levels of pulmonary TNF-α produced by P . brasiliensis infected NO-deficient mice at the 2nd week of infection . Interestingly , it has been reported that low concentrations of NO enhanced Th1 immunity by increasing the expression of IL-12 receptor in T cells although high NO concentrations are cytotoxic [44] , [46] . Interestingly , alternative mechanisms for the immunosuppressive activity of NO production have recently been described . It was demonstrated that NO suppresses NALP3 inflammasome activation by nitrosylation of NALP3 proteins resulting in decreased synthesis of mature IL-1 beta and IL-18 , impaired Th1 immunity and NK cell activation [47] . Considering this information , it is tempting to speculate that the enhanced Th1 immunity developed by iNOS-deficient mice could be associated with increased NALP3 inflammasome activity and enhanced IL-1/IL-18 production . Consistent with a prevalent Th1 pattern of immunity , by week 2 , increased levels of IFN-γ-regulated isotypes ( IgG2a and IgG3 ) , were produced by iNOS−/− mice ( data not shown ) . Thus , these data are consistent with the benign forms of murine and human PCM , which are associated with prevalent type-1 immunity [21] . Allied with type-1 cytokine production , an increased number of activated lymphocytes and macrophages was found in the lungs of iNOS-deficient mice suggesting the development of enhanced DTH reactions at the site of infection . iNOS-deficient alveolar macrophages showed fungal loads equivalent to those detected in lung tissue and did not present an increased fungal ability when activated by IFN-γ . Several investigations showed the fundamental role of NO in P . brasiliensis killing by IFN-γ activated macrophages [2] , [23] , although NO-independent mechanisms were reported to be associated with TNF-α-activated human and murine macrophages [48] , [49] . Our studies with inflammatory macrophages confirmed that iNOS−/− macrophages are refractory to IFN-γ and TNF-α activation . Furthermore , the reduced CFU numbers displayed by these cells could not be ascribed to increased IDO expression , because the CFU counts were not modified by 1MT treatment . Importantly , iNOS−/− macrophages were also shown to have a decreased ability to adhere to and ingest yeast cells , which possibly explains the low CFU numbers detected both in the in vitro and in vivo infections . Absence of iNOS expression was also associated with an M2-like behavior of macrophages . Indeed , these cells expressed high levels of arginase-1 and TGF-β and were associated with low levels of IL-12 . This anti-inflammatory behavior appears to explain why iNOS-deficient cells were not activated by IFN-γ and TNF-α and explains why the iNOS-deficient cells do not express elevated levels of IDO , an enzyme primarily induced by IFN-γ activation [45] . However , iNOS-deficient cells expressed high levels of TNF-α , which is potentially associated with an efficient activation of dendritic cells and increased migration of inflammatory cells to the site of infection . Interestingly , the behavior of iNOS−/− macrophages is similar to the behavior of macrophages and dendritic cells from mice ( strain A/J ) resistant to P . brasiliensis infection [27] , [35] . A/J cells are poorly activated by IFN-γ and IL-12 and show an impaired NO production and killing ability . Concomitant with elevated production of TGF-β , A/J macrophages produce high levels of TNF-α that contribute to the late but consistent cellular immunity and immunoprotection developed by this mouse strain [27] , [35] , [50] . Our data have also demonstrated that in pulmonary PCM , early NO production inhibits the activation and migration of CD4+ and CD8+ T cells to the site of infection . This finding was not previously described in murine PCM , but was reported in other experimental models where less severe infections , mainly due to increased IFN-γ production and CD4+ Th1-skewed immune responses , were observed in iNOS−/− mice [4] , [44] , [46] , [51] , [52] . The increased influx of T cells and macrophages was concomitant with elevated levels of pulmonary TNF-α , a proinflammatory cytokine that enhances the maturation of antigen presenting cells and induces increased expression of adhesion molecules on endothelial cells [53]–[55] . Thus , the enhanced secretion of TNF-α appeared to have amplified the afferent and the efferent phases of immunity , by increasing T cell sensitization and further migration to the site of P . brasiliensis infection . Despite the less severe infection at the 2nd week , iNOS−/− mice were not able to sustain this behavior , and elevated fungal burdens were seen in their lungs at later periods . Despite the yet significantly increased presence of activated CD4+ T cells in iNOS−/− mice , no differences in the number or activation of CD8+ T cells were detected . This decreased influx of effector CD8+ T cells was parallel to the decreased activation of macrophages , indicating , as we previously demonstrated , the important role of this T cell subset in the immunoprotection of pulmonary PCM [21] , [56] . Our studies appear to indicate that at an early phase of infection NO does not affect the expansion of CD4+CD25+Foxp3+ regulatory T cells . However , late in infection ( week 10 ) , iNOS deficiency supported the expansion of this regulatory T cell subset . These findings suggest that absence of NO production led to an early enhanced T cell immunity , but the excessive lung inflammation was lately avoided by increased expansion of Treg cells . The M2-like profile of inflammatory macrophages appears to have contributed to the expansion of Treg cells and the controlled tissue pathology observed in iNOS−/− mice . This Treg-associated mechanism controlling effector immunity was not previously described for the NO-induced immunosuppression in murine PCM , and can be added to the regulatory mechanisms mediated by unbalanced NO production . Moreover , our data on the presence of CD4+CD25+Foxp3+ Treg cells at the site of infection are in good agreement with ours [57]–[59] , and others [60] , [61] reports showing a late enhancement of Foxp3+ Treg cells associated with increased immunological responses and pathogen burden . Interestingly , the histopathological examination of lungs was consistent with the important role of TNF-α in the organization of granulomatous lesions [62] , [64] . Indeed , the high influx of inflammatory T cells and macrophages into the lungs of iNOS−/− mice , associated with the locally increased levels of TNF-α resulted in more organized lesions at week 10 , which appeared to have overridden the elevated fungal loads due to the lack of NO synthesis . Furthermore , this pattern of lesion organization appeared to be able to restrain fungal dissemination to distant organs , since despite the lately increased pulmonary fungal burdens , no differences were detected in the liver and spleen of iNOS−/− mice . The increased survival time of iNOS-deficient mice appears to underline the possible protective effect of well-organized lesions . It is known that arginine catabolism is mediated by two types of enzymes: the nitric oxide synthases convert arginine to citrulline and NO , while the arginases hydrolyze arginine to urea and ornithine . The latter component is necessary for the production of proline and is essential for the synthesis of collagen , which is the main component of the extracellular matrix ( ECM ) observed in granulomatous lesions [65] . As here shown , iNOS-deficient macrophages only express arginase-1 , which may have contributed to the expression of ECM components and the compact organization of granulomas observed in iNOS-deficient mice . It is also known that NO has a modulatory effect on the expression and activity of zinc-dependent metalloproteases ( MMPs ) , which decrease the deposition and accumulation of extracellular matrix ( ECM ) proteins [66] . Some studies have demonstrated that NO induces the activation of MMP2 and MMP9 , which are two enzymes that have an anti-fibrotic effects through the degradation of ECM proteins ( laminin , collagen , elastin , etc . ) and pro-cytokines involved in fibroblast activation and granuloma organization ( e . g . , pro-TNF-α and pro- TGF-β ) [66] , [67] , [68] . In the intraperitoneal model of PCM using WT and iNOS−/− mice [69] , NO production was associated with increased MMP9 activity and the loose organization of granulomas developed by WT mice . In the pulmonary model employed in the present study we suppose that the NO produced by the WT mice inhibited the production of TNF-α , which is required for granuloma organization , and of TGF-β , which is needed for tissue repair and Treg cell expansion . Consistent with this possibility , infected iNOS−/− macrophages expressed elevated levels of TGF-β ( Fig . 5 ) , which is a cytokine involved in fibroblast activation , enhanced synthesis of ECM , and Treg cell differentiation . The TGF-β-induced Treg cell differentiation apparently contributed to the controlled tissue pathology even in the presence of the augmented immune response of the iNOS−/− mice . The study by Nishikako et al . [69] revealed additional important information on the influence of NO in granuloma formation and disease outcome . At late stages of infection ( 120 days post-infection , not evaluated in our model ) , i . p . infected iNOS−/− mice presented decreased fungal loads in their well-organized granulomas , which demonstrated the efficiency of their inflammatory reactions . Furthermore , as observed here , iNOS−/− mice showed decreased mortality rates when compared with WT mice . In vivo depletion of TNF-α abrogated important advantages conferred by NO deficiency since only the iNOS−/− strain showed precocious mortality rates associated with non-organized pulmonary lesions . Importantly , early after interrupting anti-TNF treatment , the levels of TNF-α returned to normal levels and a massive influx of activated T cells and macrophages into the lungs occurred in both mouse strains . Therefore , it became clear that besides mortality rates and organization of lesions , TNF-α also controlled the migration of inflammatory cells to the site of infection . As a whole , these results led us to demonstrate that the concomitant deficiency of NO and TNF-α is fatal to iNOS-deficient mice , whereas the still preserved ability of NO synthesis by TNF-depleted WT mice appeared to rescue this mouse strain from precocious mortality . Depletion experiments of CD8+ T cells revealed the important role of this T cell subset in the early immunoprotection of iNOS−/− mice . Thus , the early differences in fungal loads were abrogated , and the influx of inflammatory cells was markedly impaired only in CD8-depleted mice iNOS−/− mice . Studies on the phenotype of DCs at the site of infection showed that anti-CD8 treatment did not alter the presence of CD8+ lymphoid DCs , suggesting that CD8+ T lymphocytes , and not lymphoid DCs , played an important control of fungal growth and inflammatory reactions mediated by T cells and macrophages . We could verify by intracellular cytokine staining that in iNOS−/− mice , the depletion of CD8+ T cells resulted in decreased numbers of CD4+ ( IFN-γ ) and CD8+ ( IFN-γ and TNF-α ) T cells , supporting the proinflammatory feature of this T cell subpopulation . This fact was consistent with the diminished influx of inflammatory cells observed in the lungs of CD8-depleted iNOS−/− mice . In WT mice , however , the depletion of CD8+ cells had a negligible effect , and this appears to reflect the concomitant increase of IFN-γ and TNF-α CD4+ T cells with decreased numbers of CD8+ T cells secreting the same proinflammatory cytokines in the lungs . The increased presence of IL-4+ CD4+ T cells appeared to have exerted a negligible effect in the inflammation mounted by WT mice . In conclusion , this work brought new information regarding the role of NO synthesis in experimental PCM . We demonstrated the protective effect of iNOS deficiency in pulmonary PCM . This protective effect appeared to be mediated by increased type-1 inflammatory reactions regulated by TNF-α production and expansion of IFN-γ and TNF-α-producing T cells .
|
Paracoccidiodomycosis is a human systemic mycosis endemic in Latin America that has a wide spectrum of manifestations ranging from localized to fatal disseminated forms . Both in humans and experimental models , immunoprotection is mediated by T cell immunity whereas immunosuppression is associated with the severe forms of the disease . The literature shows that nitric oxide ( NO ) produced by the enzyme nitric oxide synthase-2 ( NOS2 or iNOS ) is the major fungicidal component of phagocytic cells . The role of NO production was previously investigated in the intra-peritoneal and intravenous murine models of Paracoccidioides brasiliensis infection . The human paracoccidioidomycosis is believed to be acquired by the respiratory route , thus our study aimed to characterize the role of NO production in a pulmonary model of infection . We verified that , paradoxically , absence of NO production by iNOS- deficient mice resulted in less severe disease and increased survival times . This was associated with increased development of cellular immunity and enhanced synthesis of TNF-α which enhances cell migration to the site of infection and contributes to the better organization of lesions . Our work highlighted the deleterious effect of excessive NO production in pulmonary paracoccidioidomycosis , and demonstrated that uncontrolled fungal growth can be overridden by an efficient immune response .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"immunopathology",
"mycology",
"immunity",
"microbial",
"pathogens",
"immunology",
"biology",
"microbiology",
"immune",
"response"
] |
2013
|
TNF-α and CD8+ T Cells Mediate the Beneficial Effects of Nitric Oxide Synthase-2 Deficiency in Pulmonary Paracoccidioidomycosis
|
Chloroquine combined with primaquine has been the recommended antimalarial treatment of Plasmodium vivax malaria infections for six decades but the efficacy of this treatment regimen is threatened by chloroquine resistance ( CQR ) . Single nucleotide polymorphisms ( SNPs ) in the multidrug resistance gene , Pvmdr1 are putative determinants of CQR but the extent of their emergence at population level remains to be explored . In this study we describe the prevalence of SNPs in the Pvmdr1 among samples collected in seven P . vivax endemic countries and we looked for molecular evidence of drug selection by characterising polymorphism at microsatellite ( MS ) loci flanking the Pvmdr1 gene . We examined the prevalence of SNPs in the Pvmdr1 gene among 267 samples collected from Pakistan , Afghanistan , Sri Lanka , Nepal , Sudan , São Tomé and Ecuador . We measured and diversity in four microsatellite ( MS ) markers flanking the Pvmdr1 gene to look evidence of selection on mutant alleles . SNP polymorphism in the Pvmdr1 gene was largely confined to codons T958M , Y976F and F1076L . Only 2 . 4% of samples were wildtype at all three codons ( TYF , n = 5 ) , 13 . 3% ( n = 28 ) of the samples were single mutant MYF , 63 . 0% of samples ( n = 133 ) were double mutant MYL , and 21 . 3% ( n = 45 ) were triple mutant MFL . Clear geographic differences in the prevalence of these Pvmdr mutation combinations were observed . Significant linkage disequilibrium ( LD ) between Pvmdr1 and MS alleles was found in populations sampled in Ecuador , Nepal and Sri Lanka , while significant LD between Pvmdr1 and the combined 4 MS locus haplotype was only seen in Ecuador and Sri Lanka . When combining the 5 loci , high level diversity , measured as expected heterozygosity ( He ) , was seen in the complete sample set ( He = 0 . 99 ) , while He estimates for individual loci ranged from 0 . 00–0 . 93 . Although Pvmdr1 haplotypes were not consistently associated with specific flanking MS alleles , there was significant differentiation between geographic sites which could indicate directional selection through local drug pressure . Our observations suggest that Pvmdr1 mutations emerged independently on multiple occasions even within the same population . In Sri Lanka population analysis at multiple sites showed evidence of local selection and geographical dispersal of Pvmdr1 mutations between sites .
Malaria is one of the world’s leading causes of mortality and morbidity . Since late in the 1940s the antimalarial drug chloroquine ( CQ ) has been the primary chemotherapeutic for prophylaxis and treatment of malaria because of its good safety profile , low cost and high efficacy against the blood stages of CQ sensitive ( CQS ) Plasmodium parasites , causing the disease . Since the 1950`s , CQ treatment of P . vivax infections has been combined with the hypnozoitocidal drug primaquine ( PQ ) for clearance of the latent P . vivax liver stages , responsible for later relapses of the disease [1–3] . Compared to P . falciparum , development of CQ resistance ( CQR ) in P . vivax has been relatively slow with the first reports emerging in 1989 in Papua New Guinea ( PNG ) [4] . Since then , CQR has spread and today it is considered to be present in vivax-malaria endemic countries all over the world ( reviewed in [5] and more recently in [6] ) . Development of CQR has been slower in P . vivax than P . falciparum and this is sometimes attributed to the use of combined treatment ( CQ with PQ ) , where PQ acts synergistically with CQ against CQR parasites [7] . It is also proposed that CQR in P . vivax has a different CQR mechanism than P . falciparum [8] . Knowledge of the mode of action of CQR in P . vivax is limited . In P . falciparum , reduced CQ sensitivity is strongly associated with single nucleotide polymorphisms ( SNPs ) in the chloroquine resistance transporter-gene , Pfcrt [9;10] . However , studies of the Pfcrt orthologue in P . vivax , Pvcg10 , have not been able to find an association to CQR [8;11] . In P . falciparum is the P-glycoprotein–like molecule Pgh-1 encoded by Pfmdr1 , is also associated with CQR though it may only modulate the effects of the Pfcrt gene [12;13] . In 2005 , Brega et al . characterized the mdr-like gene Pvmdr1 in P . vivax isolates [14] , and evidence suggests that SNPs in the Pvmdr1 gene are a possible genetic determinant of CQR [11;14;15] . Cross-species comparisons led the focus in P . vivax to be primarily on the mdr-codons orthologous to codons implicated in P . falciparum CQR namely 86 , 184 , 1034 , 1042 and 1246 [14;16] . However codons 91 and 189 which are homologous to codons 86 and 184 in P . falciparum [14;17] and codons 1071 and 1079 which are homologous to codons 1034 and 1042 in P . falciparum [14] are rarely polymorphic in Pvmdr1 Instead SNPs at codons , 976 and 1076 , have been detected multiple times [11;14;15;18;19] . Suwanarusk et al . observed an association between the Y976F mutation and increased CQ IC50 in samples from Thailand and Papua province of Indonesia , and stated that the Y976F mutation is a useful tool to indicate foci of chloroquine resistance [11] . Others detect the mutations , but doubt their association with CQR [15;18;20–23] . Studies of P . falciparum drug resistance loci have used flanking microsatellite ( MS ) variation to describe selective sweeps around Pfdhfr [24] and Pfdhps [25] and Pfcrt [26] . This approach has revealed lineages of drug resistance mutant alleles which are derived from a single emergence event . Notably , in P . falciparum some resistance lineages were found to have spread across vast geographical distances [26;27] . When the same approach was repeated antifolate drug targets in P . vivax Pvdhfr [28] and Pvdhps [29] contrasting results were found . In Pvdhfr and Pvdhps there was evidence of multiple independent mutation events with little selective sweep around mutant alleles at those loci . This result may reflect the limited antifolate drug selection pressure that has historically been applied to P . vivax , or it may point to differences in transmission and selection dynamics in the two species . In this study we looked for evidence of drug selection on the CQR candidate Pvmdr using MS flanking the Pvmdr1 gene . We analysed samples collected from Pakistan , Afghanistan , Sri Lanka , Nepal , Sudan , São Tomé and Ecuador . In Pakistan approximately 83% of the malaria cases are caused by P . vivax and in Afghanistan , 95% are P . vivax . In both countries P . vivax infections are still being treated with CQ + PQ [30] . Around 83% of reported malaria infections in Sri Lanka were caused by P . vivax and CQ with PQ were still efficient and recommended treatment of P . vivax infections on the island until autochthonous cases of both P . falciparum and P . vivax in Sri Lanka fell to zero [30] . In Nepal 88% of the malaria cases are caused by P . vivax and treated with CQ + PQ [30] . In Ecuador , P . vivax accounts for 86% of all malaria infections and is treated with CQ + PQ [30] . To the best of our knowledge no cases of CQR have been reported from either Nepal or Ecuador . Only , 5% of the malaria infections in Sudan are caused by P . vivax and these are treated with artemether-lumefantrine + PQ [30] . No reports of P . vivax CQR have been published from Sudan . Falciparum malaria is the main cause of malaria in São Tomé and no recommendations are provided regarding treatment of P . vivax [30] . The objectives of the present study were to 1 ) Determine the diversity of SNPs in the Pvmdr1 gene , a putative marker of CQR , in P . vivax samples collected from Pakistan , Afghanistan , Sri Lanka , Nepal , Ecuador , Sudan and São Tomé and 2 ) Characterise flanking MS variation and use this to explore the evolutionary origin of Pvmdr mutations .
The total number of P . vivax samples analysed in this study was 267 . The samples originated from 7 countries: Pakistan ( n = 36 ) , Afghanistan ( n = 13 ) , Nepal ( n = 55 ) , Sri Lanka ( n = 136 ) , Ecuador ( n = 17 ) , São Tomé ( n = 4 ) and Sudan ( n = 6 ) . The samples were all ( with the exception of Sao Tomé , and Sudan ) derived from larger sets of PCR positive samples and the subset selected by computer randomisation . The DNA from a total of 263 P . vivax samples selected were already extracted as part of a P . vivax microsatellite study described previously [31] . Pakistan and Afghanistan: Forty-nine samples from a cluster of neighbouring sites in Pakistan and Afghanistan were analysed . Thirty-six were from Pakistan ( n = 36 ) ; Ashaghroo refugee camp in Adizai from 2003 ( n = 10 ) , Adizai Refugee Village in Peshawar from 2004–2005 ( n = 3 ) sampled as part of another study [32] and lastly Adizai , Baghicha and Khagan villages located near Peshawar 2005–2006 ( n = 23 ) described in [33] . Thirteen samples were from Afghanistan collected at the Malaria Reference Center in Jalalabad in 2004–2005 [32] . The samples from all these sites were grouped together because of similar study designs and close geographical distance between the sites . Sri Lanka: The samples from Sri Lanka were collected in 9 malaria endemic districts during 2002–2007 , see [34] . For this study , the samples were divided into 9 groups of districts , where after randomised computerisation was used to select samples from each district ( N = 136 ) . Nepal: Samples from two separate studies in Nepal were grouped together . Thirty-eight samples collected in 2009–2010 from the districts of Jhapa ( N = 34 ) and Banke , Chitwan and Dang ( N = 5 ) have been previously described in [35] . The other study collected samples in the districts of Kanchanpur ( N = 5 ) and Jhapa ( N = 12 ) in 2005–2006 as a part of a cross-sectional prevalence survey estimating the malaria burden and risk behaviour in two endemic districts of Nepal ( S . Hewitt , personal communication ) . The Kanchanpur samples were grouped with Banke , Chitwan and Dang . Ecuador: Twenty-one P . vivax samples were collected from 2007–2009 in the Province of Sucumbíos through the network of laboratories of the Ministry of Health . Sudan: Six P . vivax samples from Sudan were collected in the village of Asar in Gedaref state in 2006 as a part of an artemether-lumefantrine efficacy trial community based survey [36] . The amount of available P . vivax DNA was small , and only limited analysis was possible . São Tomé: The island São Tomé is a part of the Democratic Republic of São Tomé and Príncipe in the western equatorial coast of Central Africa . Four samples were available . As with the samples from Sudan , limited analysis was possible because only a small amount of extracted DNA-solution was available . A fragment spanning nucleotide 2596 and 3532 ( amino acids 865–1177 ) of the Pvmdr1 gene and was amplified using semi-nested primers Pvmdr1-4F [11] , Pvmdr1-AS and Pvmdr1-S [14] . Primer sequences are shown in Table 1 . Cycling conditions were as follows: 94°C for 15 min , followed by 30 cycles of 94°C for 30 s , 55°C for 1 min , and 72°C for 1 min , and finally 72°C for 10 min . The amplified Pvmdr1 fragments were sequenced on an ABI Prism 377 ( Perkin-Elmer ) using the Big Dye terminator reaction mix ( Perkin-Elmer ) . After sequencing , the individual haplotypes were aligned and analysed by use of the DNASTAR-Lasergene software . The Pvmdr1 gene is located on chromosome 10 , and sequences flanking the gene were screened for suitable microsatellite marker loci in the Salvador-1 ( Sal-1 ) reference strain ( accessed through the European Bioinformatics Institute homepage ( www . ncbi . nlm . nih . gov/ ) . Multiple repeats were identified using the software Tandem Repeats Finder [37] and semi-nested primers designed ( Table 1 ) . The primary reaction comprised of 1μl template , 0 . 5 unit Taq polymerase , 1 . 1μl Thermopol Reaction buffer ( New England Biolabs Inc , Glostrup , Denmark ) , and 0 . 4μM dNTPs , 0 . 1μM of forward ( F ) and reverse primers ( R ) with cycling conditions as follows; 2 min at 94°C and then 25 repeated cycles of 30 s at 94°C , 30 s at 42°C , 30 s at 40°C and 40 s at 65°C followed by 2 min at 65°C and a minimum of 10 min at 15°C . In the secondary PCR , the same concentrations of reagents were added , but with 0 . 15μM of reverse primers ( R ) and fluorescent-labelled inverse primers ( I ) . The cycling conditions were initiated with 2 min at 94°C followed by 25 repeated cycles of 20 s at 94°C , 20 s at 45°C , 30 s at 65°C , and finished with 2 min at 65°C and 10 min at 15°C . A panel of four MS were selected for further analysis on the basis of their successful amplification and were named according to their distance to the Pvmdr1 gene: m9 . 5 ( 9 , 489 bp downstream of Pvmdr1 ) , m10 . 1 ( 10 , 120 bp downstream ) , m10 . 4 ( 10 , 420 upstream ) and m43 . 1 ( 43 , 168 bp upstream ) , ( Fig 1 ) . PCR amplified fragments were run with LIZ-500 size standard on an ABI 3730XL genetic analyzer ( Applied Biosystems ) , and analysed using the GeneMapper software ( Applied Biosystems ) . Samples that were negative by PCR were repeat amplified with 2μl template in the first PCR . The number and length of the repeats in each of the four MS is shown in Table 1 . In the cases where multiple ( ≥ 2 ) microsatellite alleles were detected in a single sample the major/predominant allele chosen , ( ‘predominant’ is defined by the electropherogram peak height which had to be twice that of the minor allele ) . Linkage disequilibrium ( LD ) was calculated to test for a non-random association of Pvmdr1 allele and the MS alleles . Only the non-mixed samples were used in the calculations . LD was measured by the formula D' = D/Dmax , where D equals derivation of random association between alleles at different loci , and D’ measures D standardized by the maximum value ( Dmax ) , given the observed allele frequencies . LD values range from -1 to +1 , where the value +1 refers to a complete non-random association between the alleles . Values of gene diversity were calculated by expected heterozygosity by the formula He = ( n/[n-1] ) ( 1-∑pi2 ) , by use of the Arlequin software [38] , where He is expected heterozygosity , n the number of samples , and pi the frequency of the i-th allele in the sample set . Clearance for analysis of Plasmodium genes were approved by London School of Hygiene Tropical Medicine and Hygiene Ethics Board , locally by Bioethics Committee , Pakistan Medical Research Council and Directorate of Public Health , Jalalabad , Nangahar , Comitee de Bioetica Universidad San Francisco de Quito , Committee on Research and Ethical Review at the Faculty of Medicine , Peradeniya , Kandy and the Nepal Health Research Council . All data analysed were anonymized .
The analysis of Pvmdr1 in 39 samples from Nepal was previously published [35] . Of the remaining 228 samples , sequencing was successful in 173 ( 75 . 9% ) ; Pakistan ( n = 24 ) , Nepal ( n = 4 ) , Sri Lanka ( n = 120 ) , São Tomé ( n = 3 ) , Sudan ( n = 4 ) and Ecuador ( n = 17 ) . Combined with the 39 samples from Nepal , 212 Pvmdr1 sequence fragments were available for further analysis ( Table 2 , Fig 2 ) . SNP variation in fragment 2 was largely confined to three codons; T958M ( ACG→ATG ) , Y976F ( TAC→TTC ) and F1076L ( TTT→CTT ) , ( Fig 1 ) . Three novel SNPs were detected in two samples from the Jhapa district in Nepal . These were sequenced twice to confirm the results . One of the samples carried a SNP at c1080 ( S1080N , AGT→AAT ) , while the other sample possessed SNPs at c979 ( F979S , TTT→TCT ) and c980 ( M980V , ATG→GTG ) ( Table 3 ) . These two samples and the SNPs were described by Ranjitkar et al . [35] . The substitutions at codons 958 , 976 , and1076 were found in various configurations . TYF ( the wild-type ) was found only in Ecuador in 5 of 17 samples ( Fig 2 , S1 Table ) . All the remaining samples carried one of three mutant haplotypes , MYF ( single mutant , T958M ) , MYL ( double mutant , T958M and F1076L ) and MFL ( triple mutant T958M , Y976F and F1076L ) . The double mutant MYL was present in 63 . 0% of the samples ( n = 133 ) , the triple mutant MFL in 21 . 3% ( n = 45 ) , and the single mutant MYF in 13 . 3% ( n = 28 ) . Their relative abundance at the different sites is shown in Fig 2 . In Pakistan only the MYL double mutant haplotype was detected , while the Ecuador samples ( apart from the wild-type TYF ) possessed the single mutant MYF ( n = 12 ) . The Sudan P . vivax samples a mix of MYL ( n = 2 ) and MFL ( n = 2 ) haplotypes were found , while the three samples from São Tomé all possessed the MFL haplotype ( Fig 2 ) . The heterozygosity of Pvmdr1 measured as He is shown in Table 2 . Measured over all samples He was 0 . 54 . When divided by collection site , Sri Lanka was the most diverse ( He = 0 . 54 ) , and Pakistan the least diverse ( He = 0 ) with only one allele-the MYL haplotype . Although the He value for Sudan was high ( 0 . 67 ) , the sample size was small ( n = 4 ) , and the broad variation in sample size between sites precluded further in-depth analysis of difference between the sites . District level analysis was possible for Sri Lanka ( Fig 3 ) . In Sri Lanka , samples were collected in 9 districts , and despite the small sample size per district , the distribution of alleles was similar at district level when compared to the pooled sample set , with a dominating MYL haplotype , followed by MFL and MYF . The exception was Kurunegala district where the MFL haplotype was most common ( n = 15 ) , followed by the MYL haplotype ( n = 5 ) . Four MS from Pvmdr1 flanking genomic regions were genotyped in the 267 samples although these were amplified with varying successes; for m9 . 5 ( n = 190 ) , m10 . 1 ( n = 196 ) , m10 . 4 ( n = 229 ) and m43 . 1 ( n = 181 ) alleles ( Table 2 ) . The m10 . 1 locus was the most polymorphic with 26 alleles identified among 196 samples ( He = 0 . 90 ) , while m9 . 5 had 4 alleles , m10 . 4 had 13 and m43 . 1 had 3 different alleles ( Table 2 , and S1 Table ) . The number of mixed samples ( those containing more than one allele ) detected using each locus differed; only 1% were mixed at the m9 . 5 locus , while 12% were mixed at the m43 . 3 locus . When combining all 5 loci , 26% were mixed among 125 samples ( Table 2 ) . The m10 . 1 locus had a mono-A-repeat as well as an AT dinucleotide repeat , and this was reflected in the high number of different alleles ( He = 0 . 90 ) . Ecuador was an exception , only possessing 4 different m10 . 1 alleles among the 16 positive samples ( He = 0 . 68 ) . Allele size variation is shown in S1 Table , the highest number of observed MS alleles were generally of intermediate size . The combination of all 5 loci resulted in 57 different 5-loci haplotypes among the monoclonal samples ( Table 3 ) and 78 when majority alleles in the mixed genotype samples were included . All 5-locus haplotypes differed from the CQS wild-type Sal-1 haplotype . The most commonly observed 5-locus haplotype was detected in Ecuador ( n = 6 , haplotype number 8 in Table 3 ) . The other frequently observed 5-loci haplotypes were all from Sri Lanka . The distribution of the Pvmdr1 5- locus haplotypes among the sub-populations sampled in Sri Lanka is shown in Fig 4 . A large number of haplotypes occurred only once and these are indicated by grey segments in the pie charts . Haplotypes which occurred multiple times are indicated each by a different colour . The sample collections with the greatest degree of haplotype sharing were from Trincomalee and Anuradhapura ( 6 haplotypes ) and Anuradhapura and Polonaruwa ( 5 haplotypes of which only 1 was found in Trincomalee ) . Other sites appear more isolated , for instance all the haplotypes found in the district of Batticaloa were unique . Unfortunately , the small amount of DNA-solution available from Sudan and São Tomé prevented reanalysis of these samples , and no 5-loci haplotype from Sao Tomé could be created . We tested for LD , between Pvmdr1 alleles and flanking MS alleles . Significant associations are shown in S2 Table . Strong linkage associations were seen in Ecuador where MS alleles occurred in association with the MYF single mutant allele and also with the wildtype allele TYF . Other population level LD associations were observed in Sri Lanka , and Nepal . In each case different MS alleles were associated with the Pvmdr1 allele . When the 4 MS loci were combined , significant LD between certain 4-loci haplotypes and either the TYF , MYF or MFL haplotypes was found in Ecuador and Sri Lanka , ( S2 Table ) . When samples from all sites were pooled the LD analysis found significant associations of MS alleles with TYF and MFL which are likely attributable to admixture .
The aim of this study was to characterise SNPs in the Pvmdr1 gene , to examine whether the putative CQR mutations have one , few or many origins , and to determine whether there has been geographical spread of certain Pvmdr1 haplotypes . SNPs were found at three codons , T958M , Y976F and F1076L among 267 samples from Pakistan , Nepal , Sri Lanka , Ecuador , Sudan and São Tomé . Polymorphism in the last two codons has been described in multiple studies [11;14;15;18;19;22;23] , but the high prevalence of 958M which we observed ( 206/211 , 97 . 6% ) was surprising since this SNP has only been mentioned in two earlier studies; in Madagascar ( with a 100% fixation of the 958M ) [18] and in a few samples from Indonesia and Brazil [39] . Besides these studies , all others either report the presence of the wild-type T958 allele , or do not mention the locus [11;20–23] . Since the 958M mutation was present in countries with both high and low level of reported CQR over a wide time span , we hypothesize that the T958M is an allelic variant of the wildtype and most likely not associated with CQR . In the present study the T958 wild-type was only detected in Ecuador . It is also seen in the Sal-1 reference sample which originates from El Salvador , so it is possible this allele might be a characteristic of American samples while the 958M is characteristic of Asia and Africa . Rare mutations , F979S ( TTT→TCT ) , M980V ( ATG→GTG ) and S1080N ( AGT→AAT ) were found in two samples from the Jhapa district of Nepal , both possessing the MYL double mutant Pvmdr1 haplotype ( Table 3 ) ; One of these samples was mutated at codon 1080 while the other was mutated at codon 979 and codon 980 . These results have been previously published by Ranjitkar in 2011 [35] . Thus , in total only five Pvmdr1 SNPs were detected which was an unexpected result . Orjuela-Sanchez et al . ( 2009 ) [23] reported up to 24 Pvmdr1 mutations in a study of only 7 samples from Brazil , while Barnadas et al . ( 2008 ) [18] reported 21 mutations among 105 samples from Madagascar . However , both studies amplified longer fragments of the Pvmdr1 gene than the present study , which might be a part of the explanation . Genotyping of microsatellites flanking the Pvmdr1 gene revealed high levels of diversity around single , double and triple mutant alleles . There were too few wildtype TYF alleles to meaningfully compare the MS heterozygosity surrounding wildtype and mutant alleles for evidence of selective sweeps on the mutant alleles . However our finding that all 3 wildtype TYF alleles were flanked by an identical microsatellite haplotype would not support the view that reduced diversity among microsatellite haplotypes is attributable to selective sweeps , but rather suggests a tendency to clonal population structure in some populations . The evidence for association of MS alleles with particular mutations was patchy . TYF and MYF Pvmdr1 haplotypes occurred together with the “201” m9 . 5 allele , while the “204” allele at m9 . 5 was more commonly seen with MFL and MYL . No obvious pattern of distribution was seen for the other 3 MS , suggesting this association was caused by greater representation of certain mutant alleles in particular geographic localities rather than a selective sweep . The combination of the 5 Pvmdr1 loci into a 5-loci haplotype revealed 57 different haplotypes among the 125 samples positive at all loci , many of them unique . Country-wise , Nepal was the most diverse , when analysed by locus and for the combined 5-loci haplotype , whereas Ecuador was more conserved . The diminished diversity within the Ecuadorian samples is consistent with the general finding of little diversity amongst P . vivax samples from the Americas , although this is not a hard and fast rule [40] . Just 10 samples of African origin were available for this study but both double and triple Pvmdr1 mutant alleles were present , and their microsatellite fingerprint was distinct from that associated with the same alleles in Asia . Likewise , our South American sample from Ecuador ( n = 17 ) was distinct from the other populations being less diverse , and unique in having the wild-type Pvmdr1 allele . Studies of P . falciparum have revealed a contrasting pattern of resistance evolution in which relatively few resistance mutants emerge but then become globally disseminated . The pattern is consistent for both CQR [26;41] and high levels of resistance to SP [24;27;42–44] . Hawkins et al . [28;29] analysed the origin and dissemination of SPR in P . vivax by analysing SNPs in and surrounding the Pvdhfr and Pvdhps genes . They concluded that the genes are considerably more diverse than seen in P . falciparum [28;29] and that highly pyrimethamine-resistant Pvdhfr alleles arose three times in Thailand , Indonesia and PNG/Vanuatu , and that sulfadoxine resistance associated SNPs had evolved independently on multiple occasions . This is consistent with our findings on Pvmdr and may be explained by comparisons of total genomic diversity among P . vivax and P . falciparum isolates . A study by Neafsey et al . [45] reported twice as much SNP diversity , significantly higher MS diversity and a far deeper divergence among P . vivax geographic isolates than among a comparable set of P . falciparum isolates . The higher level of diversity in P . vivax can explain the multiple origin pattern of resistance emergence in of Pvmdr1 , Pvdhfr and Pvdhps . Our findings can indicate three things . First , Pvmdr1 mutant alleles have developed on multiple haplotype backgrounds by convergent evolution in Pakistan , Nepal , Sri Lanka , Ecuador , Sao Tomé and Sudan . Second , assuming that Pvmdr1 is a reliable CQR marker , there is little evidence that the variation around mutant haplotypes has been subject to a selective sweep , ( the result of positive natural selection causing diminished diversity in sequences flanking the selected marker ) . Third , the historical pattern of drug resistance emergence in P . falciparum is not repeated in P . vivax . The time-delay of around 30 years between initial reports of CQR in the two malaria species , and the two different treatment regimens ( mono-in P . falciparum and usually combined CQ/primaquine treatment in P . vivax ) might explain this , and may suggest that it is just a matter of time before the effect of selection of the markers can be seen since high grade treatment failure has not yet been reported in any of the sites sampled in this study . P . vivax is generally a chronic disease with low parasitaemia causing mild symptoms compared to P . falciparum , therefore there is less selective drug pressure on the parasite . Furthermore , the fast gametocytogenesis in P . vivax enables uptake of gametes by vectors before clinical symptoms arise and antimalarial treatment is initiated which balances the spread of sensitive and resistant P . vivax parasites . Equally , these differences may lie with the biology and transmission dynamics of the two species . Our district level analysis of Pvmdr and linked MS variation in Sri Lanka found evidence of exchange of genotypes between districts which may or may not be linked to their resistance phenotype . Latent P . vivax infections cause relapses of the disease , which may increase the possibility of migration of parasites within the human host while the broader temperature tolerance by P . vivax parasites compared to P . falciparum might increase the likelihood of gene flow between sites with varying temperature or microclimate . Finally CQR may be a complex trait including other genes in addition to Pvmdr1 and future research will hopefully illuminate the genomic-level change underpinning changes in CQ sensitivity . Meanwhile , monitoring and research of CQR is of highest importance for the public health in the afflicted areas .
|
Chloroquine combined with primaquine has been the recommended antimalarial treatment for Plasmodium vivax malaria infections for sixty years but the efficacy of this treatment regimen is threatened by chloroquine resistance . In this study we describe the prevalence of mutations in the P . vivax gene , Pvmdr1 among samples collected in seven endemic countries . The mutations are thought to be associated with chloroquine resistance and here we looked for evidence of drug selection by characterising polymorphism in DNA repeat regions ( microsatellite ( MS ) loci ) flanking the Pvmdr1 gene . Mutations in the Pvmdr1 gene were mainly identified at codons T958M , Y976F and F1076L . Just 2 . 4% of samples were wildtype at all three codons , while 63% were single mutants ( MYF ) . Clear geographic differences in the prevalence of these Pvmdr mutation combinations were observed . At the flanking MS loci , we found high levels of diversity , and significant differentiation between geographic sites . This pattern of variation could indicate directional selection through local drug pressure . In summary , our observations suggest that Pvmdr1 mutations and thus , chloroquine resistance has emerged independently on multiple occasions even within the same population .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Multiple Origins of Mutations in the mdr1 Gene—A Putative Marker of Chloroquine Resistance in P. vivax
|
Many previous studies on visual search have reported inter-trial effects , that is , observers respond faster when some target property , such as a defining feature or dimension , or the response associated with the target repeats versus changes across consecutive trial episodes . However , what processes drive these inter-trial effects is still controversial . Here , we investigated this question using a combination of Bayesian modeling of belief updating and evidence accumulation modeling in perceptual decision-making . In three visual singleton ( ‘pop-out’ ) search experiments , we explored how the probability of the response-critical states of the search display ( e . g . , target presence/absence ) and the repetition/switch of the target-defining dimension ( color/ orientation ) affect reaction time distributions . The results replicated the mean reaction time ( RT ) inter-trial and dimension repetition/switch effects that have been reported in previous studies . Going beyond this , to uncover the underlying mechanisms , we used the Drift-Diffusion Model ( DDM ) and the Linear Approach to Threshold with Ergodic Rate ( LATER ) model to explain the RT distributions in terms of decision bias ( starting point ) and information processing speed ( evidence accumulation rate ) . We further investigated how these different aspects of the decision-making process are affected by different properties of stimulus history , giving rise to dissociable inter-trial effects . We approached this question by ( i ) combining each perceptual decision making model ( DDM or LATER ) with different updating models , each specifying a plausible rule for updating of either the starting point or the rate , based on stimulus history , and ( ii ) comparing every possible combination of trial-wise updating mechanism and perceptual decision model in a factorial model comparison . Consistently across experiments , we found that the ( recent ) history of the response-critical property influences the initial decision bias , while repetition/switch of the target-defining dimension affects the accumulation rate , likely reflecting an implicit ‘top-down’ modulation process . This provides strong evidence of a disassociation between response- and dimension-based inter-trial effects .
In everyday life , we are continuously engaged in selecting visual information to achieve our action goals , as the amount of information we receive at any time exceeds the available processing capacity . The mechanisms mediating attentional selection enable us to act efficiently by prioritizing task-relevant , and deprioritizing irrelevant , information . Of importance for the question at issue in the present study , the settings that ensure effective action in particular task episodes are , by default , buffered by the attentional control system and carried over to subsequent task episodes , facilitating performance if the settings are still applicable and , respectively , impairing performance if they no longer apply owing to changes in the task situation ( in which case the settings need to be adapted accordingly ) . In fact , in visual search tasks , such automatic carry-over effects may account for more of the variance in the response times ( RTs ) than deliberate , top-down task set [1] . A prime piece of evidence in this context is visual search for so-called singleton targets , that is , targets defined by being unique relative to the background of non-target ( or distractor ) items , whether they differ from the background by one unique feature ( simple feature singletons ) or a unique conjunction of features ( conjunction singletons ) : singleton search is expedited ( or slowed ) when critical properties of the stimuli repeat ( or change ) across trials . Such inter-trial effects have been found for repetitions/switches of , for example , the target-defining color [2 , 3] , size [4] , position [5] , and , more generally , the target-defining feature dimension [6 , 7] . The latter has been referred to as the dimension repetition/switch effect , that is: responding to a target repeated from the same dimension ( e . g . , color ) is expedited even when the precise target feature is different across trials ( e . g . , changing from blue on one trial to red on the next ) , whereas a target switch from one dimension to another ( e . g . , from orientation to color ) causes a reaction time cost ( ‘dimension repetition effect’ , DRE ) [8–10] . While inter-trial effects have been extensively studied , the precise nature of the processes that are being affected remains unclear . Much of the recent work has been concerned with the issue of the processing stage ( s ) at which inter-trial effects arise ( for a review , see [11] ) . Müller and colleagues proposed that inter-trial effects , in particular the dimension repetition effect , reflect facilitation of search processes prior to focal-attentional selection ( at a pre-attentive stage of saliency computation ) [10] . However , using a non-search paradigm with a single item presented at a fixed ( central ) screen location , Mortier et al . [12] obtained a similar pattern of inter-trial effects–leading them to conclude that the DRE arises at the post-selective stage of response selection . Rangelov and colleagues [13] demonstrated that DRE effects can originate from distinct mechanisms in search tasks making different task demands ( singleton feature detection and feature discrimination ) : pre-attentive weighting of the dimension-specific feature contrast signals and post-selective stimulus processing–leading them to argue in favor of a 'multiple weighting systems hypothesis' . Based on the 'priming of pop-out' search paradigm , a similar conclusion [11] has also been proposed , namely , inter-trial effects arise from both attentional selection and post-selective retrieval of memory traces from previous trials [4 , 14] , favoring a dual-stage account [15] . It is important to note that those studies adopted very different paradigms and tasks to examine the origins of inter-trial effects , and their analyses are near-exclusively based on differences in mean RTs . Although such analyses are perfectly valid , much information about trial-by-trial changes is lost . Recent studies have shown that the RT distribution imposes important constraints on theories of visual search [16 , 17] . RT distributions in many different task domains have been successfully modeled as resulting from a process of evidence accumulation [18 , 19] . One influential evidence accumulation model is the drift-diffusion model ( DDM ) [20–22] . In the DDM , observers sequentially accumulate multiple pieces of evidence , each in the form of a log likelihood ratio of two alternative decision outcomes ( e . g . , target present vs . absent ) , and make a response when the decision information reaches a threshold ( see Fig 1 ) . The decision process is governed by three distinct components: a tendency to drift towards either boundary ( drift rate ) , the separation between the decision boundaries ( boundary separation ) , and a starting point . These components can be estimated for any given experimental condition and observer by fitting the model to the RT distribution obtained for that condition and observer . Estimating these components makes it possible to address a question that is related to , yet separate from the issue of the critical processing stage ( s ) and that has received relatively less attention: do the faster RTs after stimulus repetition reflect more efficient stimulus processing , for example: expedited guidance of attention to more informative parts of the stimulus , or rather a bias towards giving a particular one of the two alternative responses or , respectively , a tendency to require less evidence before issuing either response . The first possibility , more efficient processing , would predict an increase in the drift rate , that is , a higher speed of evidence accumulation . A bias towards one response or a tendency to require less evidence would , on the other hand , predict a decreased distance between the starting point and the decision boundary associated with that response . In the case of bias , this would involve a shift of the starting point towards that boundary , while a tendency to require less evidence would be reflected in a decrease of the boundary separation . While response bias is more likely associated with changes at the post-selective ( rather than pre-attentive ) processing stage , the independence of the response selection and the attentional selection stage has been challenged [23] . For simple motor latencies and simple-detection and pop-out search tasks [24] , there is another parsimonious yet powerful model , namely the LATER ( Linear Approach to Threshold with Ergodic Rate ) model [25 , 26] . Unlike the drift-diffusion model , which assumes that evidence strength varies across the accumulative process , the LATER model assumes that evidence is accumulated at a constant rate during any individual perceptual decision , but that this rate varies randomly across trials following a normal distribution ( see Fig 1 ) . Such a pattern has been observed , for instance , in the rate of build-up of neural activity in the motor cortex of monkeys performing a saccade-to-target task [27] . Similar to the DDM , the LATER model has three important parameters: the ergodic rate ( r ) , the boundary separation ( θ ) , and a starting point ( S0 ) . However , the boundary separation and starting point are not independent , since the output of the model is completely determined by the rate and the separation between the starting point and the boundary; thus , in effect , the LATER model has only two parameters . The evidence accumulation process can be interpreted in terms of Bayesian probability theory [26 , 28] . On this interpretation , the 'linear approach to threshold with ergodic rate' represents the build-up of the posterior probability that results from adding up the log likelihood ratio ( i . e . , 'evidence' ) of a certain choice being the correct one and the initial bias that derives from the prior probability of two choices . The prior probability should affect the starting point S0 of the evidence accumulation process: S0 should be the closer to the boundary the higher the prior probability of the outcome that boundary represents . The drift rate , by contrast , should be influenced by any factor that facilitates or impedes efficient accumulation of task-relevant sensory evidence , such as spatial attentional selection . The present study was designed to clarify the nature of the inter-trial effects for manipulations of target presence and the target-defining dimension as well as inter-trial dimension repetitions and switches . If inter-trial effects reflect a decision bias , this should be reflected in changes of the decision boundary and/or the starting point . By contrast , if inter-trial effects reflect changes in processing efficiency , which might result from allocating more attentional resources ( or 'weight' ) to the processing of the repeated feature/dimension [6] , the accumulation rate r should be changed . Note that neither the DDM nor the LATER model provides any indication of how the initial starting point might change across trials . Given that the inter-trial effects are indicative of the underlying trial-by-trial dynamics , we aimed to further analyze trial-wise changes of the prior and the accumulation rate , and examine how a new prior is learned when the stimulus statistics change , as reflected in changes of the starting point to decision boundary separation during the learning process . To address these inter-trial dynamics , we adopted the Dynamic Belief Model ( DBM ) [29] . The DBM has been successfully used to explain why performance on many tasks is better when a stimulus matches local patterns in the stimulus history even in a randomized design where it is not actually possible to use stimulus history for ( better-than-chance ) prediction . Inter-trial effects arise naturally in the DBM . This is because the DBM assumes a prior belief about non-stationarity , that is: participants are updating their beliefs about the current stimulus statistics while assuming that these can change at any time . The assumption of non-stationarity leads to something similar to exponential discounting of previous evidence , that is , the weight assigned to previous evidence decreases exponentially with the time ( or number of updating events ) since it was acquired . Consequently , current beliefs about what is most likely to happen on an upcoming trial will always be significantly influenced by what occurred on the previous trial , resulting in inter-trial effects . Thus , here we combine a belief-updating model closely based on the DBM , for modelling the learning of the prior , with the DDM and , respectively , the LATER model for predicting RTs . A very similar model has previously been proposed to explain results in saccade-to-target experiments [30] . We also consider the possibility that the evidence accumulation rate as well as the starting point may change from trial to trial . To distinguish between different possible ways in which stimulus history could have an influence via updating of the starting point and/or the rate , we performed three visual search experiments , using both a detection and a discrimination task and manipulating the probability of target presence , as well as the target-defining dimension . Based on the RT data , we then performed a factorial model comparison ( cf . [31] ) , where both the response history and the history of the target dimension can affect either the starting point or the rate . The results show that the model that best explains both the effects of our probability manipulation and the inter-trial effects is the one in which the starting point is updated based on response history and the rate is updated based on the history of the target dimension .
The singleton search was quite easy , with participants making few errors overall: mean error rates were 1 . 5% , 2 . 5% , and 3 . 3% in Experiments 1 , 2 , and 3 respectively ( Fig 2 ) . Despite the low average error rates , error rates differed significantly between blocks in both Experiments 1 and 2 [F ( 1 . 34 , 14 . 78 ) =11 . 50 , p<0 . 01 , ηp2=0 . 51 , BF=8372 , and F ( 2 , 22 ) =12 . 20 , p<0 . 001 , ηp2=0 . 53 , BF=3729 , respectively]: as indicated by post-hoc comparisons ( S1 Text ) , error rates were higher in the low-frequency blocks compared to the medium- and high-frequency blocks , without a significant difference between the latter . In addition , in Experiment 1 , error rates were overall higher for target-present than for target-absent trials , that is , there were more misses than false alarms , F ( 1 , 11 ) =11 . 43 , p<0 . 01 , ηp2=0 . 51 , BF=75 . In contrast , there was no difference in error rates between color and orientation targets in Experiment 2 , F ( 1 , 11 ) = 0 . 70 , p = 0 . 42 , BF = 0 . 33 . In Experiment 3 , there was no manipulation of target ( or dimension ) frequency , but like in Experiment 1 , error rates were higher on target-present than on target-absent trials , t ( 11 ) = 4 . 25 , p < 0 . 01 , BF = 30 . 7; and similar to Experiment 2 , there was no significant difference in error rates between color and orientation targets , t ( 11 ) = 1 . 51 , p = 0 . 16 , BF = 0 . 71 . Given the low error rates , we analyzed only RTs from trials with a correct response , though excluding outliers , defined as trials on which the inverse RT ( i . e . , 1/RT ) was more than three standard deviations from the mean for any individual participant . Fig 3 presents the pattern of mean RTs for all three experiments . In both Experiments 1 and 2 , the main effect of frequency was significant [F ( 2 , 22 ) =10 . 25 , p<0 . 001 , ηp2=0 . 48 , BF=73 , and , respectively , F ( 1 . 27 , 13 . 96 ) =29 . 83 , p<0 . 01 , ηp2=0 . 73 , BF=8 . 7*108] . Post-hoc comparisons ( see S2 Text ) confirmed RTs to be faster in high-frequency compared to low-frequency blocks , indicative of participants adapting to the stimulus statistics in a way such as to permit faster responses to the most frequent type of trial within a given block . In addition , in Experiment 1 , RTs were faster for target-present than for target-absent trials [F ( 1 , 11 ) =5 . 94 , p<0 . 05 , ηp2=0 . 35 , BF=51] , consistent with the visual search literature . In contrast , there was no difference between color- and orientation-defined target trials in Experiment 2 , and no interaction between target condition and frequency in either Experiment 1 or 2 ( S2 Text ) –suggesting that the effect of frequency is independent of the target stimuli . Comparing the error rates depicted in Fig 2 and the mean RTs in Fig 3 , error rates tended to be lower for those frequency conditions for which RTs were faster . While this rules out simple speed-accuracy trade-offs , it indicates that participants were adapting to the statistics of the stimuli in a way that permitted faster and more accurate responding to the most frequent type of trial within a given block , at the cost of slower and less accurate responding on the less frequent trial type . A possible explanation of these effects is a shift of the starting point of a drift-diffusion model towards the boundary associated with the response associated with the most frequent type of trial; as will be seen below ( in the modeling section ) , the shapes of the RT distributions were consistent with this interpretation . Without a manipulation of frequency , Experiment 3 yielded a standard outcome: all three types of trial yielded similar mean RTs , F ( 2 , 22 ) = 2 . 15 , p = 0 . 14 , BF = 0 . 71 . This is different from Experiment 1 , in which target-absent RTs were significantly slower than target-present RTs . This difference was likely obtained because the target-defining dimension was kept constant within short mini-blocks in Experiment 1 , but varied randomly across trials in Experiment 3 , yielding a dimension switch cost and therefore slower average RTs on target-present trials ( see modeling section for further confirmation of this interpretation ) . Given our focus on inter-trial dynamic changes in RTs , we compared trials on which the target condition was switched to trials on which it was repeated from the previous trial . Fig 4 illustrates the inter-trial effects for all three experiments . RTs were significantly faster on target-repeat than on target-switch trials , in all experiments: Experiment 1 [F ( 1 , 11 ) =6 . 13 , p<0 . 05 , ηp2=0 . 36 , BF=0 . 81] , Experiment 2 [F ( 1 , 11 ) =71 . 29 , p<0 . 001 , ηp2=0 . 87 , BF=2 . 6*107] , and Experiment 3 [F ( 1 , 11 ) =32 . 68 , p<0 . 001 , ηp2=0 . 75 , BF=625] . Note that for Experiment 1 , despite the significant target-repeat/switch effect , the ‘inclusion’ BF ( see Methods ) suggests that this factor is negligible compared to other factors; a further post-hoc comparison of repeat versus switch trials has a BF of 5 . 88 , compatible with the ANOVA test . The target repetition effect in all three experiments is consistent with trial-wise updating of an internal model ( see the modeling section ) . The target repetition/switch effect was larger for target-absent responses ( i . e . , comparing repetition of target absence to a switch from target presence to absence ) than for target-present responses in Experiment 3 ( interaction inter-trial condition x target condition , F ( 1 , 11 ) =14 . 80 , p<0 . 01 , ηp2=0 . 57 , BF=18 ) , while there was no such a difference in Experiment 1 , F ( 1 , 11 ) = 2 . 55 , p = 0 . 14 , BF = 0 . 43 , and also no interaction between target dimension and inter-trial condition in Experiment 2 , F ( 1 , 11 ) = 0 . 014 , p = 0 . 91 , BF = 0 . 76 . These findings suggest that , while the target repetition/switch effect as such is stable across experiments , its magnitude may fluctuate depending on the experimental condition . The interaction between target condition and inter-trial condition seen in Experiment 3 , but not in Experiment 1 , is likely attributable to the fact that color and orientation targets were randomly interleaved in Experiment 3 , so that target-present repetitions include trials on which the target dimension did either repeat or change–whereas the target dimension was invariably repeated on consecutive target-present trials in Experiment 1 . The effects of repeating/switching the target dimension are considered further below . Note that in all experiments , we mapped two alternative target conditions to two fixed alternative responses . The repetition and switch effects described above may be partly due to response repetitions and switches . To further examine dimension repetition/switch effects when both dimensions were mapped to the same response , we extracted those target-present trials from Experiment 3 on which a target was also present on the immediately preceding trial . Fig 5 depicts the mean RTs for the dimension-repeat versus -switch trials . RTs were faster when the target dimension repeated compared to when it switched , F ( 1 , 11 ) =25 . 06 , p<0 . 001 , ηp2=0 . 70 , BF=1905 , where this effect was of a similar magnitude for color- and orientation-defined targets [interaction target dimension x dimension repetition , F ( 1 , 11 ) = 0 . 44 , p = 0 . 84 , BF = 0 . 33] . There was also no overall RT difference between the two types of target [main effect of target dimension , F ( 1 , 11 ) = 0 . 16 , p = 0 . 69 , BF = 0 . 34] , indicating that the color and orientation targets were equally salient . This pattern of dimension repetition/switch effects is in line with the dimension-weighting account [8] . Of note , there was little evidence of a dimension repetition benefit from two trials back , that is , from trial n-2 to trial n: the effect was very small ( 3 ms ) and not statistically significant [t ( 23 ) = 0 . 81 , p = 0 . 43 , BF = 0 . 38] . In addition to inter-trial effects from repetition versus switching of the target dimension , there may also be effects of repeating/switching the individual target-defining features . To examine for such effects , we extracted those trials on which a target was present and the target dimension stayed the same as on the preceding trial , and examined them for ( intra-dimension ) target feature repetition/switch effects . See Fig 6 for the resulting mean RTs . In Experiments 1 and 3 , there was no significant main effect of feature repetition/switch [Exp . 1: F ( 1 , 11 ) = 0 . 30 , p = 0 . 593 , BF = 0 . 30 , Exp . 3: F ( 1 , 11 ) = 3 . 77 , p = 0 . 078 , BF = 0 . 76] , nor was there an interaction with target dimension [Exp . 1: F ( 1 , 11 ) = 2 . 122 , p = 0 . 17 , BF = 0 . 44 , Exp . 3: F ( 1 , 11 ) = 0 . 007 , p = 0 . 93 , BF = 0 . 38] . In contrast , in Experiment 2 ( which required an explicit target dimension response ) , RTs were significantly faster when the target feature repeated compared to when it switched within the same dimension , F ( 1 , 11 ) =35 . 535 , p<0 . 001 , ηp2=0 . 764 , BF=13 , and this effect did not differ between the target-defining , color and orientation , dimensions , F ( 1 , 11 ) = 1 . 858 , p = 0 . 2 , BF = 0 . 57 . Note though that , even in Experiment 2 , this feature repetition/switch effect was smaller than the effect of dimension repetition/switch ( 20 vs . 54 ms , t ( 11 ) = 5 . 20 , p<0 . 001 , BF = 122 ) . In summary , the results revealed RTs to be expedited when target presence or absence or , respectively , the target-defining dimension ( on target-present trials ) was repeated on consecutive trials . However , the origin of these inter-trial effects is unclear: The faster RTs for cross-trial repetitions could reflect either more efficient stimulus processing ( e . g . , as a result of greater ‘attentional ‘weight’ being assigned to a repeated target dimension ) or a response bias ( e . g . , an inclination to respond ‘target present’ based on less evidence on repeat trials ) , or both . In the next section , we will address the origin ( s ) of the inter-trial effects by comparing a range of generative computational models and determining which parameters are likely involved in producing these effects . Because feature-specific inter-trial effects , if reliable at all ( they were significant only in Exp . 2 , which required an explicit target dimension response ) , were smaller than the inter-trial effects related to either target presence/absence or the target-defining dimension ( e . g . , in Exp . 3 , a significant dimension-based inter-trial effect of 39 ms compares with a non-significant feature-based effect of 11 ms ) , we chose to ignore the feature-related effect in our modeling attempt . With the full combination of the four factors , there were 144 ( 2 x 2 x 6 x 6 ) models altogether for comparison: non-decision time ( with/without ) , evidence accumulation models ( DDM vs . LATER ) , RDF-based updating ( 6 factor levels ) , and TDD-based updating ( 6 factor levels ) . We fitted all models to individual-participant data across the three experiments , which , with 12 participants per experiment , yielded 5184 fitted models ( see S7 Text for RT distributions and model fits for the factor levels with no updating but with a non-decision time ) . Several data sets could not be fitted with the full memory version of the starting point updating level ( i . e . , Level 2 ) of the dimension-based updating factor , due to the parameter updating to an extreme . We therefore excluded this level from further comparison . To obtain a better picture of the best model predictions , we plotted predicted versus observed RTs in Fig 11 . Each point represents the average RT over all trials from one ratio condition , one trial condition , and one inter-trial condition in a single participant . There are 144 points each for Experiments 1 and 2 ( 12 participants x 3 ratios x 2 trial conditions x 2 inter-trial conditions ) and 108 for Experiment 3 ( 12 participants x 3 trial conditions x 3 inter-trial conditions ) . The predictions were made based on the best model for each experiment , in terms of the average AIC ( see Figs 8 , 9 and 10 ) . The r2 value of the best linear fit is 0 . 85 for Experiment 1 , 0 . 86 for Experiment 2 , and 0 . 98 for Experiment 3 , and 0 . 89 for all the data combined . Fig 12 presents examples of how the starting point ( S0 ) and rate were updated according to the best model ( in AIC terms ) for each experiment . For all experiments , the best model used starting point updating based on the response-defining feature ( Fig 12A , 12C and 12E , left panels ) . In Experiments 1 and 2 , the trial samples shown were taken from blocks with an unequal ratio; so , for the starting point , the updating results are biased towards the ( correct ) response on the most frequent type of trial ( Fig 12A and 12C ) . In Experiment 3 , the ratio was equal; so , while the starting point exhibits a small bias on most trials ( Fig 12E ) , it is equally often biased towards either response . Since , in a block with unequal ratio , the starting point becomes biased towards the most frequent response , the model predicts that the average starting point to boundary separation for each response will be smaller in blocks in which that response is more frequent . This predicts that RTs to a stimulus requiring a particular response should become faster with increasing frequency of that stimulus in the block , which is what we observed in our behavioral data . In addition , since , after each trial , the updating rule moves the starting point towards the boundary associated with the response on that trial , the separation between the starting point and the boundary will be smaller on trials on which the same response was required on the previous trial , compared to a response switch . This predicts faster RTs when the same response is repeated , in line with the pattern in the behavioral data . The forgetting mechanism used in the best models ensures that such inter-trial effects will occur even after a long history of previous updates . In Experiment 1 , the best model did not use any updating of the drift rate , but a different rate was used for each dimension and for target-absent trials ( Fig 12B ) . In Experiment 2 the best model updated the rate based on the ‘Rate with decay’ rule described above . The rate is increased when the target-defining dimension is repeated , and decreased when the dimension switches , across trials , and these changes can build up over repetitions/switches , though with some memory decay ( Fig 12D ) . Since the target dimension was ( also ) the response-defining feature in Experiment 2 , the rate updating would contribute to the ‘response-based’ inter-trial effects . In Experiment 3 , the best model involved the ‘Weighted rate’ rule . Note that the rate tends to be below the baseline level ( dashed lines ) after switching from the other dimension , but grows larger when the same dimension is repeated ( Fig 12F ) . This predicts faster RTs after a dimension repetition compared to a switch , which is what we observed in the behavioral data .
In three experiments , we varied the frequency distribution over the response-defining feature ( RDF ) of the stimulus in a visual pop-out search task , that is , target presence versus target absence ( Experiments 1 and 3 ) or , respectively , the dimension , color versus orientation , along which the target differed from the distractors ( Experiment 2 ) . In both cases , RTs were overall faster to stimuli of that particular response-defining feature that occurred with higher frequency within a given trial block . There were also systematic inter-trial ‘history’ effects: RTs were faster both when the response-defining feature and when the target-defining dimension repeated across trials , compared to when either of these changed . Our results thus replicate previous findings of dimension repetition/switch effects [6 , 9] . In contrast to studies on ‘priming of pop-out’ ( PoP ) [3 , 32–34] , we did not find significant feature-based repetition/switch effects ( consistent with [6] ) , except for Experiment 2 in which the target dimension was also the response-defining feature . The dimension repetition/switch effects that we observed were also not as ‘long-term’ compared to PoP studies , where significant feature ‘priming’ effects emerged from as far as eight trials back from the current trial . There are ( at least ) two differences between the present study and the PoP paradigms , which likely contributed to these differential effect patterns . First , we employed dense search displays ( with a total of 39 items , maximizing local target-to-non-target feature contrast ) , whereas PoP studies typically use much sparser displays ( e . g . , in the ‘prototypical’ design of Maljkovic & Nakayama [3 , 32–34] , 3 widely spaced items: one target and two distractors ) . Second , the features of our distractors remained constant , whereas in PoP studies the search-critical features of the target and the distractors are typically swapped randomly across trials . There is evidence indicating that , in the latter displays , the target is actually not the first item attended on a significant proportion of trials ( according to [35] , on some 20% up to 70% ) , introducing an element of serial scanning especially on feature swap trials on which there is a tendency for attention ( and the eye ) to be deployed to a distractor that happens to have the same ( color ) feature as the target on the previous trial ( for eye movement evidence , see , e . g . , [36 , 37] ) . Given this happens frequently , feature checking would become necessary to ensure that it is the ( odd-one-out ) target item that is attended and responded to , rather than one of the distractors . As a result , feature-specific effects would come to the fore , whereas these would play only a minor role when the target can be reliably found based on strong ( local ) feature contrast [38] . For this reason , we opted to start our modeling work with designs that , at least in our hand , optimize pop-out ( see also [39] ) , focusing on simple target detection and ‘non-compound’ discrimination tasks in the first instance . Another difference is that we used simple detection and ‘non-compound’ discrimination tasks in our experiments , while PoP experiments typically employ ‘compound’ tasks , in which the response-defining feature is independent of the target-defining feature . We do not believe that the latter difference is critical , as reliable dimension repetition/change effects have also been observed with compound-search tasks ( e . g . , [40] ) , even though , in terms of the final RTs , these are weaker compared to simple response tasks because they are subject to complex interactions arising at a post-selective processing stage ( see below and [41 , 42] ) . To better understand the basis of the effects we obtained , we analyzed the shape of the RT distributions , using the modified LATER model [26] and the DDM [21 , 22] . Importantly , in addition to fitting these models to the RT distribution across trials , we systematically compared and contrasted different rules of how two key parameters of the LATER/DDM models–the starting point ( S0 ) or the rate ( r ) of the evidence accumulation process–might be dynamically adapted , or updated , based on trial history . We assumed two aspects of the stimuli to be potentially relevant for updating the evidence accumulation parameters: the response-defining feature ( RDF ) and the target-defining dimension ( TDD; in Experiment 2 , RDF and TDD were identical ) . Thus , in our full factorial model comparison , trial-by-trial updating was based on either the response-defining feature or the target dimension ( factor 1 ) , combined with updating of either the starting point or the rate of evidence accumulation ( factor 2 ) , with a number of different possible updating rules for each of these ( 6 factor levels each ) . An additional factor ( factor 3 ) in our model comparison was the evidence accumulation model used to predict RT distributions: either the DDM or the LATER model . Finally , to compare the DDM and LATER models on as equal terms as possible , we modified the original LATER model by adding a non-decision time component . Thus , the fourth and final factor concerned whether a non-decision time component was used or whether the non-decision time was fixed to zero . Our model assumes that the starting point ( S0 ) is updated based on the observer’s current estimate of the probabilities of the response alternatives , which may depend on trial history . The assumption that the starting point is set based on the prior probabilities of the two alternative responses is consistent with a Bayesian framework of evidence accumulation , in which evidence is accumulated from the starting log prior odds until a threshold level is reached on the posterior odds before a decision is made [19 , 26 , 43] . Our model assumes that the relative frequency of the two alternative values of the RDF ( target-present vs . -absent in Experiments 1 and 3 , color vs . orientation target in Experiment 2 ) is learned from trial history . Since there is always some uncertainty about the frequency , the range of plausible values , given the trial history , is represented by a probability distribution . On the first trial , this distribution is set to a Bernoulli distribution , with a single parameter representing a prior belief about how frequently the two values of the RDF will occur before encountering the first search display . This probability distribution is then updated according to Bayes’ rule on each trial . Note that , on its own , such Bayesian updating would converge on a stable estimate and then not change much–which would predict the size of the inter-trial effects to decrease over the course of an experiment . However , we did not observe such a decrease in any of our experiments ( see S5 Text ) . For this reason , in addition to the Bayesian updating rule described above , we introduced a learning rule based on the Dynamic Belief Model [29] , which assumes there is some fixed probability on each trial that the stimulus frequencies will change and which therefore , in addition to the Bayesian updating , involves a ‘forgetting’ step that serves to reduce the weight of old information relative to the most recent one . This model allows for rapid adaptation to a change even after a long period without any change; and , importantly , it does not predict a decrease of the inter-trial effect magnitude over the course of an experiment . Considering the data from each experiment individually , we found that the best model ( with the lowest AIC ) used updating of the starting point , with partial forgetting ( i . e . , the learning rule from the DBM ) , based on the history of the response-defining feature of the stimulus array . This updating can explain both the effect of RDF frequency on RTs and the response-based inter-trial effects . The updating would result in the starting point being , on average , closer to the threshold associated with the most frequently required response in each trial block , predicting the effect of frequency on RTs . And response-based inter-trial effects arise in the model because , after each trial , the starting point is moved closer to the threshold associated with the response that was required on that trial , reducing the starting point to boundary separation if that response is again required on the next trial . The forgetting mechanism ensures that the magnitude of the starting point shifts , and therefore the predicted inter-trial effects , do not shrink towards zero over the course of the 1000 plus trials in our experiments ( in line with our data , which revealed no evidence of such a shrinkage; see S5 Text ) . Some form of forgetting mechanism is likely to be important for adapting to a changing environment [29] . It might be argued that the frequency effects and response-based inter-trial effects on the mean RTs might , potentially , be equally well explained by trial-to-trial adaptations of the rate of evidence accumulation . However , this would have predicted a different RT distribution , and our model comparison did not favor models in which the rate was updated based on response history . We therefore conclude that the most likely explanation of response-based inter-trial effects is that observers became biased towards the response to which they assigned a higher subjective probability , and that these probabilities were particularly sensitive to what happened on the most recent trials . Of course , our starting point updating model with partial forgetting , which is closely inspired by the Dynamic Belief Model [29] , is only one plausible way in which the learning of response probabilities can be implemented and linked to response biases , and other implementations remain possible . Note also that , in the present study , the feature that was critical for target detection was the same as that determining the response , which did not allow us to dissociate response repetition from target repetition effects . Further work is required to examine for such a disassociation using what is known as a ‘compound’ search task [44] . As to the dimension-based updating factor , in our model comparison , the best models differed among the three experiments . For Experiment 1 , the best model did not include dimension-based updating , most likely because this experiment did not involve random dimension switching ( switching occurred only between the last trial of one mini-block and the first trial of the next block , which were separated by a performance feedback screen ) . In Experiments 2 and 3 , in which random dimension switching did occur within trial blocks , the best models involved updating of the evidence accumulation rate , though with somewhat different updating rules . For both experiments , the best model involved a rule that increased the rate when the target dimension repeated across trials and decreased it when the dimension changed . In Experiment 2 , a partial memory of this increase or , respectively , decrease is then carried over to the next trial , regardless of whether the target on that trial is defined in the same or a different dimension to the preceding trial . We termed this ‘rate with decay’ rule . The best model for Experiment 3 , on the other hand , used an updating rule which assumes that a different rate is associated with each dimension , where , after each trial , the rate for the dimension that defined the target on that trial is increased , and that for the other dimension is decreased by an equivalent amount . This ‘weighted rate’ rule is inspired by the dimension-weighting account [6] , according to which potential target-defining dimensions share the same , limited attentional ‘weight’ resource . The two rules are similar but make significantly different predictions , for instance , when a long sequence of repeats is followed by a switch , or when a long sequence of switches occurs . The ‘rate with decay’ rule predicts the rate to be higher after a sequence of repeats followed by a single switch , compared to a switch following a run of switches–a pattern actually seen in Experiment 2 ( see S6 Text ) . The ‘weighed rate’ rule , by contrast , makes the opposite prediction–consistent with the pattern seen in Experiment 3 ( see S6 Text ) . Recall that , in Experiment 2 , the target dimension was also the response-determining feature . As a consequence , ( repeatedly ) switching the dimension and the response may give rise to a cost that carries over across trials by slowing the ( executive ) act of selecting the appropriate motor response on a given trial . This may be the case because , with choice responses , some ‘event file’ buffering the requisite S–R link might be carried over across trials and affect the speed of response decisions ( see ‘episodic-retrieval theory’ below ) . On switch trials ( ‘S’ ) , the old rule no longer applies , that is , it needs to be inhibited and replaced by a new association , where the mismatch with the old setting slows response selection . On repeated switch trials ( e . g . , ‘SSS’ ) , the link relevant on the current trial ( trial n; the same association as on trial n-2 ) might still be inhibited ( from trial n-1 , on which the rule was found to be inappropriate ) , slowing responses relative to switch trials preceded by repeated trials ( e . g . , ‘RRS’ ) where the association required on trial n is different from trial n-2 and would , thus not be inhibited on trial n-1 . Assuming that the evidence accumulation in favor of a particular target dimension feeds more or less directly into the process of making a response decision , inhibition of an S–R link might narrow the whole ‘pipeline’ of perceptual and response-related evidence accumulation , explaining why the best dimension-based updating rule in Experiment 2 involved updating of the rate . This account of the cost on repeated switch trials would be consistent with the ‘negative priming’ literature ( e . g . , [45] ) . No such cost would arise in Experiment 3 , in which the dimension was not response-defining–rather , all trials with a target present ( in whatever dimension it was defined ) required one and the same , simple target detection response . Accordingly , dimension switches were not associated with a response switch , and so there would be no need for an updating of the S–R association after switch trials ( consistent with evidence that dimensional target identity is not explicitly encoded in simple singleton detection tasks; see [9] ) . In this situation , on the dimension-weighting account , each repetition would mean that increasingly more weight is assigned to the repeated dimension and consequently less weight to the alternative dimension , which will be the target dimension on the switch trial at the end ( RRS ) . Consequently , on that trial , the rate of evidence accumulation ( for a target in the alternative dimension ) is slowed relative to an SSS sequence ( where the dimension on trial n had received a weight increase , rather than a decrease , on trial n-2 ) . Thus , the fact the best model for that experiment involved the ‘weighted rate’ rule would lend support to ‘dimension weighting’ as the best account of dimension repetition/switch effects when there is no concurrent response switching . Importantly , the ‘weighted rate’ and ‘rate with decay’ rules both involve updating of the rate of evidence accumulation ( rather than of the starting point ) . The model comparison thus clearly supports the hypothesis that the dimension repetition benefit derives from more efficient stimulus processing , rather than a response bias . Convergent evidence comes from recent studies of visual search examining event-related potentials , in which dimension-specific RT inter-trial effects were reflected in the latency and amplitude of the early sensory processing N1 [46] and the N2pc component . The N2pc is commonly taken to reflect processes of spatial-attentional selection [41 , 47] . Thus , in light of the present model comparison , the fact that repetitions versus changes of the target-defining dimension across trials shortened the N2pc latencies would support the notion that dimension repetition increases the rate of salience accumulation for attentional target selection . Our model comparison revealed that employing the LATER model for predicting RT distributions did a better job explaining the data than using the DDM . Note , though , that to keep the computational demands at a manageable level , we used a closed-form approximation of the RT distribution predicted by the DDM [48] . This approximation does not capture all features implemented in most computational realizations of the DDM; perhaps critically , it does not allow for trial-to-trial variability of the non-decision time . Applied to the present data , a DDM implementation with added trial-to-trial variability of the non-decision time might have significantly improved the performance of this model ( whereas it would likely have made less of a difference to the LATER model ) –thus reducing the difference in AIC between the LATER model and the DDM . Adding trial-to-trial variability of the non-decision time to the future model implementations may also be important theoretically , as it may be possible to explain some of this variability by adding updating rules that operate on the non-decision time . Critically though , for all the other factors in our model comparison , the best-performing levels turned out the same , whether the DDM or the LATER model was used . Note that , while we tested a large number of possible models , there potentially are other models that might perform even better . In particular , a model that allows several updating rules to operate at once would likely perform somewhat better than our winning model . In the present study , we limited our comparisons to parsimonious models with one updating rule based on the RDF and one based on the TDD , assuming that manipulation of the RDF or the TDD only affects one distinctive process that is reflected in either the starting point S0 or the accumulation rate r . However , it remains possible that the RDF and/or the TDD influence RTs through more than one mechanism in parallel–in which case our model comparison would have identified only that mechanism which accounts for the largest portion of the inter-trial effects . In future work , it will be interesting to determine whether a model which permits the RDF and/or the TDD to operate through more than one mechanism can explain the data significantly better . In our model , we treated target-absent trials similar to target-present trials , given that pop-out targets are detected efficiently ( based on spatially parallel search ) , that is: with pop-out targets , a target-presence versus -absence decision can be made by setting a single threshold on the search-guiding overall-saliency map [49] . Indeed , our model predicts RTs well on both target-absent and target-present trials . However , deciding that a target is absent in a non-pop-out search task may be quite different . In a non-pop-out search display , every item in the search display would in principle need to be processed to ( reliably ) arrive at a correct target-absent decision [50] , though some process terminating the search ( and triggering a target-absent decision ) prior to exhaustive scanning of all display items may also be involved [16 , 51] . In any case , to model non-pop-out search , a more complex model may be required in which multiple stages of evidence accumulation typically occur before a response is triggered , corresponding to checking individual items to determine whether or not they are the target . While we examined a number of different updating rules in our model comparison , we are not suggesting that these covered all possibilities; that is , we cannot rule out that there may be updating rules that would perform even better . While our winning model was based on the Dynamic Belief Model [29] , a very similar model has been proposed by Anderson and Carpenter [30] , which also involves a combination of Bayesian updating and forgetting of old trials , and this could have served as an equally good starting point for our model . Another , similar model was proposed by Mozer et al . [52] . Unlike the present model , this does not involve a hyperprior on the stimulus category probability with Bayes’ rule; rather , it updates the probability more directly , using a weighted-averaging rule , with the weight assigned to older trials decaying exponentially . This rule is close to the forgetting rule of the Dynamic Belief Model . Mozer et al . [52] showed that their model can qualitatively reproduce the pattern of results from a number of ‘priming of pop-out’ and visual search experiments [3 , 4 , 52 , 53] . Different to our model , the model of Mozer et al . learns conditional probabilities , which they argued was essential for explaining interactions between the inter-trial effects for different features of the stimuli in some of the experiments they modeled . While learning of conditional probabilities was not necessary to explain the results from the three experiments reported here , any more complete model of inter-trial effects in visual search may well need to incorporate conditional probabilities to provide a truly general account . Another noteworthy difference to our model is that the model of Mozer et al . only included the learning of probabilities without specifying how these learned probabilities influence the perceptual decision process . Consequently , they could not make quantitative predictions about RTs and their distributions . In contrast , our model makes quantitative predictions because it combines a Bayesian updating rule with a model of the perceptual decision process ( either DDM or LATER ) . Another modeling framework that has previously been applied to explaining inter-trial effects in visual search is the ‘Theory of Visual Attention’ ( TVA ) [54] . TVA models the rate at which visual categorizations of the type “object x has feature i” are made and encoded into visual short-term memory ( mediating overt responses ) . Each visual object receives an attentional weight , which is the product of the strength of the sensory evidence that the object belongs to category i and the current importance of attending to category i , referred to as the ‘pertinence’ of the category , summed over all relevant visual categories ( i . e . , categories for which there is sensory evidence ) . The scaling factors in our dimension-weighted rate updating rule , representing the current weight or importance assigned to each dimension , play a similar role to the pertinence values in TVA . Ásgeirsson et al . [55 , 56] have shown that color priming effects in visual search can be well explained by TVA , by assuming that the pertinence of a given feature increases or decreases when the target or , respectively , a distractor possesses that feature . Similarly , our dimension-weighted rate rule assumes that the scaling factor increases for a given dimension when the target is defined in this dimension , and decreases when the target is defined in a different dimension . Our finding that this was the best rule for explaining performance in Experiment 3 is thus broadly consistent with the TVA-based model proposed by Ásgeirsson et al . [48 , 49] . However , our model also differs from theirs in a number of respects . First , in our model , the scaling factors were associated with dimensions rather than individual features ( recall that , in our paradigms , feature-specific inter-trial effects are relatively unsubstantial compared to dimension-specific effects; see also [6] ) . Second , the model of Ásgeirsson et al . only considered effects from a single trial back , while our dimension-weighted rate rule can model longer-term effects ( of course , it would be possible to combine TVA with a similar rule to take longer-term inter-trial history into account ) . Third , unlike the model of Ásgeirsson et al . , our model did not include ‘spatial weights’ associated with potential target locations . Ásgeirsson et al . showed that their TVA-based model performed better when taking spatial weighting into account . Note , though , that spatial weighting is likely to be more important with sparse displays and a limited set of locations ( six in Ásgeirsson et al . ) , compared to the dense displays used in our experiments [57] . Finally , by modelling full RT distributions , we could make a distinction between two different ways in which the speed of a perceptual decision could be increased: by increasing the rate at which relevant sensory evidence accumulates or by decreasing the amount of evidence required to make a decision ( through a shift of the starting point ) . TVA does not make any equivalent distinction . Another framework for understanding inter-trial effects in visual search is offered by the episodic-retrieval account [14 , 58]–though the evidence for this account derives exclusively from compound-search tasks not investigated here . Huang et al . [4] argued that repetition effects in visual search are well explained by episodic-retrieval theory , based on the finding that repetition of a task-irrelevant feature ( in their experiments: color ) speeded search only when the target-defining feature ( size ) was also repeated ( participants had to respond to the orientation of a size-defined target , irrespective of the target color ) . When the target-defining feature changed , RTs were slower if the task-irrelevant feature was repeated . The episodic-retrieval account can explain this pattern by assuming that participants retrieve an episodic memory trace of the target from the previous trial , which influences a post-selective process of verifying whether a candidate target is the actual target . If the retrieved memory trace completely matches the target on the current trial , the decision will be fast; by contrast , a partial match ( i . e . , a target of the same size but a different color ) gives rise to ‘inconsistency’ and may thus be slower to process than a complete mismatch , explaining the interaction between repetition of target-defining and task-irrelevant features in the study of Huang et al . [4] . A similar result was reported by Töllner et al . [46] , though for two task-relevant target attributes . They observed a partial-repetition cost when the response-defining feature ( target orientation ) changed across trials while the target-defining dimension ( color or shape ) was repeated . However , the latency of the N2pc was affected only by repetition/switch of the target-defining dimension , independently of whether the response-defining feature repeated/changed–leading Töllner et al . to conclude that at least one critical component of the target repetition/switch effect arises at a ( pre-attentive ) stage of saliency coding , leading up to target selection . The partial-repetition effect , by contrast , arises at a post-selective stage where the response-defining target feature is analyzed and a response decision is determined . This process is modulated by ‘linked expectancies’ between the dimension and the response: when the dimension is repeated , the system expects the response to be repeated as well , yielding a cost when the response actually changes . –Our best-fitting model , while predicting a RT cost when the dimension or the response changes ( compared to when both are repeated ) , does not predict a larger cost when either one or the other changes , compared to when both change ( instead , the dimension and response change costs would be additive ) . To account for such partial-repetition cost effects , further modeling work is required based on RT performance in simple-detection and compound-search tasks that make the same demands with regard to target selection , but different demands with regard to response selection ( i . e . , simple detection of a target-defining attribute vs . discrimination of a separate , response-defining feature ) , as well as RT performance in a non-search task that makes no demands on target selection , but similar demands to compound search on response selection ( along the lines of [12] ) . RTs could then be modeled , for instance , as a series of two diffusion processes ( one for target selection and one for response selection ) , where parameters of the second process ( r , θ , or S0 ) might be set conditional upon repetition/switch of the target-defining attribute . Such a model might then also be able to account for partial-repetition costs attributable to completely ( detection- and response- ) irrelevant target attributes [4] , over and above those caused by relevant features [46] , perhaps by making updating based on irrelevant features conditional on relevant features [52] . In conclusion , we found that RTs in pop-out visual search are faster when the response required on a given trial occurred frequently in the recent past , and particularly when the same response is repeated from the previous trial . By performing a factorial model comparison , we showed that these effects are best explained by updating of the starting point of an evidence accumulation process , that is , they reflect a bias towards a response that is more likely to occur , given the recent history . We also found that reaction times are faster when the target-defining dimension is repeated , even when this is unrelated to the response . Our model comparison showed that this effect is best explained by trial-to-trial updating of the evidence accumulation rate . This suggests that dimension repetition/switch effects do not reflect a response bias , but rather reflect more efficient processing when the same dimension is repeated .
All participants gave informed consent prior to the experiment . The study was approved by the LMU Department of Psychology Ethics Committee and conformed to the Helsinki Declaration and Guidelines . Bayesian ANOVA and associated post-hoc tests were performed using JASP 0 . 86 ( http://www . jasp-stats . org ) with default settings . All Bayes factors for main effects and interactions in the ANOVA are ‘inclusion’ Bayes factors calculated across matched models . Inclusion Bayes factors compare models with a particular predictor to models that exclude that predictor . That is , they indicate the amount of change from prior inclusion odds ( i . e . , the ratio between the total prior probability for models including a predictor and the prior probability for models that do not include it ) to posterior inclusion odds . We used inclusion Bayes factors calculated across matched models meaning that models that contain higher order interactions involving the predictor of interest were excluded from the set of models on which the total prior and posterior odds were based . Inclusion Bayes factors provide a measure of the extent to which the data support inclusion of a factor in the model . Bayesian t-tests were performed using the ttestBF function of the R package ‘BayesFactor’ with the default setting ( rscale =“medium” ) . To find the model that best explained our data , we performed a factorial model comparison . Full descriptions of the four factors and their levels are given in the modelling section . Here we describe the general procedure used for the model fitting , which was the same for all models . Each model consisted of an evidence accumulation model: either the LATER model or the DDM , and two updating rules , each of which specified how one aspect of stimulus history should affect the trial to trial change of a parameter of the evidence accumulation model . There was one such updating rule for the response defining feature and one for the target defining dimension , and in each case one of the factor levels specified that no updating at all should take place . For the DDM , we used a closed-form approximation [48] , adding a scaling parameter that determined the size of the random component of the drift diffusion model . This was necessary since our rule for updating the starting point made the scale non-arbitrary . Models were fitted using maximum likelihood , using the R function ‘constrOptim’ to find minimum value of the negative log likelihood . Error trials and outliers were excluded from the calculation of the likelihood , but were included when implementing the updating rules . Outliers were defined as trials with reaction times more than 1 . 5 interquartile ranges below the mean or longer than 2 seconds . To make sure we found the best possible fit for each combination of factor levels , we used an inner and an outer optimization process . The inner optimization process was run for each combination of parameters that was tested by the outer optimization process , to find the best possible values of the inner parameters for those values of the outer parameters . The inner parameters were the parameters of the evidence accumulation model itself , except for the non-decision time which was an outer parameter ( because one level of one of the factors specified that the non-decision time should be fixed to zero ) . For the LATER model , the inner parameters were the starting point boundary separation , and the mean and standard deviation of the distribution for the rate . For the DDM , the inner parameters were the starting point boundary separation , the rate , and the scaling parameter . These parameters could differ between target absent trials , as well as between the two different target dimensions , meaning that there were nine inner parameters for Experiments 1 and 3 and six for Experiment 2 ( where there were no target absent trials ) . The outer parameters were the non-decision time ( when this wasn't fixed to zero ) , and 0 to 2 parameters for each updating rule ( see the modelling section for details ) . This means that models could have 0 to 5 outer parameters in total depending on the factor levels .
|
When a perceptual task is performed repeatedly , performance becomes faster and more accurate when there is little or no change of critical stimulus attributes across consecutive trials . This phenomenon has been explored in previous studies on visual ‘pop-out’ search , showing that participants can find and respond to a unique target object among distractors faster when properties of the target are repeated across trials . However , the processes that underlie these inter-trial effects are still not clearly understood . Here , we approached this question by performing three visual search experiments and applying mathematical modeling to the data . We combined models of perceptual decision making with Bayesian updating rules for the parameters of the decision making models , to capture the processing of visual information on each individual trial as well as possible mechanisms through which an influence can be carried forward from previous trials . A systematic comparison of how well different combinations of models explain the data revealed the best model to assume that perceptual decisions are biased based on the response-critical stimulus property on recent trials , while repetition of the visual dimension in which the target differs from the distractors ( e . g . , color or orientation ) increases the speed of stimulus processing .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"learning",
"decision",
"making",
"reaction",
"time",
"social",
"sciences",
"neuroscience",
"learning",
"and",
"memory",
"cognitive",
"neuroscience",
"cognitive",
"psychology",
"mathematics",
"probability",
"distribution",
"computer",
"vision",
"cognition",
"memory",
"vision",
"computer",
"and",
"information",
"sciences",
"target",
"detection",
"probability",
"theory",
"psychology",
"biology",
"and",
"life",
"sciences",
"sensory",
"perception",
"physical",
"sciences",
"cognitive",
"science"
] |
2018
|
Inter-trial effects in visual pop-out search: Factorial comparison of Bayesian updating models
|
The self-fertile nematode worms Caenorhabditis elegans , C . briggsae , and C . tropicalis evolved independently from outcrossing male-female ancestors and have genomes 20-40% smaller than closely related outcrossing relatives . This pattern of smaller genomes for selfing species and larger genomes for closely related outcrossing species is also seen in plants . We use comparative genomics , including the first high quality genome assembly for an outcrossing member of the genus ( C . remanei ) to test several hypotheses for the evolution of genome reduction under a change in mating system . Unlike plants , it does not appear that reductions in the number of repetitive elements , such as transposable elements , are an important contributor to the change in genome size . Instead , all functional genomic categories are lost in approximately equal proportions . Theory predicts that self-fertilization should equalize the effective population size , as well as the resulting effects of genetic drift , between the X chromosome and autosomes . Contrary to this , we find that the self-fertile C . briggsae and C . elegans have larger intergenic spaces and larger protein-coding genes on the X chromosome when compared to autosomes , while C . remanei actually has smaller introns on the X chromosome than either self-reproducing species . Rather than being driven by mutational biases and/or genetic drift caused by a reduction in effective population size under self reproduction , changes in genome size in this group of nematodes appear to be caused by genome-wide patterns of gene loss , most likely generated by genomic adaptation to self reproduction per se .
Self reproduction increases the probability of homozygosity at single loci , reducing the effective size of the population by a factor of two [1–3] . At the level of whole genomes , the reduced probability that two loci will be heterozygous within a single individual greatly decreases the efficacy of recombination , thereby increasing linkage disequilibrium within the population [4] . Increased homozygosity can expose heretofore recessive mutations to natural selection , potentially accelerating the rate of evolution at those loci , while increased linkage disequilbrium makes self-reproducing ( “selfing” ) species more susceptible to selective sweeps and/or background selection generated by advantageous and deleterious mutations [5] . What then should be the genomic consequences of a transition in mating system from outcrossing to selfing ? The reduction in population size means that genetic drift should become more prominent , allowing for the accumulation of slightly deleterious features , such as repetitive elements , which should lead to an increase in genome size [6] . Similarly , increased linkage disequilibrium means that new advantageous mutations are more likely to bring along deleterious elements via hitchhiking as they increase in frequency [7] . Alternatively , any systematic mutational bias in the direction of DNA deletion would have a greater chance of succeeding within selfing species if such deletions are mildly deleterious [8] , and even more so if the transition to selfing means that certain biological functions related to outcrossing ( such as mate finding ) are no longer needed . Further , increased linkage within selfing lineages increases the probability of co-inheritance of the host genome and selfish genetic elements such as transposable elements , which should lead to an increase in the efficacy of selection against the selfish elements and therefore a reduction in genome size if such elements are a significant fraction of the original ancestral genome [9] . Finally , in species with sex chromosomes , selfing has the potential to equalize the effective population size of sex chromosomes , which tend to have an Ne that is 3/4 as large as autosomes because of the reduced chromosome count in the heterogametic sex [1] . Although this is not usually an issue in plants , for which there are few species with sex chromosomes [10] , in animals this change in the ratio of effective population size , as well as other sex-chromosome specific effects such as the lack of dominance in the heterogametic sex , could influence the rate of molecular evolution on sex chromosomes [11] following a transition to selfing . Thus , shifts in mating systems could potentially lead to either increases or decreases in genome size depending on the functional role and genomic context of a given segment of DNA . Determining what actually occurs in nature therefore depends both on a well-documented evolutionary transition in mating system and a set of well-annotated genomes that allow genetic function to be appropriately classified and compared . Nematodes in the genus Caenorhabditis have made the transition from outcrossing to selfing three separate times independently [12 , 13] . The well known model system C . elegans , as well as the species C . briggsae and C . tropicalis , reproduce primarily through self fertile hermaphrodites , which are essentially sperm-producing females derived from male-female ( gonochoristic ) ancestors . Males capable of mating with the hermaphrodites are also present at low frequencies within these species , but importantly , hermaphrodites are incapable of mating with each other . The genome sizes in Caenorhabditis nematodes are smaller by 20–40% for self-fertile hermaphrodites ( C . elegans , 100 . 4Mb; C . briggsae , 108Mb; C . tropicalis , 79Mb ) than the flow-cytometry estimated genome sizes of the larger Ne outcrossers ( C . remanei , 131Mb; C . brenneri , 135Mb; C . japonica , 135Mb ) [14–16] . A similar pattern of genome reduction has been observed in multiple self-reproducing plant species as well [9] , which raises the possibility that genome size reduction may be a general syndrome associated with the transition to self reproduction . Here , we combine existing genome assemblies with new functional annotation for each of these species in order to examine common features of genomic evolution that are shared across the three transitions in mating system from outcrossing to selfing within this genus . However , outcrossing nematode genome assemblies remain problematic because of remarkably high levels of DNA sequence variation in these species , including some of the most polymorphic animals currently known [17 , 18] . This extreme polymorphism presents a particularly unique problem for genome assembly because the worm’s small size precludes sequencing a single animal , and so DNA must be extracted from populations that tend to remain highly polymorphic despite laboratory inbreeding [19] . Thus , in order to address questions about changes in finer-scale genomic structure under the transition to selfing , we also present a high-quality draft genome sequence for the outcrossing C . remanei and use this sequence to test theoretical predictions regarding the influence of self-fertilization on genome evolution . The combination of general comparisons across the phylogeny with specific comparisons between C . elegans , C . briggsae , and C . remanei shows that , while these nematodes share the general pattern of genome size reduction with plants , they appear to achieve it in different ways and that the particulars of the changes likely result from an interaction between genomic architecture and changes in population size and the frequency of interactions between the sexes .
We analyzed the genome content of all Caenorhabditis members of the Elegans supergroup with genome sequences available on Wormbase [20]: the self-fertile hermaphrodites C . briggsae , C . elegans , and C . tropicalis and the outcrossing C . remanei , C . japonica , C . brenneri , and C . sinica ( formerly C . sp . 5 [21] ) . In addition , we analyzed the outcrossing C . angaria as an outgroup [22] . We performed each analysis on both the extant C . remanei assembly and our new de novo assembly presented below . The results presented here are based on the de novo assembly , with the analysis of previously assembled ( highly polymorphic ) C . remanei genome sequence presented in the Supplemental Materials ( S1 Table; S1 Fig ) . For each genome used here , we also generated a de novo functional annotation for each species using the same pipeline so as to minimize annotation bias from influencing the results ( S2 Table ) . Overall , consistent with previous genome size estimates based on flow cytometry [14 , 15] , self-fertile species within this group have substantially smaller genomes than their most closely related outcrossing relatives ( Fig 1 , Table 1 , S2 Table ) . Taking the average genome size of the outcrossing species to be 130Mb , the genomes of C . elegans , C . briggsae and C . tropicalis have been reduced by 23% , 17% and 39% respectively via the transition to selfing . These are likely to be upper bounds on the actual size reduction because of likely over-assembly of most of the outcrossing species ( see below ) . Although the branch lengths between each is actually fairly long ( on the order of tens of millions of years ) , the actual time of the shift to selfing within any given lineage is likely to not have been more than ~4 million years based on the rate of evolution of codon usage bias [23] . To a first order , the overall pattern of genome shrinkage is roughly proportional when summed across broad functional categories , including the total size of exons , introns , intergenic regions , and repetitive elements within each genome ( Fig 1; Table 1 ) . Importantly , it does not appear that genome shrinkage is dominated by changes within a single functional class , such as repetitive elements . Thus , this comparative analysis suggests that genomic change following the transition to selfing is generated by a general reduction in genome size across coding and non-coding regions . A more precise analysis requires a careful comparison within each functional category , which in turn requires more complete genome assemblies for outcrossing species than are currently available within this group . To this end , we generated a de novo high quality assembly of C . remanei before completing the remainder of the tests . To create a reliable and well-assembled genome sequence for C . remanei , we first aimed to remove residual polymorphism from extant laboratory strains . We used a novel breeding scheme to create nearly isogenic strains for deep coverage genome sequencing , genetic mapping , and high-quality genome assembly . Specifically , we performed sequential inbreeding and selection over 50 generations to purge deleterious mutations and create 2 highly inbred C . remanei lines ( New York PX356 and Ohio PX439 ) . We then assembled a de novo draft genome sequence for PX356 from ~560x coverage of paired-end shotgun sequence , and ~75x coverage of 3 sizes of mate pair libraries . We used sequenced mRNA extracts from a mixed-stage population of C . remanei to annotate protein-coding genes . We estimated residual polymorphism in our inbred PX356 strain to occur at just ~0 . 01% of sites in well-assembled genic regions . In comparison , analyses of the previously assembled draft C . remanei genome [19] found allelic dimorphism for 4 . 7% of defined C . elegans orthologous genes and a sizable portion of DNA aligning to the C . elegans Chromosome IV ( ~10% of the total genome ) . Unfortunately , cryptic reproductive incompatibilities between PX356 and PX439 led to significant segregation distortion for several linkage groups in our genetic map . Overall , our assembly appears to provide good coverage for linkage groups orthologous to C . elegans Chromosomes II , IV , and X , with more fragmented coverage of Chromosomes I , III , and V ( Fig 2 ) . Nevertheless , gene assemblies within these large fragments are excellent . This therefore represents the first well-assembled genome from a highly polymorphic outcrossing species from this group . When looking at the evolution of genome structure we concentrate on the subset of well-assembled chromosomes , while when looking at the evolution of gene structure , we include the entire genome assembly . We first tested the hypothesis that change in genome size in selfers is driven by a reduction in transposable element ( TE ) abundance [24] . TEs vary widely in structure , mobility , distribution , and diversity [25] and , depending on the dynamics of these factors and population size , transposons are predicted to increase genome size in selfers relative to outcrossers [26] or outcrossers relative to selfers [27] . Within this group of nematodes , the C . briggsae genome is 8Mb larger than the C . elegans genome largely owing to repeat content [28] , with roughly 8% of the C . briggsae genome composed of Tc1-IS630-Pogo DNA transposons [29] . We found that the repeat content of C . remanei is intermediate between C . elegans and C . briggsae ( Fig 1; Table 1; S2 Table ) , and therefore expansion of repetitive DNA within C . remanei can not explain genome size differences . The self-fertile C . tropicalis has the smallest genome ( 79Mb ) of the Elegans supergroup , as well as the smallest repeat content . However , the self-fertile C . briggsae has a repeat content larger than any of the outcrossing Caenorhabditis with the exception of C . japonica . We therefore find no evidence that repeat expansion and/or shrinkage explains the majority of genome size differences between outcrossing and self-fertile Caenorhabditis , although it is a minor contributing factor for C . elegans and C . tropicalis . This is in stark contrast to plants , in which it appears that reduction in TE content is one of the major factors driving the evolution of smaller genome size within the self-compatible species [30] . Accumulated biases in insertion or deletion mutations may grow or shrink genomes across different scales . For example , Hu et al . [30] examined the basis for genome size differences between the self-fertile plant Arabidopsis thaliana ( with a 125Mb genome ) and the outcrossing A . lyrata ( with a 207Mb genome ) , and found hundreds of thousands of small deletions in the self-fertile A . thaliana . Alternatively , insertions and deletions could occur at the level of individual genes; A . thaliana has 17% fewer genes than A . lyrata[30] . Thomas et al . [31] found that selfing Caenorhabditis have smaller transcriptomes than related outcrossers and a specific reduction in expression of genes that show sex-biased expression in outcrossing Caenorhabditis . Large-scale rearrangements may also account for genome size differences , and comparison between the A . thaliana-A . lyrata genomes discovered 3 large deletions in the selfing species [30] . Biases in the distribution of indels can be driven by selection or via neutral processes . Rapid growth and reproduction may favor small genomes [32] , and self-fertile organisms with these life cycles [33] may experience selection for DNA loss . Parasites are expected to be more frequent in outcrossing populations than selfing , and theoretical studies indicate that parasites may select against gene loss in hosts [34] . Alternatively , neutral differences in mutational processes may result in genome size differences [35 , 36] . Size transmission bias , whereby the XX hermaphrodites tend to inherit chromosomes shortened by deletions and XO males tend to inherit longer chromosomes with transgenic insertions , has been reported in C . elegans[37] . Although the mechanism through which this occurs is not yet known , simulations indicate that the androdioecious mating system of self-fertilizing Caenorhabditis , with males contributing few offspring to most generations , could rapidly lead to reduced genome size via this mechanism [37] . First , we explored the role of large-scale insertions and deletions contributing to genome size differences within C . elegans , C . briggsae , and C . remanei , as this analysis requires well-assembled chromosomes . These species have diverged from a common ancestor at least 30 million years ago [38] , which means that few sequences other than those within conserved coding regions can be aligned between them . In order to align large genomic regions , we identified 15 , 699 orthologous genes among the three species , finding that >90% of these orthologous genes are found on the same chromosome in all three ( Fig 2 ) . The X chromosome in particular retains a striking conservation of synteny , with the exception of an apparent C . remanei-specific ~3 . 6Mb region of divergence ( Fig 2C ) . This latter portion of the C . remanei X chromosome contains 786 annotated genes , but only 48 of these are orthologous to genes in C . elegans ( 47 to genes in C . briggsae ) , and of these , only 17 orthologous genes are found on the C . elegans X ( the same 17 genes have orthologous counterparts on the C . briggsae X ) . Fifty of the genes in this region are orthologous to genes in C . brenneri but in the absence of a C . brenneri genetic map we do not know if these are located on the X chromosome . Genes within a highly divergent syntenic region may retain similar biological functions despite a lack of clear orthologous counterparts . We used Interproscan v5 . 3 to assign putative biological functions to protein domains for the clusters of genes with no apparent syntenic relationships across species . For example , roughly 7% of the genes in the C . elegans genome are seven transmembrane G protein-coupled serpentine receptors ( 7TM GPCRs ) [39] . These chemosensory genes are responsible for recognition of food , environment , and other animals , and thought to be found in large numbers in Caenorhabditis because of the soil/decaying-plant-dwelling nematodes need to respond to environmental cues [39] . There are 19 families of serpentine receptors in C . elegans . The majority of members within a family occur as clusters on chromosome arms , thought to be the result of local gene duplication events [40] . The C . remanei genome contains 59 serpentine receptor class g ( srg ) genes . Six srg genes ( 10 . 17% of the total ) are located in this region of the X chromosome ( 2 . 74% of the total genome ) , and the genes surrounding each of these srg genes show no functional or sequence similarity to the genes surrounding the 72 C . elegans srg genes or the 60 C . briggsae srg genes . Thus , this region of the X appears to largely hold C . remanei-specific genes and to not be translocated relative to other chromosomes in other species . We next tested the accumulated insertion/deletion hypothesis for genome size change by focusing on indel size biases in smaller blocks of aligned sequence . Consistent with the larger-scale comparisons , we found no difference in indel size bias among species when we analyzed individual syntenic blocks of DNA between species pairs ( S2 Fig ) . In particular , there is no evidence that these aligned regions tend to be systematically smaller in C . elegans and C . briggsae relative to C . remanei . Although there are a few specific differences among species , overall we do not find evidence that either large-scale rearrangements or small-scale indels are an important contributor to genome size differences among these species . Our de novo C . remanei genome assembly predicts a complement of 25 , 415 protein-coding genes , compared to 20 , 532 in C . elegans and 21 , 936 in C . briggsae ( Table 2 ) . The genome-wide unspliced transcript footprint in C . remanei , comprising the combination of exons , introns and untranslated regions ( UTR’s ) , comprised ~69 . 33Mb ( 58 . 51% of the assembled genome; 52 . 92% of the estimated genome size ) , which is 18 . 83% larger than the equivalent footprint in C . briggsae ( 58 . 09Mb [28] , 53 . 79% of the assembled genome ) , 18 . 74% larger than C . elegans ( 58 . 39Mb [20] , 58 . 16% of the assembled genome ) . Consequently , the transcribed genic footprint explains the difference in assembled genome size between C . briggsae and C . remanei and ~60% of the difference in assembled genome size between C . elegans and C . remanei . Similarly , the outcrossing C . brenneri has 30 , 667 protein-coding genes and an unspliced transcript footprint of 70 . 5Mb ( Fig 1 ) , although these are likely overestimates due to allelism in the assembly [19] . Consistent with this pattern , we predict 22 , 326 coding genes within the self-fertile C . tropicalis ( see Table 1 for all species ) . These results are consistent with an analysis of gene content within these species assessed by whole-genome transcriptional analysis [31] . Thus , while there is no evidence for differences in repeat content or indel biases among these species , there is strong evidence that genome size differences result from differential protein-coding gene content in self-fertile and outcrossing Caenorhabditis . Intergenic distances vary widely within Caenorhabditis genomes , with some genes located in co-transcribed operons that are separated by a few nucleotides and other genes separated by many kilobases of sequence . Autosomal intergenic spacing for C . remanei exceeded that of both C . briggsae and C . elegans , despite these being lower bound values for C . remanei because of the potential for unincluded , unassembled regions probably underestimate C . remanei intergenic distances ( Fig 3; Table 2 ) . Across the entire genome ( including scaffolds not included in linkage groups ) the total intergenic content of C . remanei was 0 . 79Mb larger than that of C . briggsae and 10 . 73Mb larger than that of C . elegans . In an outcrossing population with a 50:50 sex ratio there are 3/4 the number of X chromosomes in a population for any given autosome . The effective population size of the X is also reduced by variance in male mating success , and under the effective population hypothesis these forces should result in increased genetic drift and the proliferation of weakly deleterious elements [6] . In contrast , the effective population size of the X chromosome should be equivalent to the autosomes in self-fertile organisms [41] . However , we found no difference in the relative ratio of intergenic regions in the selfing versus outcrossing species ( Fig 3 ) . Presumably either the evolutionary forces that are responsible for maintaining variation in the size of intergenic spaces are not sensitive to changes in effective population size , the timescale since the advent of selfing has not been long enough for the genomic features of the sex chromosome and autosomes to equilibrate in the selfers [23] , or , as we discuss below , there are other genetic differences between these chromosomes that drive this pattern . In order to analyze gene structure , we identified a set of co-orthologs conserved as pairs between C . remanei , C . briggsae and C . elegans ( Fig 4 ) . We found that among our co-ortholog coding sequences the average number ( 6 ) and length ( ~200bp ) of exons per gene was similar for the three species . The total length of gene transcripts was longer on autosomes in C . remanei than in C . briggsae or C . elegans but smaller on the X chromosome in C . remanei than in C . briggsae or C . elegans ( Fig 4 , Table 2 ) . The protein sequences were not significantly different between the X chromosome and autosomes in any of the three species and among the species only C . briggsae had protein sizes that differed significantly from the other species . Thus there is an interesting interaction between chromosomal gene structure and mating system , with intron size expanding on the X with the advent of selfing . Although this could reflect an accumulation of slightly deleterious DNA in the selfing species because of a decrease in effective population size [42] , we would expect the same expansion to occur within autosomes . Another alternative is that increased selection on male function on males within C . remanei in turn drives stronger X-specific genomic evolution , although this seems unlikely given the fact that it appears that the X is actually enriched for genes affecting female/hermaphroditic function as opposed to male function [43] . Given that these explanations based purely on population genetics do not appear to fit the data , another explanation based on genetic differences between the chromosomes seems more likely . In C . elegans , it is known that genetic map is much more uniform on the X chromosome than on the autosomes ( which tend to have very little recombination toward their centers ) [44] . Indeed the molecular machinery that generates chromosome pairing and crossing over is different for the X than the autosomes [45] . Because the effective recombination rate is lower in selfers than in outcrossers , the difference in intron size between the X and autosomes within C . elegans and C . briggsae may reflect differential sensitivity to changes in effective recombination across chromosomes , coupled with the fact that the X and autosomes now experience similar effective population sizes under selfing . There may therefore be a complex interaction between recombination , drift and selection on the X that is driving this unusual pattern . Distinguishing among hypotheses will require a more careful analysis of the pattern of selection operating on the X and autosomes . Caenorhabditis genomes have large numbers of nematode-specific and species-specific proteins [46] , and high divergence makes it difficult to conclusively identify individual genes that are present in outcrossing Caenorhabditis but lost in the selfers . To accommodate this , we characterized functional divergence between self-fertile and outcrossing Caenorhabditis by analyzing putative protein domains in the genomes of Caenorhabditis and the distantly related P . pacificus ( Fig 5 ) . We found no functional groups that were significantly enriched in the outcrossing Caenorhabditis relative to the selfing Caenorhabditis . There are numerous species-specific differences , however . For example , we identified between 191 and 1 , 721 proteins with F-box domains ( IPR001810 ) in C . briggsae , C . sinica , C . remanei , C . tropicalis , C . brenneri and C . elegans , 5–18 times as many as identified in C . japonica , C . angaria and P . pacificus . Domains found in large numbers in members of the Elegans supergroup encompass functionally diverse proteins ( i . e . , protein kinase domains , C-type lectins , and zinc fingers ) , and proteins known to be important in Caenorhabditis , including the 7TM GPCRs introduced earlier , F-box domains , and the Caenorhabditis-specific Domain of Unknown Function DUF38 ( Fig 5A ) . These proteins are clearly fundamental for Caenorhabditis biology , but for the most part there is little functional information known about these rapidly evolving protein families . To more closely examine a particularly hyper-diverse gene family , we identified 1 , 499 serpentine receptor genes in C . elegans , 1 , 125 in C . briggsae and 1 , 026 in C . remanei and the relative distribution among the families is similar across the Elegans supergroup ( Fig 5B ) . Despite these large numbers there is functional evidence , direct or inferred , for only ~10 GPCRs [47] . Further analysis of similarities across whole molecular pathways are provided in the Supplementary Material ( S1 Text , S3 Fig ) . We did find a small number of genes that are present in outcrossing species but lost in within selfing species , although no obvious classes of genes reveal themselves as being specifically lost within these groups ( S1 Text ) . Overall , then , broad genomic comparisons do not reveal any systematic gain or loss of functional categories within and between mating system types , although each individual genome can show dramatic differences within any given gene family . Within-species polymorphism across Caenorhabditis varies by several orders of magnitude , with selfing species being relatively depauperate of variation [16] and outcrossing species being among the most polymorphic animals yet observed [18] . At least one likely reason for these dramatic differences in polymorphism are differences in effective population size among species , which has been estimated as <10 , 000 in C . elegans[48] , <60 , 000 in C . briggsae[49] and >1 , 000 , 000 in C . remanei[17] . We would therefore expect that deleterious elements would be more likely to accumulate and expand the genomes of self-reproducing species because of their small population sizes . Instead , we find just the opposite . Genomes are smaller in selfers and intergenic spaces and protein-coding genes are larger on the X chromosome than autosomes . In the outcrossing C . remanei protein-coding genes are smaller on the X chromosome despite the reduced effective population size of the X versus the autosomes in male-female species . The transition from outcrossing to self-fertilization is common in plants , and plant genome size is similarly positively correlated with outcrossing [9] , although this relationship is weakened when corrected for phylogenetic relatedness [50] . It is possible that changes in genome size may be a consequence of ecological shifts that accompany life history differences being selfing and outcrossing species . However , the genome of the model self-fertile Arabidopsis thaliana ( 125Mb ) is smaller than its outcrossing relative A . lyrata ( 207Mb ) largely due to numerous small and large-scale deletions , including 17% fewer genes [30] . TEs play a major role in plant genome size evolution [24 , 51 , 52] and , while the genome of A . lyrata does show an increase in TE activity relative to A . thaliana , a comparison between Capsella rubella , a plant that became self-fertile less than 200 , 000 years ago , and the related outcrossing C . grandiflora reported few differences in TE content [53] . We find no evidence that the genome size differences between selfing and outcrossing species are mediated by TE activity and/or other forms of small indels . Instead , DNA loss in Caenorhabditis appears to have occurred specifically at the level of individual genes . Natural selection might drive genome reduction , or genome shrinkage could accrue through deletion biases and genetic drift . DNA content is positively correlated with increased cell size and negatively correlated with cell growth and division , metabolic rate [54] , and developmental rate [32] , but it is unclear why self-fertility would necessarily lead to increased selection on these traits . Alternatively , selective sweeps on new beneficial mutations could lead to the fixation of weakly deleterious deletions because of increased linkage disequilibria in selfers . The fixation of such deletions would be particularly facilitated in a non-adaptive manner if deletion per se acted as a directional process . For example , deletions predominate over insertions in C . briggsae nuclear [55 , 56] and mitochondrial DNA [57] , however insertions predominate over deletions in C . elegans mutation accumulation lines [55 , 58 , 59] and under temperature stress [56] . Perhaps most interestingly , there is evidence in C . elegans that autosomal deletions are preferentially transmitted to X-bearing sperm ( and thereby hermaphrodites in this XO sex determination system ) [37] . This kind of bias could rapidly reduce genome size following the transition to self-fertilization . However , rather than observing systematic reduction in gene size with the selfing species , we find that both C . elegans and C . briggsae have larger introns on the X than C . remanei while maintaining similarly sized genes on the autosomes . Thus , while there are definitely deletions across the whole genome , they are at the level of whole genes instead of being randomly spread across functional elements such as introns , as would be expected if the genome size reduction were driven by a directional mutation process . Instead , it appears that adaptation to self-fertility per se is the most likely explanation for the reduction in genome size . For instance it may be advantageous to lose/alter systems directly related to maintaining outcrossing , such as mating in the case of nematodes or floral characteristics in the case of plants [23 , 60] . In keeping with this , Thomas et al . [31] found that genes with higher degrees of sex-specific expression tend to be lost more frequently than other genes within these same species . Similar loss of function changes appear to be common in other cases of adaptive phenotypic evolution [61] . Overall , then , these results suggest an evolutionary model for genome reduction following the evolution of selfing within this group of nematodes: 1 ) relaxed selection on specific genes , like those involved in facilitation of outcrossing [31 , 53]; 2 ) deletion of genes and their surrounding intergenic sequences; and 3 ) accumulation of these deletions resulting in derived decreases in genome size . There are several other genera of nematodes within this family that also show variation in mating systems [62 , 63] , so it should be possible to discern if this is a general pattern of genome loss within self-fertilizing species . Lynch [42] proposed the effective population size hypothesis as a first step in transforming “the descriptive field of comparative genomics into a more mechanistic theory of evolutionary genomics . ” Our results indicate that self-fertile organisms experience loss of DNA as a general feature of the transition in mating systems but that these losses are not driven by changes in effective population size per se . The specific mechanisms by which this occurs appear to vary across groups , from plants to animals . Linking changes within specific genomic functional classes with the dynamics of natural selection and fitness differences that favor DNA loss in self-fertile organisms is the next step in understanding the influence of mating system on genome evolution .
C . remanei strains were cultured using standard methods adopted from techniques developed for C . elegans[64] . The C . remanei PX356 strain was created from the canonical isofemale line EM464 originally isolated from New York ( USA ) , which was initially named C . vulgaris but has since been synonymized with C . remanei[65] . The existing polymorphic assembly of C . remanei is based on a partially inbred line derived from this same strain ( EM4641 ) . To overcome the extreme inbreeding depression usually observed within C . remanei , we derived 200 independent lines from the original EM464 population and subjected them to brother-sister mating in order to allow the independent fixation and loss of as many recessive deleterious alleles as possible . All but two of the lines went extinct by generation 7 . These two remaining lines were then crossed together and maintained as an outcrossing population for 20 generations to increase the probability that deleterious alleles alternatively fixed in the two lines could recombine . An additional 100 lines were derived from this secondary population and then subject to brother-sister mating for 23 generations . The sole surviving line from this procedure was deemed the new canonical EM464-derived line: PX356 . In order to create an alternative mapping line , a similar procedure was conducted for an isofemale line ( PB259 ) originally collected from a forest in Ohio ( USA ) by Scott Baird ( Wright State University ) . This resulted in an additional , divergent inbred line: PX439 . Genomic DNA was isolated from either starved L1 larvae following a bleach “hatch-off” or from mixed stage populations from a sucrose float using the DNeasy Blood and Tissue kit ( Qiagen ) with the C . elegans supplemental protocol . Total RNA was isolated from mixed stage populations using Trizol ( Invitrogen ) . mRNA was purified using Dynabeads Oligo d ( T ) 25 ( Invitrogen ) and fragmented ( Ambion ) before cDNA synthesis [66] . Paired-end genomic DNA sequencing libraries were constructed using the Nextera DNA sample preparation kit ( Illumina ) or the NEBNext DNA library kit for Illumina ( NEB ) as per the manufacturers protocols . Using the NEBNext kit , genomic DNA was fragmented with a Bioruptor sonicator ( Diagenode ) set on high for ten 30 second ON/OFF cycles . Final libraries were size selected on 2% agarose gels with an average genomic insert size of 180 bp as per ALLPATHS-LG recommendations . All libraries were quantified by qPCR ( Life Technologies ) and the proper size range was confirmed using a fragment analyzer ( Advanced Analytical Technologies ) . Libraries were sequenced as 2 X 101nt reads using an Illumina HiSeq instrument . The mate pair libraries were constructed using standard molecular techniques following the manufacturers recommendations . In brief , genomic DNA was sheared on the low setting for 5 seconds using a Bioruptor sonicator ( Diagenode ) and purified using the Desalting and Concentrating DNA section for the QIAEX-II kit ( Qiagen ) . DNA was end-repaired using the End-it kit from epicentre . Following purification , the DNA was biotin labelled with 1mM dNTP ( 4% biotin ) , purified and run on a 0 . 6% agarose gel . Size ranges of 3 , 5 and 7 kb were isolated and purified using the standard QIAEX II kit . Circularization was carried out overnight at 16 deg . C with ~200ng DNA using T3 ligase ( Enzymatics ) with T4 ligase buffer . The non-circularized DNA was digested with 3μl of DNA exonuclease ( NEB ) and placed at 37 deg . C for 20 min . and heat inactivated for 30 min . at 70 deg . C . The DNA was then sheared to ~400 bp using a Bioruptor for ten 30 sec . ON/OFF cycles at high and then purified . The biotin labelled DNA pieces were isolated using Dynabeads M-280 strepavidin ( Invitrogen ) . While on the beads , the DNA fragments were end-repaired , A-tailed , and t-overhang adapters were added before PCR enrichment for 15 cycles using Phusion polymerase ( NEB ) . Libraries were isolated from 2% agarose gels at ~ 400 bp average size range and eluted in EB buffer with 1% tween 20 . Final libraries were validated for correct size and molar concentration as noted above . Mate-paired libraries were sequenced as 2 X 101nt reads using an Illumina HiSeq instrument ( cDNA synthesis and RNA sequencing libraries were prepared as previously reported [66] ) . We constructed the C . remanei assembly from ∼ 560x coverage of 180bp paired end fragments designed to have overlapping reads on both ends . This high depth of coverage will necessarily include a large number of sequencing errors which would have complicated assembly , and a large number of repeat elements which would have increased assembly time . We pre-filtered the 180bp paired end fragments by kmer frequency spectra ( here , k = 15 ) to address these biases . We removed reads with greater than 12 rare kmers ( singletons ) to eliminate possible errors and reads with greater than 51 abundant kmers ( occurring more than 20 , 000 times in our dataset ) to eliminate possible over-represented repeats ( additional details are given in S1 Text ) . The final short-read dataset averaged 416x coverage across the estimated genome size ( 131Mb ) . We also used mate pair libraries with inserts of varying sizes: 31 . 5x coverage of 0 . 7–2kb insert paired end fragments , 29x coverage of 2–4kb insert paired end fragments , and 15x coverage of 4–7kb insert paired end fragments . We sequenced 101bp reads with 6bp inline barcodes which resulted in 95bp sequences . We used the assembly software ALLPATHS-LG [67] , which performs its own kmer spectra correction of sequencing reads ( k = 25 ) , uses a de Bruijn graph algorithm to build contiguous sequences from the 180bp reads , and constructs scaffolds with mate pair sequences . The initial heterozygosity rate was estimated as 1/176bp and we used the haploidify option to address residual heterozygosity in our inbred strain ( S3 Table ) . We used a multi-step decision tree to identify possible contaminant sequences and removed 17Mb of contaminant scaffolds ( details are given in S1 Text; S4 Fig ) . After assembly we aligned the paired end sequences to our final linkage groups and scaffolds with GSNAP [68] and used SAMtools [69] to summarize single nucleotide polymorphisms ( SNPs ) in a . vcf ( variant call format ) file . We analyzed this file with a custom pipeline of perl scripts that divide the annotated genomic regions into introns , exons , transposable elements , intergenic regions , transcription start sites , 5’ UTRs and 3’ UTRs and measures the number of polymorphic sites in each of these categories . Read alignments were discarded if the per-base sequencing coverage exceeded 600 ( as these may be poorly assembled , collapsed repetitive regions ) or if the base quality fell below 90% certainty ( Q10 Phred+33 ) . Our estimated residual polymorphism therefore applies only to the well-assembled genic regions and excludes intergenic sequences , transposable elements and other poorly assembled repeats . Our kmer frequency spectra filtering removed over-represented sequences . In order to eliminate the possibility that this filtering may have created a bias in assembling repetitive elements we also de novo assembled the entire set of 180bp paired end fragments and mate pair libraries with ALLPATHS-LG [67] . We analyzed the repeat content of this genome sequence ( methods are given below in the section Characterizing repeats ) and found that the repeat content of this sequence was lower than our kmer-filtered genome sequence ( 13 . 92% vs . 15 . 73% ) . The kmer-filtered assembly was 118 . 5Mb with 4Mb of gapped sequence and 1 , 600 scaffolds ( S3 Table ) . The unfiltered assembly was similar in size ( 119 . 1Mb ) but contained 7Mb of gapped sequence and was fragmented into 3 , 921 scaffolds . We concluded that kmer frequency spectra did not specifically eliminate repeat elements from our genome but it did remove noise and facilitate a higher-quality , ungapped genome sequence . We generated 64 recombinant inbred lines ( RILs ) from a cross between parental strains PX356 and PX439 following an advanced intercross method [70] . We then used restriction site associated DNA ( RAD ) markers [71] , generated with EcoRI , to identify SNP markers within each of these strains , as well as in the parental lines . We used the software Stacks[72] to assign sequencing reads to genetic loci and identified 25 , 447 mappable ( i . e . , AA x BB where A is the PX356 genotype and B is the PX439 genotype ) polymorphic markers . The parental strains appear to be partially reproductively isolated and so many of these SNP markers showed extensive segregation distortion ( S5 Fig ) . We used SAMtools [69] to align the markers to our assembled scaffolds and identified scaffolds where >5 markers had parental genotype frequencies of >80% or <20% in the RILs . We eliminated all markers located on these scaffolds , markers that contained only duplicate data and those with low representation in the RILs ( present in <40 lines ) . We constructed a genetic map in R/qtl [73] from the resulting 330 distinct SNP markers ( S6–S7 Figs ) . We added markers with duplicate data back to the dataset and the final genetic map contains 2 , 688 SNP markers across 65 Mb ( S8–11 Figs ) . The genetic map identified 4 scaffolds that were incorrectly joined in the assembly , and we broke these on the basis of parental genotype frequency in the RILs and synteny with the existing C . remanei draft assembly . We aligned the 3 large linkage groups to the C . elegans and C . briggsae chromosomes X , II , and IV but smaller linkage groups were not definitively assigned to chromosomes . The PX356 genome was repeat masked with RepeatMasker v4 . 0 . 5 [74] using a custom C . remanei repeat library created by RepeatModeler v1 . 0 . 8 . In order to compare repeat content between nematode species we also created custom repeat libraries for C . briggsae , C . sinica , C . tropicalis , C . brenneri , C . elegans , C . japonica , C . angaria , and P . pacificus , and repeat masked each genome with the custom repeat library and RepeatMasker . We compared these custom repeat characterizations to the repeats currently annotated in each genome ( Wormbase release WS244 ) and found that each custom characterization was within 1–2% of the currently annotated content with the exception of C . japonica for which we predicted 42% repeat content . This is consistent with previous findings that the allelism in the C . japonica assembly produced over-representation of repeat regions [19] . We present the currently annotated content in Fig 1 . We sequenced mRNA from our PX356 nematodes with a Illumina Hi-Seq machine and generated 16 , 250 , 052 paired end sequences . We assembled these data with the genome-independent Trinity RNA-seq assembly software [75] to generate a set of putative transcripts . We used the software MAKER2 [76] to annotate putative protein-coding loci , with the Trinity transcript set as EST evidence and the protein sequences from C . briggsae , C . elegans , C . brenneri , and the Uniprot/Swiss-Prot [77 , 78] database ( with Caenorhabditis removed ) as protein homology evidence . The maximum intron size was specified as 5 , 000 nucleotides for evidence alignments . We used SNAP [79] and Augustus [80]ab initio gene prediction software within MAKER2 to generate a putative set of predictions , and used these initial predictions to re-train SNAP and Augustus to produce C . remanei-specific gene prediction models . We predicted 26 , 339 transcripts from 25 , 415 protein-coding genes and used a manually curated set of 169 miRNA sequences to annotate putative miRNA loci . Roughly half of the annotated genes ( 12 , 323 ) reside on the 13 linkage groups . Although the assembly is 9 . 5% shorter than the flow-cytometry estimated genome size , analysis of a core set of genes that are thought to be conserved in single copy in eukaryotes [81] indicates that 95 . 16% are in complete form in the assembled sequence and the remaining 4 . 84% are partially complete . We used the same annotation pipeline to annotate protein-coding genes in the genomes of C . angaria , C . brenneri , C . briggsae , C . elegans , C . japonica , C . sinica and C . tropical . We found that the re-annotation estimated a similar protein-coding footprint with the exception of C . brenneri , for which we predicted >15Mb additional genic content . This is consistent with previous findings that >30% of the genes in the C . brenneri genome are found in two copies [19] . We identified protein motifs and domains with InterProScan [82] ( version 5 . 3 ) , which searches against public protein databases including ProDom [83] , PRINTS [84 , 85] , Pfam [86] , SMART [87] , PANTHER [88] and PROSITE [89] . We used the Repbase ( version 16 . 10 ) database of repetitive elements and a library of de novo repeats identified with RepeatModeler [29] to analyze transposable elements and genomic repeat content . We used OrthoMCL ( version 1 . 4 ) to identify orthologous gene clusters between C . remanei , C . briggsae , C . elegans , and C . brenneri . The C . brenneri genome assembly is highly fragmented ( 3305 scaffolds ) and conflated by ~30% allelic assembly artifacts to be 55Mb larger than the flow-cytometry estimated genome size [19] , but is the only other outcrossing Caenorhabditis in the Elegans group with a genome sequence suitable for comparative analyses . Briefly , OrthoMCL [90] uses BLASTP [91] to calculate pairwise protein sequence similarities , and Markov clustering of the similarity scores to define orthologous proteins among the species and paralogous proteins within each proteome . In order to compare genome content between nematode species we used our MAKER2 gene annotation pipeline to identify protein-coding genes in the Wormbase release WS244 genome sequences of C . briggsae , C . sinica , C . tropicalis , C . brenneri , C . elegans , C . japonica , C . angaria , and P . pacificus ( S2 Table ) . We found that each custom characterization was within 3–5% of the currently annotated content , with the exception of C . brannier for which we predicted a larger gene content . This is consistent with previous findings that >30% of C . elegans orthologous genes are found in two copies in the C . brenneri assembly [19] . We present the currently annotated content in Fig 1 . We used BedTools v2 . 22 . 1 [92] to identify intergenic spaces , genic regions , and exonic regions in the genome sequences of C . elegans , C . elegans , and C . remanei . We used the statistical computing language R to calculate descriptive statistics , Kruskal-Wallis rank sum tests , and Pairwise Wilcoxon rank sum tests . We were not able to use parametric statistics because our data were nucleotide counts , each of the distributions showed severe heteroscedasticity , and the sample sizes were unbalanced . Nonparametric statistical tests can not address interaction terms or multi-factor comparisons and we performed one-way tests and Bonferroni corrected our p-values for multiple comparisons . Assembly and annotation are at www . wormbase . org and are deposited in GenBank under BioProject ID PRJNA248909 . De novo annotations of existing genomes are published at figshare . com at dx . doi . org/10 . 6084/m9 . figshare . 1399184 , dx . doi . org/10 . 6084/m9 . figshare . 1396472 , dx . doi . org/10 . 6084/m9 . figshare . 1396473 , dx . doi . org/10 . 6084/m9 . figshare . 1396474 , dx . doi . org/10 . 6084/m9 . figshare . 1396475 , dx . doi . org/10 . 6084/m9 . figshare . 1396476 , dx . doi . org/10 . 6084/m9 . figshare . 1396477 , and dx . doi . org/10 . 6084/m9 . figshare . 1396478 . Our analysis pipeline is available at github . com/Cutterlab/popgenome_pipeline .
|
Closely related species can vary widely in genome size , yet the genetic and evolutionary forces responsible for these differences are poorly understood . Among Caenorhabditis nematodes , self-fertilizing species have genomes 20–40% smaller than outcrossing species . Constructing a high quality de novo genome assembly in C . remanei , we find that this outcrossing species has many more protein coding genes than the self-fertilizing Caenorhabditis . Intergenic spaces are larger on the X chromosome and smaller on autosomes for both selfing and outcrossing Caenorhabditis , but protein-coding genes are larger on the X chromosome in the self-fertile C . briggsae and C . elegans and larger on autosomes in the outcrossing C . remanei . This contrasting pattern of contracting genomes and expanding genes is likely mediated by changes in the balance between genetic drift and natural selection accompanying the transition to self-fertilization .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Reproductive Mode and the Evolution of Genome Size and Structure in Caenorhabditis Nematodes
|
Adaptive behavior often exploits generalizations from past experience by applying them judiciously in new situations . This requires a means of quantifying the relative importance of prior experience and current information , so they can be balanced optimally . In this study , we ask whether the brain generalizes in an optimal way . Specifically , we used Bayesian learning theory and fMRI to test whether neuronal responses reflect context-sensitive changes in ambiguity or uncertainty about experience-dependent beliefs . We found that the hippocampus expresses clear ambiguity-dependent responses that are associated with an augmented rate of learning . These findings suggest candidate neuronal systems that may be involved in aberrations of generalization , such as over-confidence .
Successful behavior in new situations often requires us to apply ‘rules-of-thumb’ . However , acquiring and applying abstract rules from limited experience presents a fundamental computational problem [1]: in which both over- or under-generalization must be avoided [2] , . Despite their importance , little is known about how neuronal systems learn these rules , and how the delicate balance between past and present information is maintained . Evolutionary arguments suggest that the use of previously learned rules when generalizing to new situations increases adaptive fitness by optimizing behavior [7] . This raises the key question of whether and how generalization is optimized [8] . In this work , we examine whether human subjects combine previously learned rules and current information in an optimal way and identify the brain systems that underlie this combination . Using Bayesian learning theory to specify optimal generalization , we looked for its neural correlates . In particular , we drew on existing evidence that points to the hippocampus as a key structure that is implicated in learning the specifics of a new situation , when previously learned rules may not apply [9] , [10] . Probabilistic inference in a natural environment is confounded by multiple sources of uncertainty [11] , [12] , [13] , [14] , including objective randomness and subjective ignorance [14] . Uncertainty is a key concept here because the confidence about prior beliefs should be weighed against the confidence about new information , when deciding whether to generalize those beliefs to a new situation . Classical reinforcement learning models ( e . g . [15] , [16] ) do not represent uncertainty or use generalization to guide learning and behavior: these schemes simply learn the expected value of action-states and only prosper in environments where the current state is sufficient to specify a successful action: see [17] for a critique and extension . Having said this , several other RL schemes are based on some form of non-probabilistic function approximation and therefore support generalization ( see Chapter 8 in [18] for discussion and recent RL approaches in neuroscience that consider generalization in the spatial [19] and temporal [20] case ) . While recent RL developments in neuroscience incorporate some notion of uncertainty [21] , learning and generalization are typically non-probabilistic . In this work we ask if learnt generalizations are accompanied with due uncertainty [22] , as prescribed by probability theory . At the behavioral level , human subjects readily abstract probabilistic rules and use them to generalize [8] . Furthermore , they can distinguish different sources of uncertainty: the unavoidable or irreducible randomness of certain events versus subjective ignorance about the world [12] , [13] , [23] , [24] , [25] . The latter resembles the concept of subjective ambiguity in economics and represents uncertainty about objective risks . For example , the risk ( or irreducible randomness ) associated with a fair coin toss is high ( 50∶50 ) ; however , there may be subjective ambiguity as to whether the coin is itself fair . This paper examines the function and mechanisms of generalization in the face of ambiguity . While there are good reasons to restrict the term ambiguity to complete ignorance [26] , we use the term more inclusively to denote the level of uncertainty about the outcome probabilities . This is akin to estimation [14] or second-order [26] uncertainty ( i . e . , uncertainty about uncertainty ) . Ambiguity is subjective and reference-dependent: it ranges from complete ignorance to near certainty and , crucially , can be reduced by generalization in a Bayes-optimal fashion [8] . In other words , if subjects consider their current situation in the light of past experience , they can exploit similarities between the past and present to reduce their ambiguity [27] , [28] . In our example , ambiguity about a new coin will be reduced by observing the random behavior of similar coins . This ability to generalize over similar situations is seen readily in behavior and learning [8] , [14] . In this study , we examined the neuronal correlates of generalization with a special focus on the hippocampus: The hippocampus is involved in generalization [29] , [30] , [31] , [32] , [33] , [34] and shows activations that are sensitive to objective uncertainty or risk [23] , [35] . In this paper , we asked if hippocampal responses also report subjective uncertainty or ambiguity that changes with experience . Specifically , we tested for ambiguity-dependent hippocampal responses , when probabilistic nature of outcomes had to be learned . Furthermore , we hoped to show behaviorally that learning rates were greater in contexts that had more ambiguity . We addressed these questions using a model of our experimental task and , tested whether Bayesian updates or learning could explain behavioral and neurophysiological responses , as measured with fMRI .
Nineteen subjects ( age 19–31 , 11 female ) were recruited from the UCL psychology Dept subject pool . All subjects gave informed consent , before reading a brief description of the task which was then performed under fMRI . The study protocol was approved by the local UCL ethics committee . While our goal was to identify domain-general computational processes , the paradigm was framed as a social inference task: Subjects were told that two groups of thirty individuals had completed a marketing survey . Subjects were then asked to guess , over ten consecutive trials , whether each individual would choose a ‘blue’ or ‘purple’ product . Subjects were told they would be paid ‘in proportion to the number of correct guesses’ and that the two groups were ‘geographically and economically unlike one another’ . Trial cues ( individuals ) were faces from the Sterling data-set , whose group membership was indicated by the symbol ‘*’ or ‘o’ ( see Figure 1 ) . Each trial comprised the following sequence: 1 ) an individual's face was presented along with the symbol indicating their group membership; 2 ) the response options ( blue and purple squares ) were then presented , after which 3 ) the subject responded and 4 ) received feedback about whether their guess was correct or incorrect . The timeline for a single trial is shown in Figure 1 . If subjects did not guess within one second , they were shown the instruction ‘ACT FASTER ! ’ . The subject's guess was highlighted until feedback was delivered . Correct guesses were signaled with an auditory beep ( 500 milliseconds of 500 Hz sine wave ) and accumulated in a score bar at the bottom of the screen . Incorrect guesses were indicated by a 500 millisecond burst of white noise ( with no increase in their score ) . Unbeknown to subjects , individuals from one group had similar preferences , while the other group had more between-individual variability . This meant that subjects had to make guesses about choices in two distinct contexts established by the group an individual belonged to: in the generalization context ( GC ) , all individuals chose ‘purple’ with probability . In the ambiguous context ( AC ) , ‘blue’ was probabilistically chosen ( ) by half of the group members and ‘purple’ ( ) by the other half . To reiterate , subjects were presented with the same face ten times and had to guess whether the individual preferred blue or purple . Each individual was identified as belonging to one group or the other . Every individual preferred one color that was chosen 80% of the time . In the generalization context , all group members preferred the same color , while in the ambiguous context , individual group members preferred blue or purple with equal probability . In both contexts , subjects could learn about any given individual over ten trials . The generalization context therefore contained a probabilistic rule prescribing the best guess , even in the absence of learning about an individual's preferences . Conversely , in the ambiguous context , subjects had to learn about individual preferences because their group membership provided no clues about what they would preferentially choose . Trials were arranged into blocks , in which the same individual was presented for ten consecutive trials . The blocks alternated between AC and GC , with a new individual ( face ) for each block . This resulted in trials , for thirty individuals , presented ten times for two groups . The blue and purple options were presented with equal probability on the left and right of the screen on each trial . Individuals ( faces ) were randomly reassigned to either group , between subjects . All subjects experienced the same feedback contingencies ( with randomly reassigned cues ) . Subjects had three short breaks during the task: for each they were first cued ‘PLEASE HAVE A SHORT REST AND RELAX’ before being prompted to restart thirty seconds later: ‘OK ! PLEASE PRESS ANY KEY TO CONTINUE’ . Bayesian learning theory predicts that subjects should learn more quickly about a new individual from the ambiguous group , relative to an individual from the generalization group . This is based upon the assumption that subjects are making Bayes-optimal guesses using a notion of group or context . The increase in learning rate with higher levels of ambiguity is related to increases in learning rate in situations with a high degree of volatility [39] ( see below ) . At the neuronal level , we predicted that increases in learning rate would selectively engage hippocampal processing in the ambiguous context . In other words , hippocampal activation should track changes in ambiguity about an individual's preference as it alternates between AC ( high ambiguity ) and GC ( low ambiguity ) blocks . To quantify ambiguity , we assumed subjects were ideal Bayesian observers who used a model of probabilistic outcomes . We focused on two alternative models to predict subject responses , M1 and M2 . Under M1 , Bayesian learning combines new information with existing generalizations based on group membership . Conversely , M2 accumulates information about every individual independently , without the benefit of generalization . To make optimal guesses about the choices of each group member , subjects have to infer their preferences i . e . the probability that this individual will choose a particular option , say ‘purple’ . We denote this probability with . The information following each trial is equivalent to observing the outcome of a biased coin . We use the random variable to denote whether the choice of the individual was ‘purple’ ( ) or ‘blue’ ( ) : in trial ( subjects encountered individuals in each of the two groups ) . In what follows , we consider alternative models that subjects might have used to infer the . We start with a model that permits generalization and then turn to a version that precludes generalization . We also consider a few alternative models that can be considered as special cases that are of interest from an RL perspective .
Figure 2 shows the value ( expected reward ) of each choice according to the two main learning models we considered , together with a typical subject's guesses . Model 1 ( M1 ) generalizes , while Model 2 ( M2 ) cannot . For each subject , we used logistic regression to explain their choices in terms of these predictions and a constant term . Using a between-subject summary-statistic approach , we applied a two-tailed Student's t-test to the subject-specific logistic regression coefficients associated with the predictions of M1 ( red-dashed curve , Figure 2 ) . We rejected the null hypothesis that this effect was equal to zero ( , ) . Interestingly , the size of the M1 regression coefficient predicted the total number of rewards obtained by each subject ( correlation , ) . This illustrates that generalization is evident behaviorally and pays off . A secondary behavioral analysis assessed the specificity of M1 predictions by examining the explanatory power of M1 in the context of the alternative models , M2 to M5 . For each subject , we used logistic regression to explain subject's choices as a mixture of predictions from five models ( M1 to M5 ) , plus a constant term . Having estimated the logistic regression model for each subject , we again considered the subject-specific estimates for the coefficient reporting on M1 predictions . A two-tailed Student's t-test on the M1 coefficients was highly significant , . No other model coefficients reached significance . To summarize , we used standard regression techniques to ask if , having accounted for competing models , a component of choice behavior reflects Bayes-optimal generalization ( M1 ) . Specifically , we included several model predictions in one linear model and estimated the partial regression coefficient for the predictor of interest ( action-values derived from M1 ) . One can therefore [54] conclude that , over and above competing models , behavior can be predicted by M1 . Because Models M2 and M3 have a free parameter this conclusion is conservative: having been pre-fit to subject's behavior , these models have an explanatory advantage that is unavailable to M1 ( or M4 and M5 ) . In contrast to M2 ( Eq . 8 ) , M1 attempts to explain behavior via abstract computational principles , not detailed mechanisms . Its predictions have no free parameters . Rather , its predictions are based only on the subject's observations under ideal Bayesian assumptions . We have demonstrated that this model predicts behavior , above and beyond that explained by the other models considered . In what follows , we now ask whether the brain encodes ambiguity [see e . g . [39] for a similar approach] . While M1 differs from other models in many ways , the important aspect for the fMRI analysis is that M1 provides an ambiguity measure . We therefore tested the null hypothesis that the fMRI signal is sensitive to ambiguity , as quantified by the Shannon entropy of prior belief ( see above ) . In our fMRI data , fourteen subjects satisfied the inclusion criteria for a second-level between subject analysis ( no interruptions to the scanner session or rapid head movement , as estimated by co-registration ) . We conducted regional and whole-brain analyses . All fMRI results presented here are based on the same general linear model , including the confounding factors ( i . e . , with nine regressors ) . In view of our specific hypothesis , region of interest ( ROI ) analyses asked whether activity within bilateral hippocampi tracked ambiguity about the current contingencies . Figure 4a shows the anatomy of the ROI . Figure 3 depicts ambiguity about a new individual ( alternating block-wise between GC and AC blocks ) . As discussed , this dictates the relative influence of the current observation on belief updates ( higher when there is high ambiguity ) . The parameter estimates associated with the entropy regressor above were averaged over bilateral hippocampal voxels for each subject , using the AAL atlas [55] . We applied a two-tailed Student's t-test to these subject-specific summaries , testing the null hypothesis that hippocampal responses do not covary with ambiguity . We were able to reject this null hypothesis with a correct . Repeating the analysis on unilateral right and left hippocampus separately provided similar results ( , , respectively ) . ( These latter two results examine the separate contribution of each hemisphere to our bi-lateral effect . These tests are not statistically independent of the bi-lateral test and were not subject to additional correction . ) There was no significant difference between left and right hippocampi . Our results therefore suggest that neuronal activity encodes the same sorts of variables that arise in our Bayes-optimal computations and , consequently , may be performing some form of approximate Bayesian inference . As with the behavioral data , we next examined the between-subject correlation between the hippocampal ambiguity coefficients and the total number of rewards attained in the experiment ( correlation , ) . Testing for separate correlations in left and right hippocampal effects gave respectively: , and , ( ) . In an exploratory whole brain analysis , we then smoothed the data with a Gaussian Kernel and re-estimated the general linear model above using a conventional SPM analysis with whole brain correction for multiple comparisons [56] . Two right-hemisphere clusters survived correction for cluster-extent ( using a height threshold of 3 ) . The first region ( FWE corrected ) subsumed a right hippocampal region , mostly hippocampus and amygdala , but also putamen , as defined with the AAL atlas [55] . The second region ( FWE corrected ) encompassed the fusiform gyrus and precuneus , with a spill-over into a calcarine region . These regions are shown in maximum intensity projection format in Figure 4b ( this display format shows voxels with maximum intensity that fall on parallel lines traced from the viewpoint to the plane of projection as in a standard X-Ray ) . Orthogonal views of the anterior activation at its local maximum are shown in Figure 4c . For illustration purposes , Figure 4d shows the mean times series in this anterior region , averaged over all subjects . All of the above fMRI analyses were based on the same model , which included the nuisance regressors listed in Relating model predictions to data: fMRI . None of these nuisance effects could explain the variation in hippocampal responses that was explained by our Bayes-optimal generalization model ( M1 ) .
Behaviorally , we have shown that subjects learn action-reward relationships in a manner that enables them to generalize rules to new situations . Crucially , this enables subjects to adapt their learning rate to provide an optimal balance between pre-existing generalizations and new information . We established this by showing that the accuracy of subjects' guesses evolved over trials in a way that was predicted by Bayes-optimal generalization , using a statistical model equipped with prior beliefs that allowed for contextual ambiguity . Furthermore , we established that a significant component of hippocampal responses could be explained by fluctuations in ambiguity under this model . These regionally specific responses were also significant in a whole brain SPM analysis . We provide empirical support for a model that explains how experience moderates decision making . In this model , the bias towards rule-based choices is determined by low ambiguity . We show that both learning and hippocampal responses are attenuated when the underlying rule is learned and applied in an unambiguous context . Conventional ‘model-free’ reinforcement learning cannot easily explain such effects because these schemes do not include contextual ambiguity . As noted in the introduction , one recent variant of reinforcement learning [43] is relevant here: In this two-system learning theory , generalization between observable cues rests both on their perceptual similarity and their predictive similarity ( do cues look the same ? do they predict the same outcomes ? ) . The authors of [43] contrast normal learning with under/over-generalization or ‘under/over willingness to generate a new state’ p 97 . We have used a single model that formalizes this optimality by drawing on principles of optimal probabilistic generalization ( see [42] for a related model ) . As in [43] , our model generalizes by classifying observable cues before acting . Unlike [43] , it invokes an explicit representation of subjective ambiguity to mediate and optimize this generalization . There remains an interesting challenge to relate our formulation and results to classical RL schemes . Interestingly the authors of [43] speculate that the neuronal systems mediating generalization depend on the hippocampus ( and PFC ) ; because these systems are flexible , the rules by which observable cues are classified can easily be changed to permit new discriminations . These speculations are entirely consistent with our findings . As in previous treatments [14] , we distinguish uncertainty about objective , observable events ( e . g . , the risk of getting ‘tails’ in a fair coin flip ) from subjective ambiguity about unobservable states or parameters ( is the coin really fair ? ) . While the hippocampus has been implicated in the former [23] , [35] , [57] , the latter is central to computational accounts of contextual learning and inference; e . g . [1] , [22] . Using a Bayes-optimal model , our work provides the first evidence that the hippocampus tracks contextual ambiguity about hidden or latent variables . Previous work [11] , [14] , [39] , [58] has addressed how ambiguity mediates the influence of uncued temporal variability ( volatility ) on learning . We asked if variability in response requirements to different cues influences creates ambiguity and influences learning . In the current study , we manipulated the uncertainty about the behavioral contingencies over contexts , rather than time , and showed that associative learning adapts accordingly . Further work could examine whether neuromodulatory manipulations influence this effect; e . g . , by selectively facilitating synaptic gain as predicted by [11] , [59] . The role of dopamine deserves special attention , given prior work with Pavlovian or simpler instrumental tasks [60] . Additionally , given that the amygdala is able to modulate memory storage in non-amygdala brain areas [61] , multi-region in vivo recordings could disclose interactions with the hippocampus in these tasks . Interestingly , the amygdala activation in our whole-brain analyses is consistent with previous work implicating the amygdala in the representation of ambiguity [24] , [62] . However , previous studies were unable to address whether ambiguity regulates learning , as predicted theoretically . In line with Bayesian learning theory , our results suggest that learning ( updating beliefs ) can be guided by optimal probabilistic constraints , generalized from previous experience . The learning rate in ( model-free ) reinforcement learning prescribes the sensitivity of belief updates to current information . When this information is under or over-weighted , inefficient learning ensues . While classical RL is non-probabilistic ( i . e . has a degraded uncertainty representation [22] ) , it may in principle address this challenge by incorporating something akin to an ‘ambiguity-dependent’ or ‘surprise-dependent’ learning rate . For example , attempts have been made to optimise learning rates [63] , [64] in both stationary and non-stationary settings [65] . Bayesian learners use the rules of probability to achieve this balance by weighing new information against pre-existing generalizations . The relative weight of the latter depends upon ambiguity ( the relative confidence in prior beliefs about the current context ) . When pre-existing beliefs are held with a high degree of confidence , they generally accommodate new observations , by down-weighting their impact . Such abilities to balance different sources of information and constraints are at the heart of adaptive behavior [66] . For example , appropriate social behavior requires communal norms , while retaining sensitivity to individual inclinations and preferences . The ( social ) learning task in this paper is a first step in this direction . Conversely , aberrant generalization has widespread consequences [2] , [3] , [5] , [67] . The framework used in this study may provide an experimental framework to quantify dysfunctional generalization in specific patients; e . g . , over-generalized schemata which persist despite contradictory evidence , as seen in depressive and delusional states and its associated pathophysiology at the neuronal level .
|
Intelligent behavior requires flexible responses to new situations , which exploit learned principles or abstractions . When no such principles exist , the imperative is to learn quickly from scratch . Behaviorally , we show that subjects learn action-reward relationships in a manner that enables them to generalize rules to new situations . Our fMRI results show that when subjects have no evidence that such a rule exists , medial temporal lobe responses ( that reflect uncertainty ) predict their augmented learning .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"psychology",
"social",
"and",
"behavioral",
"sciences",
"biology",
"neuroscience"
] |
2012
|
Learning and Generalization under Ambiguity: An fMRI Study
|
Antibiotic regimens often include the sequential changing of drugs to limit the development and evolution of resistance of bacterial pathogens . It remains unclear how history of adaptation to one antibiotic can influence the resistance profiles when bacteria subsequently adapt to a different antibiotic . Here , we experimentally evolved Pseudomonas aeruginosa to six 2-drug sequences . We observed drug order–specific effects , whereby adaptation to the first drug can limit the rate of subsequent adaptation to the second drug , adaptation to the second drug can restore susceptibility to the first drug , or final resistance levels depend on the order of the 2-drug sequence . These findings demonstrate how resistance not only depends on the current drug regimen but also the history of past regimens . These order-specific effects may allow for rational forecasting of the evolutionary dynamics of bacteria given knowledge of past adaptations and provide support for the need to consider the history of past drug exposure when designing strategies to mitigate resistance and combat bacterial infections .
Antibiotic resistance is a growing healthcare concern whereby bacterial infections are increasingly difficult to eradicate due to their ability to survive antibiotic treatments [1] . There have been reported cases of resistance for nearly every antibiotic we have available [2] . Coupled with the fact that the antibiotic discovery pipeline has slowed over the past few decades [3] , there is a dire need to find better treatment strategies using existing antibiotics that can slow or even reverse the development of resistance . Adaptive laboratory evolution is a technique that can be used to study and test evolutionary principles in a highly controlled laboratory setting [4] . Microorganisms with short generation times such as bacteria are especially amenable to adaptive laboratory evolution and can be adapted to an environment through repeated cycles of growth in a specific media environment , dilution of the culture , and subsequent passaging into fresh media [5] . Multiple replicates for each condition can be evolved in parallel to investigate the reproducibility of evolutionary dynamics [6] . The evolutionary trajectories of the bacteria can be measured as they adapt to different nutrients and stressors over time [7] . Whole-genome resequencing on the evolved strains can subsequently be used to determine the mutations that occurred that may be associated with the observed phenotypes [8 , 9] . Adaptive laboratory evolution can be used to study the development of antibiotic resistance in bacterial pathogens [10] . Resistance to antibiotics is an evolutionary response of bacteria to withstand and survive the effects of the stressor . Deliberately evolving bacteria to withstand antibiotics through experimental evolution can yield insights into the evolutionary dynamics and trajectories of this adaptive process [11 , 12] . These evolution experiments can provide a longer-term perspective , which can yield information for the design of novel treatment strategies that can reduce the rate of resistance evolution or potentially even reverse the effects of resistance [13–15] . Recent studies have explored how adaptation to an antibiotic can cause bacteria to concurrently become more susceptible or more resistant to other drugs , an effect termed collateral sensitivity or collateral resistance [14 , 16 , 17] . Collateral sensitivities between drugs have been used to design drug cycling strategies and to explain the decreased rate of adaptation to certain antibiotics [12 , 14 , 18–23] . Drug deployment strategies that exploit such collateral sensitivities between pairs of antibiotics to minimize resistance evolution have been tested in vitro . A recent study determined the collateral sensitivity drug interaction network in Escherichia coli and demonstrated how an alternating sequential treatment of 2 reciprocal , collaterally sensitive antibiotics can slow down the rate of resistance evolution [14] . In this drug cycling strategy , the development of resistance to 1 drug concurrently increased the sensitivity to the second drug , and this allowed wild-type cells to outcompete the resistant cells when exposed to the second drug . In a different study , evolution experiments of alternating sequential therapies of pairs of antibiotics were performed on Staphylococcus aureus , and the study showed that the alternating treatments slowed the rate of resistance development compared to single-drug treatments [12] . Consistent with the E . coli study , this study found that collateral sensitivities could explain the evolutionary constraints in the cases in which alternating treatment resulted in decreased resistance development compared to the single-drug treatment . Most of the prior studies that test the use of alternating antibiotic therapies to reduce the rate of resistance development employ an adaptive laboratory evolution scheme in which the drugs are switched at daily or subdaily intervals with the purpose of testing if rapidly changing antibiotic environments can diminish the rate of drug-resistance adaptation [12 , 19 , 23 , 24] . In this study , we expand on these prior works , but we are not focused on studying the evolutionary dynamics of bacteria adapted to rapidly changing drug environments . Rather , we explore the evolutionary dynamics of sustained , longer treatments of drugs and how the development of high levels of resistance to 1 drug influences the subsequent dynamics of sustained adaptation to a second drug . In clinical settings , when antibiotic cycling strategies are employed , they are used typically at the level of the hospital ward , and the cycling of antibiotics are often done at monthly intervals [25 , 26] . The rationale here is that if resistance to 1 drug arises after frequent use in a ward , switching to an antibiotic of a different class may allow resistance rates to the withdrawn drug to stabilize or even fall , enabling the first drug to be efficiently reintroduced again at a later time [27] . This practice of cycling drugs of different classes over the course of monthly intervals is done empirically , and it remains unclear how these regimens constrain the evolutionary dynamics of antibiotic resistance development . Here , we explore the evolutionary trajectories of bacteria as they evolve high levels of resistance to 1 antibiotic and the subsequent trajectories as the selection pressure from the first drug is withdrawn and replaced with the sustained pressure of a different drug . It remains unexplored how prior adaptation to 1 drug environment affects the evolutionary dynamics of a bacterial population during subsequent adaptation to a second drug in terms of the amount of resistance it can potentially develop and the resistance profile of the first drug . Collateral sensitivities and collateral resistances between 2 drugs have been studied in the context of adaptation to single drugs [21] , i . e . , as bacterial populations evolve and become resistant to 1 drug , do the cultures concurrently become more resistant or sensitive to other drugs ? In this study , we focus not on if bacteria become concurrently more resistant or sensitive to other drugs , but rather , if adaptation to 1 drug constrains or potentiates the evolutionary dynamics to sustained adaptation to a second drug . How does the history of adaptation to 1 drug influence the subsequent adaptation to a second drug ? If there are such historical dependencies , can we use this knowledge to design sequential therapies that slow down the evolution of resistance to the drugs used ? What happens to the previously developed resistance once the drug pressure is taken away or switched to a different drug ? Do compensatory adaptations sustain the high resistance , or do the bacteria revert and become susceptible again [28] ? The answers to these questions are important for understanding how bacteria adapt to different antibiotic environments . Bacterial pathogens have complex evolutionary histories and elucidation of any historical dependencies of antibiotic resistance evolution would allow for rational forecasting of future resistance development and would aid in the design of strategies for mitigating antibiotic resistance .
To test how different antibiotic-resistance backgrounds affect the subsequent adaptation dynamics when evolved to a new antibiotic , we used a laboratory evolution approach to evolve P . aeruginosa to all 2-drug sequences of the 3 clinically relevant drugs piperacillin ( PIP ) , tobramycin ( TOB ) , and ciprofloxacin ( CIP ) . In each of the experimental sequences , P . aeruginosa was subjected to 20 days of adaptation to each drug by serially passaging parallel replicate cultures to increasing concentrations of the drugs followed subsequently by 20 more days of adaptation to each of the 3 drugs or to lysogeny broth ( LB ) media without a drug ( Fig 1A ) . Additional parallel replicates were adapted to LB media without a drug for 40 days as a control . For each drug treatment , changes in the resistance to the other 2 drugs were concurrently measured ( Fig 1B ) . Minimum inhibitory concentration ( MIC ) gradients in microtiter plates were used to simultaneously measure the drug resistance level and to propagate the bacteria daily . To adapt the bacteria to a drug , a sample is taken from the population from the well of the highest drug concentration that allowed for growth ( i . e . , MIC/2 ) and then used to inoculate a new MIC gradient . This serial dilution cycle is done daily . More explicitly , 20 μl of culture is sampled from the well of the highest concentration that allowed for growth , then diluted in 5 ml of fresh LB media , and then this diluted culture is used to inoculate a new MIC gradient . This dilution protocol results in a daily dilution factor of the bacterial population of 1/500 ( Materials and methods , Fig 1B ) . S1 Data provides the estimated number of generations per day for the evolved lineages based on the daily measurements of the OD600 . For each lineage , the OD600 values are fairly consistent from day to day , and so with a dilution factor of 1/500 , the cultures undergo approximately 9 generations of growth per daily dilution cycle ( S1 Data ) . We observed differences in final resistance levels to the different drugs depending on the history of past treatments ( or lack of treatments ) , an effect we call drug-order–specific effects of adaptation . Our results show that a history of past drug adaptation can affect the rate at which resistance can potentially arise when subsequently adapted to a new antibiotic . Furthermore , in some cases , adaptation to a second drug or to LB can partially or fully restore sensitivity to the first drug . These observations suggest that in order to limit the rate of development of antibiotic resistance , it is important to consider which drugs a bacterial population may have been exposed to in the past when choosing which drugs to subsequently deploy . The 3 drugs tested have different mechanisms of action and are clinically used to treat P . aeruginosa infections [29] . Piperacillin is a beta-lactam that inhibits cell wall synthesis [30]; tobramycin is an aminoglycoside that binds to the prokaryote ribosome and inhibits protein synthesis [31]; and ciprofloxacin is a fluoroquinolone that binds DNA gyrase and inhibits DNA synthesis [32] . We chose to study these 3 antibiotics because of their common use in the clinical setting to treat P . aeruginosa infections [29] , their diverse mechanisms of action , and their well-studied resistance mechanisms [33] . Adaptive evolution for 20 days to these drugs individually resulted in 1-drug–resistant mutants denoted as PIPR , TOBR , and CIPR . The Day 20 PIPR , TOBR , and CIPR populations had averages of 32- , 64- , and 64 times higher MICs to piperacillin , tobramycin , and ciprofloxacin , respectively , compared to their initial levels . To determine if the population MICs that were measured during the course of the adaptive laboratory evolution experiments were representative of individual colony MICs , we retrospectively measured the MICs of cultures grown from multiple revived colonies from the saved frozen stocks ( S2 Data ) . Overall , 73% of the retrospectively measured MICs were within one 2-fold dilution step of the originally measured population MICs , which suggests that the reported MIC values for each of the evolved lineages are well representative of the bacterial populations ( S1 Fig ) . By following how the resistance to each of the 3 drugs changes for each of the drug sequences ( Fig 2; S2 and S3 Figs and S3 Data ) , we observed 3 types of drug-order–specific effects in the MIC profiles ( Fig 3 ) . Note that for now , we focus on summarizing the different drug-order–specific effects ( as seen by the changes in drug MICs ) , and later , we discuss several hypotheses for the underlying mechanisms of the drug-order–specific effects based on analysis of the genomic mutations of the adapted lineages . In the first type of drug-order–specific effects , adaptation to a second drug or to LB restores the susceptibility to the first drug ( Fig 3A ) . In these experiments , we were first interested to see if the increases in MICs of the 1-drug–resistant lineages ( Day 20 PIPR , TOBR , and CIPR ) were permanent or transient . By evolving them to LB and hence removing the selection pressure of the drug for 20 days , we observed that the high MICPIP was maintained in Day 40 PIPRLB ( Fig 3A [top] , p = 0 . 80; Fig 2A ) , while MICTOB declines ( leading to partial resensitization ) in Day 40 TOBRLB ( Fig 3A [middle] , p < 0 . 0001; Fig 2E ) , and MICCIP declines ( although not significantly ) in Day 40 CIPRLB ( Fig 3A [bottom] , p = 0 . 18; Fig 2I ) . Thus , for these 3 treatments , removal of the antibiotic pressure can maintain the high resistance or lead to resensitization in a drug-specific manner . Similar trends were seen in a recent adaptive evolution study whereby P . aeruginosa was evolved to tobramycin , ciprofloxacin , piperacillin/tazobactam , meropenem , and ceftazidime , followed by subsequent adaptation in the absence of the drug ( growth medium only ) to determine the effects of removing the drug selection pressure [34] . Similar to the patterns seen in our study , they observed that the tobramycin-resistant cultures partially resensitized , the ciprofloxacin-resistant cultures had a modest resensitization , and the 3 beta-lactam–evolved cultures maintained high levels of resistance . Next , we were interested to see if evolving the 1-drug–resistant lineages to the other 2 drugs would show the same patterns seen as when evolved to LB . Interestingly , we saw unique outcomes for each of the 3 lineages . When Day 20 PIPR was evolved to tobramycin , the MICPIP of Day 40 PIPRTOBR remained high ( p = 0 . 90 ) , similar to how the MICPIP of Day 40 PIPRLB remained high ( Fig 3A [top] ) . This result suggests that subsequent tobramycin adaptation has no role in altering the high piperacillin resistance . This specific order of drug treatments can then result in multidrug-resistant P . aeruginosa cultures that are resistant to both piperacillin and tobramycin . On the other hand , when Day 20 PIPR was evolved to ciprofloxacin , the resulting cultures became resensitized to piperacillin ( Fig 3A [top] , p < 0 . 05 ) , and the MICPIP declined to levels comparable to those of the initially susceptible cultures ( MICPIP of Day 1 PIPR versus Day 40 PIPRCIPR , p = 0 . 80 ) , indicative of a full resensitization . Since resensitization did not occur after subsequent adaptation to tobramycin or LB , we suspect that the subsequent ciprofloxacin adaptation had an active role in the resensitization to piperacillin in such a way that tobramycin and LB did not . These results show that if a piperacillin-resistant culture ( that is also sensitive to tobramycin and ciprofloxacin ) is evolved to tobramycin , multidrug resistance can occur . However , if it is evolved to ciprofloxacin , despite the fact that ciprofloxacin resistance increases , the culture becomes susceptible to piperacillin again , making piperacillin a potentially rational choice for further treatment . When Day 20 TOBR was evolved to ciprofloxacin , partial resensitization occurred ( MICTOB of Day 20 TOBR versus Day 40 TOBRCIPR , p < 10−5 ) , and the MICTOB of Day 40 TOBR-CIPR fell to a comparable level as that of Day 40 TOBRLB ( p = 0 . 98 ) ( Fig 3A [middle] ) . This result suggests that the resensitization seen during the subsequent ciprofloxacin adaptation is not caused by the selection pressure of ciprofloxacin , but rather by the absence of the selection pressure of tobramycin . On the other hand , evolving Day 20 TOBR to piperacillin also led to a partial resensitization ( MICTOB of Day 20 TOBR versus Day 40 TOBRPIPR , p < 0 . 05 ) but not as much as it did when Day 20 TOBR was evolved to ciprofloxacin ( MICTOB of Day 40 TOBRPIPR versus Day 40 TOBRCIPR , p < 0 . 01 ) and LB ( MICTOB of Day 40 TOBRPIPR versus Day 40 TOBRLB , p < 0 . 05 ) . Because of this difference , we suspect that the maintenance of the comparably high tobramycin resistance is a consequence of the piperacillin selection pressure , since we observed that adaptation in the absence of the drug pressure in LB led to substantially greater resensitization . This case highlights how the removal of all drug pressures may lead to the resensitization of the culture more than with the treatment of the culture to a new drug . In conjunction with the result that subsequent tobramycin adaptation of Day 20 PIPR still maintained a high MICPIP , this case then also shows how regardless of the order , sequential adaptation to piperacillin and tobramycin leads to multidrug resistance of the 2 drugs . Lastly , when Day 20 CIPR was evolved to piperacillin and tobramycin , both treatments lead to a partial resensitization to ciprofloxacin ( Fig 3A [bottom] ) . During subsequent tobramycin adaptation , the decrease in the MICCIP from Day 20 CIPR to Day 40 CIPRTOBR ( p < 0 . 01 ) was marginally more than the decrease in the MICCIP from Day 20 CIPR to Day 40 CIPRPIPR ( p < 0 . 05 ) during subsequent piperacillin adaptation . As mentioned above , subsequent adaptation of Day 20 CIPR to LB led to a decrease in MICCIP that was not statistically significant; however , we argue that the decrease is comparable to that seen when adapted to piperacillin and tobramycin as the final MICCIP of Day 40 CIPRLB was not significantly different than that of Day 40 CIPRPIPR ( p = 0 . 93 ) and that of Day 40 CIPRTOBR ( p = 0 . 53 ) . Hence , in this case , evolution of a ciprofloxacin-resistant culture to either a different drug or to a no-drug condition led to a partial resensitization of ciprofloxacin . Interestingly , we also observed that the resensitization that occurred during subsequent piperacillin adaptation happened more quickly than the resensitization that occurred during subsequent tobramycin and LB adaptation ( Fig 2I ) . After 5 days of subsequent piperacillin adaptation ( Day 25 CIPRPIPR ) , the MICCIP was significantly different than that of Day 20 CIPR ( p < 0 . 001 ) , while this was not the case after 5 days of subsequent tobramycin ( p = 1 . 00 ) or LB ( p = 0 . 57 ) adaptation . These cases in which partial or full resensitization to the first drug occurs after adaptation to a second drug or LB highlight opportunities in which resistance to 1 drug can potentially be reversed by treating with a second drug or by removing the drug pressure completely . In the second type of drug-order–specific effects , prior adaptation to a first drug reduces the rate of subsequent adaptation to a second drug ( such that the endpoint level of resistance to that second drug is lower compared to the amount of resistance developed when the Day 0 Ancestor is directly evolved to that second drug ) . We observed that evolution first to piperacillin reduces the rate of subsequent evolution to tobramycin ( Fig 2D and 2E ) . That is , the MICTOB of Day 40 PIPRTOBR was less than that of Day 20 TOBR ( Fig 3B , p < 0 . 05 ) . This observation suggests that , in some cases , different bacterial populations may evolve resistance to a given antibiotic at different rates depending on the history of prior adaptations that the populations have experienced . Having knowledge of prior adaptations may then potentially be used to slow down the development of resistance to a drug if that drug is selected rationally . Interestingly , we observed no cases in which prior drug adaptation led to an enhancement in the rate of adaptation to a second drug . The last type of drug-order–specific effects is when the final MIC of a drug is different after adaptation to a 2-drug sequence compared to after adaptation to the opposite order of the 2 drugs ( Fig 3C ) . This third type of drug-order–specific effect exists as a consequence of a combination of the first type of effect ( resensitization of the 1-drug–resistant lineages during subsequent adaptations to other drugs ) and specific cases of collateral sensitivities during the adaptation of certain lineages . First , the MICPIP was higher when piperacillin was used after ciprofloxacin ( Day 40 CIPRPIPR ) compared to when piperacillin was used before ciprofloxacin ( Day 40 PIPRCIPR ) ( Fig 3C [top] , p < 0 . 05 ) . In this case , adaptation to piperacillin first led to high levels of piperacillin resistance , and subsequent adaptation to ciprofloxacin led to the resensitization to piperacillin as discussed before ( Fig 2A ) . On the other hand , even though adaptation to ciprofloxacin first led to a collateral sensitivity to piperacillin ( S4A Fig [right] , p < 0 . 01 ) , subsequent adaptation to piperacillin resulted in a final MICPIP comparable to that of Day 20 PIPR ( Fig 2C ) . Next , we observed that during the adaptation to tobramycin followed by ciprofloxacin and vice versa , the final MIC values of piperacillin and ciprofloxacin were different depending on the order of adaptation to the 2 drugs ( Fig 3C [bottom and middle] ) . With regards to the difference seen in the final MICCIP ( Fig 3C [bottom] , p < 0 . 05 ) , the partial resensitization to ciprofloxacin starting from Day 20 CIPR during subsequent tobramycin adaptation ( Fig 2I ) resulted in the MICCIP to be less than adaptation to tobramycin first , followed by ciprofloxacin ( Fig 2H ) . Finally , it was interesting that even though piperacillin was not the direct selection pressure , there was a difference in the final MICPIP level whether ciprofloxacin adaptation occurred after tobramycin adaptation or vice versa ( Fig 3C [middle] , p < 0 . 01 ) . In this case , initial adaptation to tobramycin first did not affect the MICPIP ( Fig 2B ) , but subsequent adaptation to ciprofloxacin resulted in a collateral sensitivity to piperacillin ( S4C Fig , p < 0 . 01 ) . On the other hand , as previously mentioned , adaptation to ciprofloxacin first initially resulted in the collateral sensitivity to piperacillin ( S4A Fig [right] , p < 0 . 01 ) ; however , the MICPIP returned to baseline values during subsequent adaptation to tobramycin ( Fig 2C ) . Thus , regardless if ciprofloxacin adaptation occurred before or after tobramycin adaptation , ciprofloxacin adaptation led to piperacillin collateral sensitivity . However , in order to take advantage of this collateral sensitivity , ciprofloxacin adaptation should be used after tobramycin adaptation , rather than vice versa . In a contrasting example , we also found it interesting that while ciprofloxacin adaptation also led to collateral sensitivity of tobramycin , subsequent piperacillin adaptation did not cause the MICTOB to return to baseline levels ( Fig 2F ) in the manner in which subsequent tobramycin adaptation returned the MICPIP to baseline values ( Fig 2C ) . Altogether , these cases highlight how treating an infection with a sequence of 2 drugs can result in different resistance profiles depending on the order used . All the cases of collateral sensitivity that were observed occurred during ciprofloxacin treatment whereby ciprofloxacin adaptation resulted in a lower MIC of piperacillin or tobramycin compared to baseline levels ( S4 Fig ) . First , adaptation to ciprofloxacin starting from the Day 0 Ancestor resulted in collateral sensitivity to both piperacillin ( Fig 2C; S4A Fig [right] , p < 0 . 01 ) and tobramycin ( Fig 2F; S4A Fig [left] , p < 0 . 0001 ) . Next , adaptation to ciprofloxacin starting from both the 1-drug–evolved lineages Day 20 PIPR ( Fig 2D ) and Day 20 TOBR ( Fig 2B ) resulted in collateral sensitivity to tobramycin ( S4B Fig , p < 0 . 01 ) and piperacillin ( S4C Fig , p < 0 . 01 ) , respectively . These results suggest that regardless of historical background , ciprofloxacin adaptation results in collateral sensitivity to the other 2 drugs . While we observed that collateral sensitivity of other drugs occurs only during ciprofloxacin adaptation , a recent study in which P . aeruginosa ATCC 27853 was evolved to different antibiotics reported that evolution to tobramycin resulted in collateral sensitivity to piperacillin-tazobactam and ciprofloxacin , whereas we did not observe this effect [34] . Also , this study did not observe that adaptation to ciprofloxacin resulted in collateral sensitivity to piperacillin and tobramycin , as we reported here . We suspect that these inconsistences may be due to strain-specific differences in the different P . aeruginosa strains used ( strain PA14 was used in this study ) . We were interested in measuring the fitnesses of the evolved lineages to see if the adaptations to the different treatments altered their growth dynamics . We measured the growth curves ( OD600 ) of the 68 evolved replicate lineages as well as the Day 0 Ancestor in quadruplicate grown in LB for 24 hours ( S5 Fig ) . The exponential growth rates were subsequently calculated from the growth curves ( S6A Fig , S1 Text and S4 Data ) . While we observed many different growth rates across the different lineages , we did not observe any correlation between the growth rate and the change in MIC between the Day 20 1-drug–evolved lineages and the subsequent Day 40 lineages ( i . e . , altered growth rates could not explain the cases in which subsequent adaptation led to the maintenance of high resistance or resensitization to the first drug ) ( S6B Fig and S1 Text ) . We hypothesized that genomic mutations acquired during adaptive evolution contributed to the drug-order–specific effects observed in the MIC profiles . We sequenced genomes of the Day 0 Ancestor , Day 20 PIPR , TOBR , CIPR , and LB Control lineages and the Day 40 1-drug–and 2-drug–evolved lineages , as well as the LB Control lineages . Genome sequencing of the Day 20 and Day 40 mutants revealed a total of 201 unique mutations across the 56 samples consisting of 77 SNPs , 31 insertions , and 93 deletions ( Fig 4; S7 Fig , S1 and S2 Tables ) . The 77 SNPs were found within 49 genes . Two SNPs were synonymous , and 6 were intergenic . To test how representative the sequencing results were of the mutant populations , we used PCR and Sanger sequencing to test for the presence of specific mutations in multiple colonies of different lineages after reviving the lineages from the saved frozen samples . We used the primers from S3 Table to test for the presence of 1 mutation from 1 replicate of each lineage , with 4 colonies of each lineage . Overall , while there may be limited heterogeneity in the populations with respect to a few of the mutations , the large majority of the mutations were homogeneous in the populations and fixed within the lineages ( S1 Text ) . While some genes were mutated during evolution to all drugs , other mutations were drug-specific and were related to the drugs’ primary mechanisms of action as would be expected ( S4 Table ) . Genes encoding transcriptional regulators for multidrug efflux pumps were commonly mutated during evolution to all 3 drugs ( mexC , mexR , mexS , nalC , nalD , nfxB , parS ) [35] . Ribosomal proteins ( rplJ , rplL , rpsL , rplF ) [36] and NADH dehydrogenase subunits ( nuoB , nuoG , nuoL , and nuoM ) [37 , 38] were frequently mutated during tobramycin evolution . The most commonly mutated gene was fusA1 , which encodes elongation factor G and was mutated in 11 different replicate lineages adapted to tobramycin . fusA1 has been observed to be mutated in clinical isolates of P . aeruginosa [39–41] , as well as in adaptive evolution studies to aminoglycosides in P . aeruginosa [34] and E . coli [12 , 16 , 18] . Mutations in fusA1 may also contribute to altered intracellular ( p ) ppGpp levels , which may modulate virulence in P . aeruginosa [41] . Mutations in gyrA and gyrB were observed during ciprofloxacin evolution , but none were observed in parC and parE ( the other genes of the quinolone resistance-determining region [29] ) . Lastly , genes encoding peptidoglycan synthesis enzymes ( dacC , mpl ) and beta-lactamase regulators ( ampR ) were mutated during piperacillin treatment . Many of these genes have also been observed to be mutated during human host adaptation of P . aeruginosa [42] , highlighting the importance of several of these clinical resistance determinants ( S1 Text ) . We next analyzed the genomic mutations to see how the historical context affects which mutations occur during adaptation to a drug . For example , how do the mutations that occur during adaptation to piperacillin only ( Day 20 PIPR and Day 40 PIPR ) compare to the mutations that occur during piperacillin adaptation when there is a prior history of adaptation first to tobramycin ( Day 40 TOBRPIPR ) or ciprofloxacin ( Day 40 CIPRPIPR ) ? To this end , we first categorized the genes in which mutations occurred into 23 broad categories based on the available literature and on the PseudoCAP functional classifications from the Pseudomonas Genome Database [43] ( Table 1 ) . Next , for each lineage , we tallied the number of times a gene in a functional category was mutated across the 4 biological replicates for each of the lineages ( Fig 5 ) . For a complete list of genes in each functional classification and descriptions of the genes , see S2 Table . We observed several general trends in the genes mutated during adaptation to the 3 drugs , depending on their historical context . In the lineages adapted to piperacillin , we saw history-dependent trends in the mutated genes that were related to multidrug efflux pumps ( Fig 5 , dashed-black box ) . While all the piperacillin-adapted lineages had mutations in genes related to the MexAB-OprM efflux pump ( which is the primary efflux pump of piperacillin [44] ) such as nalD and mexR ( whose products repress the expression of mexAB-oprM [45] ) , the Day 40 CIPRPIPR lineage had additional mutations in the structural subunit genes of the other efflux pumps MexCD-OprJ ( mexC ) and MexEF-OprN ( mexF ) . Lastly , no mutations in genes related to the MexXY-OprM pump were observed in any of the piperacillin-adapted lineages . With regard to adaptation to piperacillin only , most of the mutations that occurred in genes related to MexAB-OprM occurred within the first 20 days , with only a few additional mutations occurring between Day 21 and 40 . Regardless of historical context , metabolic and cell wall genes tended to be frequently mutated in piperacillin-adapted lineages , whereas metabolic and cell wall genes did not seem to be consistently mutated across the tobramycin-adapted and ciprofloxacin-adapted lineages . This result is perhaps due to the fact that the primary target of piperacillin is cell wall ( peptidoglycan ) synthesis , which is largely a metabolic process . Interestingly , we also observed that the lineages adapted only to piperacillin ( Day 20 PIPR ) sustained large chromosomal deletions that were not seen in the lineages in which there was prior tobramycin or ciprofloxacin adaptation ( Day 40 TOBRPIPR and Day 40 CIPRPIPR ) . We discuss and explore the potential implications of these large deletions below . The tobramycin-adapted lineages consistently had mutations occur in ribosomal subunit genes and other ribosomal machinery genes , regardless of historical context . In the lineages adapted only to tobramycin , mutations in genes related to the ribosome , membrane , energy , and NADH dehydrogenase tended to occur by Day 20 , followed by mutations in efflux pump–related genes by Day 40 . The mutations in genes related to membrane , NADH dehydrogenase , and energy likely reflect the unique requirement of the proton-motive force for the uptake of aminoglycoside antibiotics [46] , and the mutations occurring during tobramycin adaptation may contribute to the resistance by reducing the proton-motive force [16] . While we observed mutations in the NADH dehydrogenase genes in the lineages adapted only to tobramycin , we saw no such mutations in the lineages in which prior piperacillin or ciprofloxacin adaptation occurred ( Day 40 PIPRTOBR and Day 40 CIPRTOBR ) . Also , while efflux pump–related genes were mutated in the Day 40 TOBR and Day 40 CIPRTOBR lineages , no such mutations were seen in the Day 40 PIPRTOBR lineages in which prior adaptation to piperacillin occurred ( Fig 5 , dashed-purple boxes ) . The mutations in the ciprofloxacin-adapted lineages were fairly consistently distributed regardless of historical context . For all ciprofloxacin-adapted lineages , mutations were seen in genes related to DNA/RNA synthesis as expected , as well as in genes related to membrane , flagella , efflux pumps , metabolism , and transcriptional regulators . Mutations related to the MexAB-OprM , MexCD-OprJ , and MexEF-OprN efflux pumps ( mostly in genes encoding negative regulators of the pumps ) are seen in the ciprofloxacin-adapted lineages , reflecting the ability of these different pumps to extrude ciprofloxacin; however , no mutations were seen in genes related to MexXY-OprM , even though this pump is also known to contribute to fluoroquinolone resistance [44] . Further experiments in measuring the gene expression of the different efflux pumps may help elucidate the roles that these pumps play in contributing to the different drug-order–specific effects . Next , we sought to determine if the patterns in mutated genes could explain the mechanisms of some of the drug-order–specific effects that were observed in the MIC time courses . We first discuss the cases of resensitization or maintenance of high resistance in which the 1-drug–evolved lineages were subsequently adapted to the other 2 drugs or to LB ( Fig 3A ) . While subsequent adaptation of Day 20 PIPR to LB and tobramycin maintained high piperacillin resistance , subsequent adaptation to ciprofloxacin led to full resensitization to piperacillin ( Fig 3A [top] ) . We hypothesize that these differences stem from the different efflux pump-related genes that were mutated in these lineages ( Fig 5 , dashed-purple boxes ) . Evolution of the Day 0 Ancestor to piperacillin resulted in 2 different SNPs in nalD , and 1 SNP in mexR across the 4 biological replicates of Day 20 PIPR , likely leading to the overexpression of the MexAB-OprM efflux pump [45] . We suspect that MICPIP remained high during subsequent adaptation to LB and tobramycin due to continued overexpression of MexAB-OprM . However , when Day 20 PIPR was adapted to ciprofloxacin , several mutations occurred in genes related to other efflux pumps , including 1 in mexA , 2 in nfxB , and 2 in mexS ( Fig 5 , dashed-purple boxes ) . In particular , mexS encodes a negative regulator of the expression of MexEF-OprN , and mutations in this gene likely lead to the overexpression of the efflux pump [47] . Interestingly , expression of MexEF-OprN has been observed to correlate inversely with the expression of MexAB-OprM [47 , 48] . Hence , we suspect that the resensitization to piperacillin when Day 20 PIPR was subsequently adapted to ciprofloxacin may be have been due to a concurrent decrease in MexAB-OprM expression ( leading to reduced piperacillin efflux ) as MexEF-OprN expression increased . That is , it is possible that the mutations that occurred during ciprofloxacin adaptation that led to the overexpression of MexEF-OprN negated the effects of the mutations that occurred during prior piperacillin adaptation that led to overexpression of MexAB-OprM . Furthermore , we observed no mutations in efflux pump–related genes in Day 40 PIPRTOBR ( Fig 5 , dashed-purple boxes ) , which supports the notion that because no mutations occurred , which would have negatively correlated with the expression of MexAB-OprM , expression of this efflux pump was maintained throughout the subsequent adaptation to tobramycin , and hence the MICPIP remained high . We observed that subsequent adaptation of Day 20 TOBR to LB and ciprofloxacin resulted in a partial resensitization to tobramycin , and that while subsequent adaptation to piperacillin also led to a significantly lower MICTOB , it was not as low as that of Day 40 TOBRLB and TOBRCIPR ( Fig 3A [middle] ) . In this case , the partial resensitization during subsequent adaptation to LB may be attributable to adaptive resistance of aminoglycosides in P . aeruginosa . Adaptive resistance is a phenomenon in which resistance to a drug is transiently induced in the presence of the drug , and resistance recedes upon the removal of the drug [49] . In contrast to acquired resistance , which is mediated through genetic mutations , adaptive resistance is explained by phenotypic alterations that allow for temporary increases in resistance . P . aeruginosa is known to exhibit adaptive resistance to aminoglycosides [50 , 51] , and it is primarily mediated through up-regulation of MexXY-OprM during drug exposure and subsequent down-regulation after the removal of the drug [52] . We suspect that the partial resensitization during subsequent ciprofloxacin adaptation is also a consequence of adaptive resistance once the tobramycin selection pressure is removed . We further speculate that during the initial adaptation to tobramycin , the increase in tobramycin resistance was a combination of adaptive resistance and acquired resistance from accumulation of the mutations as seen in Day 20 TOBR . Thus , the resensitization during subsequent LB and ciprofloxacin adaptation was not a full resensitization but rather a partial one , perhaps reflecting the remaining contribution of the acquired resistance . Lastly , with regards to Day 40 TOBRPIPR , it is unclear how subsequent piperacillin adaptation seemingly resulted in the maintenance of high MICTOB compared to that of Day 40 TOBRLB and TOBRCIPR . We hypothesize that the subsequent piperacillin adaptation somehow counteracted the resensitization effects of adaptive resistance , even when the tobramycin selection pressure was removed . The mechanism of ciprofloxacin resensitization when Day 20 CIPR was subsequently adapted to LB , piperacillin , and tobramycin remains unclear ( Fig 3A [bottom] ) . While reversion of aminoglycoside sensitivity has been the most characterized case of adaptive resistance in P . aeruginosa , other studies have suggested that adaptive resistance may be prevalent in other classes of antibiotic classes as well , and that it may be mediated by epigenetic processes such as methylation and stochastic gene expression [53] , particularly affecting the expression of efflux pumps [54] . It could be possible that adaptive resistance partially explains the resensitization to ciprofloxacin . We also note that qualitatively , there was much more variability in the MIC time courses between the individual replicates of the CIPR lineages , as seen by the larger error bars in Fig 2I , compared to that of the PIPR ( Fig 2A ) and TOBR ( Fig 2E ) lineages . Taken together , further investigation of the partial ciprofloxacin resensitization is needed . While we observed clear cases of collateral sensitivity develop to piperacillin and tobramycin during the course of ciprofloxacin adaptation ( S4 Fig ) , other adaptive evolution studies of P . aeruginosa evolved to ciprofloxacin showed mixed results . In one study , the adaptation of P . aeruginosa ATCC 27853 to ciprofloxacin showed no change in the MIC of 3 different beta-lactams ( including piperacillin-tazobactam ) , nor of tobramycin [34] . In another study , while no statistical significances were assigned , adaptation of P . aeruginosa PAO1 to ciprofloxacin appeared to result in slight collateral sensitivities to piperacillin-tazobactam and tobramycin in some of their replicates . Nevertheless , in our study , we hypothesize that the collateral sensitivity to piperacillin and tobramycin during ciprofloxacin adaptation is attributable to the mutations seen in nfxB ( which encodes a transcriptional repressor that regulates MexCD-OprJ [55] ) in the Day 20 CIPR lineages . Three of the Day 20 CIPR replicates had deletions in nfxB ( 15 , 13 , and 16 base pairs ) , likely resulting in the inactivation of NfxB and concomitant up-regulation of MexCD-OprJ and increased ciprofloxacin resistance [56] . In fact , nfxB mutants have been reported to be hypersusceptible to certain beta-lactams and aminoglycosides [57 , 58] . Lastly , with regards to the decreased rate of tobramycin adaptation given a history of prior piperacillin adaptation ( Fig 3B ) , we attribute this effect to the large chromosomal deletions that were sustained in 3 of the 4 Day 20 PIPR replicates . The consequences of these deletions are discussed in more detail in the subsequent sections of the manuscript . In summary , based on the genomic mutations , we have presented our interpretations of potential mechanisms that contribute to the drug-order–specific effects . These include how historical context can influence the frequency of mutations in certain genes , the varying contributions of adaptive and acquired resistance to total resistance , and specific cases of inverse correlation of the expression of different efflux pumps . While mutations are likely not the sole determinants of the differences [34 , 59] , many of the observed genomic mutations can partially explain the drug-order–specific effects . One striking mutation we observed was that 3 of the 4 replicates of Day 20 PIPR ( Day 20 PIPR-1 , -2 , and -3 ) had large , approximately 400 kbp deletions ( corresponding to approximately 6% of the genome ) in a conserved region of the chromosome ( Fig 4 [large red rectangles]; S5 Data ) , suggestive of selective genome reduction [60–63] , and have been associated with directed repeats [64] and inverted repeats [65] at the boundaries of the deletions . These large deletions were also fixed in the corresponding Day 40 PIPRTOBR , Day 40 PIPRCIPR , and Day 40 PIPRLB lineages . Interestingly , the 3 PIPR lineages with these large deletions hyperproduced the brown pigment pyomelanin during piperacillin evolution , and this visually observable phenotype also persisted when evolved to tobramycin ( PIPRTOBR ) , ciprofloxacin ( PIPRCIPR ) , and LB ( PIPRLB ) . The loss of hmgA as part of the large chromosomal deletions correlates exactly with the pyomelanin phenotype of these lineages . Indeed , hmgA mutants of P . aeruginosa hyperproduce pyomelanin [66] . This observation shows that evolving to piperacillin results in a high probability of sustaining large deletions spanning hmgA , which results in the pyomelanogenic phenotype . However , when we evolved the Day 20 TOBR and CIPR lineages to piperacillin to yield the Day 40 TOBRPIPR and Day 40 CIPRPIPR lineages ( 4 replicates each ) , none of them became pyomelanogenic , suggesting that prior history of tobramycin or ciprofloxacin adaptation leads to a lower propensity of becoming pyomelanogenic when subsequently evolved to piperacillin . Interestingly , 1 of the Day 20 TOBR replicates became pyomelanogenic when subsequently evolved to ciprofloxacin ( Day 40 TOBRCIPR-2 ) . Hence in this study , pyomelanin hyperproduction is a consequence of piperacillin and ciprofloxacin evolution , yet the likelihood to evolve this visually striking and observable phenotype depends on the history of prior drug adaptation . While the 3 PIPR lineages that produced pyomelanin were not significantly more resistant to piperacillin than the nonpyomelanogenic PIPR lineage , pyomelanin-producing strains have been observed clinically [60] and have been shown to be more persistent in chronic lung infection models [66] . We tested the reproducibility of this example of a phenotypic dependence on the history of drug adaptation with a higher throughput approach . Starting with clonal populations of Day 0 Ancestor , Day 20 TOBR , and Day 20 CIPR , we seeded 92 replicate populations of each lineage into microplates , and we used a 96-pin replicating tool to serially propagate these populations and evolve them to increasing concentrations of piperacillin daily . The lineages that started from Day 0 Ancestor had the highest propensity to become pyomelanogenic ( Fig 6A ) compared to lineages starting from Day 20 TOBR ( Fig 6B ) or Day 20 CIPR ( Fig 6C ) . Still , certain lineages starting from Day 20 TOBR and Day 20 CIPR did also produce pyomelanin , albeit with less propensity than starting from Day 0 Ancestor ( Fig 6D; S8–S10 Figs ) . To explore the relevance of our laboratory evolution results clinically , we tested for the drug-order–specific MIC evolutionary dynamics in clinical isolates . We first tested the evolutionary dynamics of clinical isolates that were resistant to piperacillin but susceptible to tobramycin and ciprofloxacin . We evolved 3 piperacillin-resistant clinical isolates of P . aeruginosa to piperacillin , tobramycin , and ciprofloxacin for 10 days and tracked how the piperacillin resistance changed in these lineages . If the results from the adaptive evolution experiment applied to these piperacillin-resistant clinical isolates , then we would expect that evolving to tobramycin would not affect the high piperacillin resistance , but evolving to ciprofloxacin would restore susceptibility to piperacillin . As mentioned previously , evolving Day 20 PIPR to LB did not result in a reduction of MICPIP , which suggests that the resensitization to piperacillin when Day 20 PIPR was evolved to ciprofloxacin was a consequence of the switch to the ciprofloxacin drug pressure . Of the 3 isolates we tested , the evolutionary dynamics of 2 of these isolates matched these expectations ( Fig 7; S11 Fig and S3 Data ) . After normalizing to Day 1 MIC values , the MICPIP after 10 days of ciprofloxacin adaptation was significantly less than the MICPIP after 10 days of LB adaptation in isolate #2 ( Fig 7B , p < 0 . 05 ) and in isolate #3 ( Fig 7C , p < 0 . 001 ) , indicating resensitization to piperacillin during ciprofloxacin adaptation . This observation suggests that the MIC evolutionary dynamics we observed are not limited to laboratory strains of P . aeruginosa and may be observed in diverse strains of P . aeruginosa , including those originating from human patients . Note that these 3 clinical isolates were isolated from different patients , and their phylogenetic relatedness between each other and to the laboratory PA14 strain used in our study is untested . In isolate #1 , there was no significant difference in the normalized MICPIP values after 10 days of adaptation to tobramycin , ciprofloxacin , and LB ( Fig 7A , p = 0 . 237 , one-way ANOVA ) . Interestingly , this isolate evolved higher levels of piperacillin and ciprofloxacin resistance than the other 2 isolates ( S11 Fig and S3 Data ) , which suggests the possibility that adaptation to ciprofloxacin in these higher piperacillin-resistant cultures could still result in a restoration of piperacillin susceptibility . In the next set of evolution experiments , we investigated the role that the large chromosomal deletions play in a drug-order–specific effect . We had observed that compared to the Day 20 PIPR replicate that did not have a large deletion , the 3 Day 20 PIPR replicates with the large deletions , when subsequently evolved to tobramycin , developed less tobramycin resistance ( S3 Data and S12 Fig ) . This observation suggests that the large deletions were involved in reducing the subsequent rate of tobramycin resistance evolution given a prior history of piperacillin adaptation . A recent study isolated 4 pairs of clinical isolates of P . aeruginosa , in which each pair consisted of a pyomelanogenic ( PM ) isolate and a “parental wild-type ( WT ) ” nonpyomelanogenic isolate [64] . In each of the 4 pairs , the only genomic difference between the pyomelanogenic ( denoted APM , BPM , CPM , and DPM ) and its corresponding parental wild-type isolate ( denoted AWT , BWT , CWT , and DWT ) was the presence of large chromosomal deletions that overlap with parts of the deletions seen in Day 20 PIPR-1 , -2 , and -3 ( Fig 8E; S5 Data ) . Indeed , all of the large deletions encompass hmgA , whose loss accounts for the pyomelanin phenotype . We used these 4 pairs of clinical isolates to test the hypothesis that the large deletions play a role in lowering the rate of tobramycin resistance evolution . We evolved the 4 pairs of isolates to tobramycin using the same daily serial passaging technique used throughout this study and tracked the MICs of tobramycin , piperacillin , and ciprofloxacin over the course of 15 days ( Fig 8; S3 Data and S13 Fig ) . At the end of the 15 days , we saw that APM , BPM , and CPM had lower relative increases in MICTOB , compared to AWT ( p < 0 . 01 ) , BWT ( p < 0 . 05 ) , and CWT ( p < 0 . 05 ) , respectively ( Fig 8A–8C ) . These data provide support for the idea that the large chromosomal deletions do indeed play a role in reducing the rate of tobramycin adaptation , and potentially even in limiting the maximum level of tobramycin resistance that can be developed comparatively . In the case of the fourth pair , we saw that DWT and DPM had comparable increases in MICTOB over the course of the tobramycin adaptation ( Fig 8D , p = 1 . 00 ) . It can be speculated that some combination of the presence or loss of specific genes in DPM led to this evolutionary trajectory that is different from the other 3 pyomelanogenic isolates . We would also like to point out that within each pair , the “WT” and “PM” isolates vary in initial Day 1 MICTOB . The BPM and BWT pair was the most disparate pair , as BPM had a much lower MICTOB than BWT ( S13 Fig ) . Interestingly , a recent study also observed large genomic deletions spanning hmgA when P . aeruginosa PAO1 was evolved to meropenem , which is another beta-lactam antibiotic [65] . These mutants were also pyomelanogenic . The large deletions in both our study as well as this recent study also span mexX and mexY , which encode portions of the efflux pump that is a significant determinant of aminoglycoside resistance [67] . The loss of these genes in the 3 PIPR replicates may partially explain why subsequent tobramycin adaptation is limited compared to the replicate that did not sustain the large deletion .
This study presents evidence of how the evolutionary history of bacterial adaptation to antibiotics can complicate strategies for treating infections and for limiting the further development of multidrug resistance . Exposing bacteria to fluctuating environments has been shown to be a potentially good strategy for slowing down the development of resistance [12 , 19 , 68] . More broadly , mechanisms of memory and history dependence in bacterial systems are being uncovered to better understand the dynamics of bacterial survival and adaptation in changing environments [69–71] . For example , a recent study showed that the survival of Caulobacter crescentus in response to a high concentration of sodium chloride stress depended on the duration and timing of an earlier treatment of a moderate concentration of sodium chloride and that this effect was linked to delays in cell division , which led to cell-cycle synchronization [72] . Another study described what they call response memory , which is when a gene regulatory network continues to persist after the removal of its external inducer . The study showed that in E . coli , lac induction transiently continued when the environment was switched from lactose to glucose , which may be beneficial when the environment fluctuates over short timescales [73] . The results of those studies as well as the results from this study challenge the notion that bacteria respond solely to their present environment . Bacteria can encounter different stressors over time such as osmotic , oxidative , and acidic stress , and other studies have looked at how adaptation to 1 stressor protects the bacteria against other stressors [7 , 74] . Another example of bacteria adapting to changing environments is how P . aeruginosa , which can be found in the natural environment in the soil and water , can readily adapt to a human host under the right conditions and consequently become pathogenic [75] . There are several factors involved in the emergence of antibiotic resistance that are clinically important that are not considered in this study . We have not taken into account any pathogen/host interactions , such as the role of the immune system . We also do not take into consideration the pharmacokinetics of the drug and the time-dependent fluctuation of drug concentration as experienced by the bacteria in a human-host environment . Furthermore , the dosages of clinical regimens are typically much higher than the wild-type MIC , and the evolutionary dynamics of the bacteria under these conditions may be different from those seen in our study , in which the drug pressure is slowly increased over time . We neglect to consider the role of horizontal gene transfer , which is a common mechanism of antibiotic resistance transfer , and focus rather on the role of de novo mutations acquired during adaptation . Because of the nature of the serial passaging method , we may be selecting for fast growers that may not necessarily have mutations that confer the most amount of resistance in terms of the MIC . We used a strong selection pressure in this study by propagating from the highest concentration of drug that showed growth , but it has been shown that weak antibiotic selection pressure can greatly affect the adaptive landscape [76 , 77] . Lastly , these bacteria were evolved to 1 antibiotic at a time , and we do not know how different mutant lineages would adapt if competed against each other . It would be interesting in the future to conduct competition experiments to measure the fitness of the different lineages with respect to each other . While adaptive evolution of clinical isolates suggests that the drug-order–specific effects are clinically relevant , actual clinical studies must be performed to test the true clinical applicability of these effects . A major challenge that still needs to be addressed is how to translate the results of in vitro adaptive evolution experiments to effective therapies that can be used in a clinical setting [78] . For example , while we observed that piperacillin adaptation often led to the large chromosomal deletions and concomitant pyomelanin hyperproduction , the clinical isolates that had the large deletions ( Fig 8 ) were not necessarily resistant to piperacillin . On the other hand , the other set of clinical isolates , which did have resistance to piperacillin , did not have the large deletions ( Fig 7 ) . Disparities between the phenotypic and genotypic adaptations such as this will need to be studied further in terms of strain-specific differences , actual history of antibiotic exposure , and other factors that are beyond the scope of this study . Despite these caveats , there are several key factors of this study that provide confidence in the claims made . We saw consistency in the parallel replicates for the treatment lineages . An interesting exception is Day 40 PIPRTOBR-4 , which had a higher final tobramycin resistance compared to Day 40 PIPRTOBR-1 , -2 and -3 , which we believe is attributed to the large genomic deletions seen in the first 3 replicates but not in the fourth replicate . We observed parallel evolution in which several genes were mutated independently across multiple lineages , and overall , there were less than 15 mutations per 20 days of evolution , which suggests positive selection . Furthermore , many of the mutated genes are also observed in clinical isolates of P . aeruginosa , further giving credence to the clinical relevance of these mutations . As mentioned previously , studies that have looked at alternating treatments of antibiotics have primarily looked at the effects of rapid switching , typically at daily or subdaily intervals . One of such recent studies evaluated how E . coli responded to 136 different sequential treatments of subinhibitory concentrations of doxycycline and erythromycin , with each treatment consisting of 8 “seasons” of 12-hour-long adaptation periods to 1 of the drugs [19] . Using final optical density as an endpoint metric , the study found that 5 of the sequential treatments could clear the bacteria at the end of the eighth season . Interestingly , one of those 5 successful treatments consisted of 4 seasons of erythromycin , followed subsequently by 4 seasons of doxycycline . On the other hand , the treatment consisting of 4 seasons of doxycycline followed by 4 seasons of erythromycin did not manage to clear the bacteria at the end of 8 seasons . While the experimental setup is much different compared to that of this present study in terms of organism , antibiotics used , duration of treatment , and endpoint metric , these 2 treatments ( 4 seasons of erythromycin then 4 seasons of doxycycline and vice versa ) are quite analogous to the types of treatments tested in our present study . The fact that these authors found a difference in the outcomes of this pair of opposite sequential treatments may suggest that drug-order–specific effects similar to those presented in our study may play a role in the evolutionary dynamics of their experiments . Cycling between 2 drugs that exhibit collateral sensitivity to one another has been proposed and tested as a strategy to slow down the rate of resistance development [14] . Studies that have systematically tested for collateral sensitivities across a variety of antibiotics in E . coli have consistently discovered that when E . coli is adapted to drugs of the aminoglycoside class , it develops collateral sensitivity to several other drugs of different classes including beta-lactams , DNA synthesis inhibitors , and protein synthesis inhibitors [14 , 16 , 77] . In our present study , we tested 1 aminoglycoside ( tobramycin ) , and we did not observe any collateral sensitivity arise to piperacillin or ciprofloxacin during adaptation to tobramycin . Instead , we saw collateral sensitivity to piperacillin and tobramycin arise as P . aeruginosa was adapted to ciprofloxacin , which is a DNA synthesis inhibitor . While we only tested 1 drug in each of 3 drug classes , the dissimilarity of collateral sensitivity profiles between those studies and this present study may highlight how collateral sensitivity profiles may be organism-specific and drug-specific . Further supporting this idea , these prior studies also showed that while adaptation to drugs of the aminoglycoside class as a whole tended to lead to collateral sensitivity to other drug classes , not every aminoglycoside drug that was tested induced the same collateral sensitivity profiles . While we did observe cases of collateral sensitivity , the main focus of our study was not to look at how resistance profiles to other drugs concurrently change during adaptation to 1 drug , but rather to see how adaptation to 1 drug influences the future evolutionary dynamics as the resistant population adapts to a new drug . Additionally , we wanted to see how adaptation to the second drug affected the resistance profile of the drug that the bacteria originally developed resistance to . Our sustained drug adaptation scheme can be thought of as being more akin to month-long antibiotic cycling at the level of the hospital ward or the environments that bacteria in persistent chronic infections are exposed to . The history-dependent evolutionary dynamics presented in this study highlight the complexity of bacterial adaptation to multidrug therapies , serve as a framework for forecasting evolutionary trajectories based on genetic and phenotypic signatures of past adaptation , and ultimately help elucidate our fundamental understandings of the evolutionary forces that drive resistance adaptation . Asymmetrical evolutionary responses in changing environments have been studied in terms of collateral sensitivity/resistance [14 , 16] , temperature [79] , other abiotic stresses [7] , and in cancer treatments [80] . Here , we present the concept of drug-history–specific effects in multidrug resistance adaptation , whereby the history of adaptation to 1 antibiotic environment can influence the evolutionary dynamics during subsequent adaptation to another antibiotic environment . These history-specific effects have direct clinical implications on optimizing antibiotic treatment strategies to slow and prevent the emergence of dangerous multidrug-resistant bacterial pathogens .
The set of P . aeruginosa clinical isolates collected from the University of Virginia Health System ( presented in Fig 7 ) were deidentified , did not require Institutional Review Board approval for their use , and were anonymized . The Hocquet P . aeruginosa clinical isolates ( which were originally collected by the authors of the Hocquet study [64]; presented in Fig 8 ) also did not require Institutional Review Board approval for their use and were anonymized . We evolved , in parallel , 4 independent replicates for each lineage in the primary adaptive evolution experiment and 3 independent replicates for each of the clinical isolates to balance the statistical power of the conclusions with the technical feasibilities of the daily serial propagations . In the primary adaptive evolution experiment , we concluded the 1-drug evolution at the end of 20 days because the resistance levels of the evolved lineages to their respective drugs were saturated or close to saturated at that point . The clinical isolates from Fig 7 and from Fig 8 were evolved for 10 and 15 days , respectively , because the similarities and differences of the drug-specific effects to those of the primary adaptive evolution experiment were readily apparent at that point . MIC plates were made daily using the broth microdilution method with the standard 2-fold dilution series [81] . LB was used as the growth medium for all experiments ( 1% tryptone , 0 . 5% yeast extract , 1% NaCl ) . Antibiotics tested include piperacillin sodium , tobramycin , and ciprofloxacin HCl ( Sigma ) . Aliquots of 1 mg/ml and 10 mg/ml antibiotic stocks were made by diluting the antibiotic powders in LB and were stored at −20°C . New frozen drug aliquots were used on a daily basis . A frozen stock of P . aeruginosa PA14 was streaked on an LB agar plate , and a single colony was inoculated into 4 ml of LB , which was then grown overnight at 37°C , shaking at 125 RPM . This antibiotic-susceptible culture , denoted as the Day 0 Ancestor , was diluted to an OD600 of 0 . 001 ( approximately 106 CFU/ml ) and then inoculated into 3 identical MIC plates consisting of concentration gradients of piperacillin and tobramycin . A sample of the ancestor was saved in 25% glycerol and stored at −80°C . The 3 MIC plates were used to serially propagate cultures evolved to LB media , piperacillin , and tobramycin , with 4 biological replicates per condition . Wells for growth control ( media + culture ) and sterility control ( media ) were included in each MIC plate . For adaptation to LB media , bacteria were sampled from the growth control well . MIC plates were placed in a plastic container ( to prevent evaporation ) and incubated at 37°C with shaking at 125 RPM ( Thermo Scientific MaxQ 4000 ) . MIC plates were incubated daily for approximately 23 hours . At the end of incubation , growth in the MIC plates was determined using a plate reader ( Tecan Infinite M200 Pro ) . Growth was defined as OD600 > 0 . 1 after background subtraction . We recorded the MIC of each lineage for each drug , which was defined as the lowest antibiotic concentration tested that did not show growth ( S1 and S3 Data ) . To propagate , cultures were passaged from the highest concentration that showed growth ( i . e . , MIC/2 ) from the corresponding MIC drug gradient . For adaptation to LB , cultures were passaged from the growth control well that contained only LB without any drug . For each culture to be passaged , the culture was first diluted by a factor of 1/250 in fresh LB ( e . g . , 20 μl of the culture was diluted in 5 ml of LB ) , which was then inoculated in fresh piperacillin and tobramycin drug gradients in the new day’s MIC plate . Wells of the MIC plate thus contained 100 μl of double the final concentration of the antibiotic and 100 μl of the diluted culture . Hence , the cultures were diluted by a total factor of 1/500 daily . Daily samples were saved in 25% glycerol and stored at −80°C . For Day 21 , the piperacillin and tobramycin evolved cultures were subcultured in additional MIC plates such that they could subsequently be evolved to tobramycin and piperacillin , respectively . A similar protocol was used to establish the ciprofloxacin-evolved lineages ( CIPR ) . Starting with a clonal population of the Day 0 Ancestor , 4 replicates were established and propagated daily under ciprofloxacin treatment for 20 days . CIPR was then subpassaged to piperacillin and tobramycin to establish the CIPRPIPR and CIPRTOBR lineages in addition to continued ciprofloxacin evolution . To establish the PIPRCIPR and TOBRCIPR lineages , bacteria from the frozen stocks of Day 20 PIPR and TOBR were revived on LB agar plates , and clonal populations were evolved to ciprofloxacin to establish these lineages . Similarly , to establish the PIPRLB , TOBRLB , and CIPRLB lineages , bacteria from the frozen stocks of Day 20 PIPR , TOBR , and CIPR were revived on LB agar plates , and clonal populations were evolved to LB . Lastly , the MIC to ciprofloxacin was retrospectively measured for the Control , PIPR , TOBR , PIPRTOBR , and TOBRPIPR lineages . Frozen stocks were revived and plated on LB agar plates . The notation for the day numbering is such that Day X PIPR means X days exposure to piperacillin . For consistency , stocks were revived from Days 0 ( Ancestor ) , 5 , 10 , 15 , 19 , 20 , 25 , 30 , 35 , and 39 for Control , PIPR , and TOBR . One day of exposure to ciprofloxacin would yield Days 1 , 6 , 11 , 16 , 20 , 21 , 26 , 31 , 36 , and 40 MICs to ciprofloxacin . For PIPRTOBR and TOBRPIPR , stocks were similarly revived from Days 20 , 25 , 30 , 35 , and 39 and exposed to ciprofloxacin to measure Days 21 , 26 , 31 , 36 , and 40 MICs to ciprofloxacin . S3 Data shows the MICs to piperacillin , tobramycin , and ciprofloxacin , respectively , for all the lineages . Note that not all drug MICs were measured on a daily basis for all lineages . During analysis of the mutations , we deduced that there were some cross-contaminations between replicates in a few lineages . Namely , we saw sets of mutations that were identical in 2 replicates . We believed that the most likely explanation was that the following 7 lines were cross-contaminated sometime between Day 21 and Day 40: CIPRPIPR-3 , CIPRPIPR-4 , TOBR-1 CIPRTOBR-1 , CIPRTOBR-2 , CIPRTOBR-4 , and CIPR-3 , where the number denotes the replicate . To redo these lineages , the corresponding Day 20 replicate frozen stocks were revived on LB agar plates . Then clonal populations were used to redo the propagation as described before . For example , CIPR-3 was evolved to piperacillin for 20 days to redo CIPRPIPR-3 . We performed Sanger sequencing of replicate-specific mutations ( S3 Table ) on the Day 40 mutants to confirm successful propagation of the cultures . Frozen samples of Day 0 Ancestor , Day 20 Control , PIPR , TOBR , CIPR , Day 40 Control , PIPR , TOBR , CIPR , PIPRTOBR , PIPRCIPR , TOBRPIPR , TOBRCIPR , CIPRPIPR , and CIPRTOBR were streaked on LB agar plates and incubated at 37°C . Agar plates were submitted to Genewiz Incorporation for sequencing service . A single colony from each plate was chosen for DNA extraction , library preparation , multiplexing , and sequencing using 101-bp paired-end reads with the Illumina HiSeq 2500 platform . Reads were aligned to the reference P . aeruginosa PA14 genome ( NC_008463 . 1 ) with coverage ranging from 113X to 759X . This large range is due to the fact that we submitted samples for sequencing in 3 batches and had different numbers of samples for each batch but had relatively the same number of reads per batch . Nevertheless , the coverage was more than sufficient to identify the SNPs , insertions , and deletions in the genomes . The sequencing reads for Day 0 Ancestor and the 56 drug-evolved lineages are available via the NCBI SRA database ( www . ncbi . nlm . nih . gov/sra ) , accession number SRP100674 , BioProject number PRJNA376615 . Reads were aligned and mutations were called using the breseq pipeline [82] using default settings . All reported mutations were visually inspected by viewing the read alignments in IGV and the breseq output files , and mutations with less than 80% frequencies were not counted . The full list of mutations is presented in S1 and S2 Tables . The circos software package [83] was used to plot the mutations by genomic position for Fig 4 and the positions of the large chromosomal deletions in Fig 8 . We confirmed some of the mutations using Sanger sequencing . For each of the Day 20 PIPR , TOBR , and CIPR replicates , we chose 1 mutation each to confirm ( S3 Table ) . We also used these to confirm that replicates were not contaminated before submitting them for whole-genome sequencing . These mutations were also confirmed in each of the Day 40 lineages that were derived from the Day 20 PIPR , TOBR , and CIPR replicates . Clonal populations of Day 0 Ancestor , Day 20 TOBR-1 , -2 , -3 and -4 , and Day 20 CIPR-1 , -2 , -3 , and -4 were grown in LB starting from the frozen samples . These cultures were diluted in LB to OD600 of 0 . 001 . On Day 1 , in 96-well plates , 100 μl of the diluted cultures were inoculated with 100 μl of 4 μg/ml piperacillin ( to yield a final concentration of 2 μg/ml piperacillin ) . Ninety-two wells were used to establish independent replicate populations exposed to piperacillin . Cultures were incubated at 37°C with shaking at 125 RPM . On Day 2 , replicate populations were passaged using a 96-pin replicator tool ( V&P Scientific , VP246 , 100–150 μl per pin ) into 200 μl of 4 μg/ml piperacillin . This protocol was continued until Day 10 with a final concentration of 20 μg/ml piperacillin . For each plate , 2 wells were used as sterility controls ( only LB ) , and 2 wells were used as growth controls ( LB with bacteria , without drug ) . Photographs were taken daily ( S8–S10 Figs ) , and the number of visibly brown wells was recorded . Three clinical isolates of P . aeruginosa with high piperacillin resistance and low tobramycin and ciprofloxacin resistance were obtained from the University of Virginia Health System and were evolved to the 3 drugs in the same manner as the main adaptive evolution experiment starting from frozen samples . These isolates were first confirmed to actually be P . aeruginosa with PCR by using primers that specifically amplify the 16S rRNA region of P . aeruginosa [84] . Three replicates of each isolate were evolved to each of the 3 drugs for 10 days , and their MICs to the 3 drugs were measured as before . In separate subsequent experiments , the 3 clinical isolates were evolved to LB with 3 replicates each . The MICPIP was measured for 10 days ( S3 Data ) . This measurement was done by inoculating into piperacillin concentration gradients to measure the MICPIP but sampling and passaging from the “growth control” well ( LB with bacteria , without drug ) to adapt to LB . The 4 pairs of clinical isolates of P . aeruginosa from the Hocquet study [64] were evolved to tobramycin for 15 days with 3 parallel replicates each , with the exception of BPM , which had 2 replicates due to cross-contamination in the third replicate . The MICs for piperacillin and ciprofloxacin were also measured every 5 days ( S3 Data ) . At the end of the 15 days of evolution , primers amplifying part of the hmgA gene were used to check for the presence of the gene in the “WT” isolates and the absence of the gene in the “PM” isolates ( S3 Table ) . All statistical comparisons of MIC values were performed on the log2 transformed values . Unless noted otherwise , one-way ANOVAs were performed on the MICs of the relevant lineages . If the p-value from the ANOVA was less than 0 . 05 , a post-hoc Tukey’s honest significant difference ( HSD ) multiple comparisons test was then performed to determine which pairs of treatments were significantly different from each other . The Tukey’s HSD tests report 95% confidence intervals for the true mean difference for each pairwise comparison . If the confidence interval does not contain 0 , then the 2 groups being compared have significantly different means at the p = 0 . 05 level . To also assess the comparisons using nonparametric statistic tests , Kruskwal-Wallis tests followed by post-hoc Dunn’s multiple comparisons tests were also performed . All of the Kruskal-Wallis tests yielded comparable results to the one-way ANOVA at the alpha = 0 . 05 significance level , and the conclusions are the same for the key comparisons that drive the results highlighted in the manuscript . For a complete set of calculations , see S2 Text . For the comparisons presented in Fig 3 , treatments being compared consist of those listed on the x-axis of each graph in the figure . For the comparisons presented in Fig 7 , the raw MIC values for each lineage were first normalized by subtracting the average Day 1 MIC of each of their respective lineages . For each of the 3 clinical isolates , a one-way ANOVA and a Kruskal-Wallis test were performed on the Day 10 MICPIP values of the lineages evolved to LB , tobramycin , and ciprofloxacin ( piperacillin-adapted lineages were excluded in the comparisons ) . The Tukey’s HSD test and Dunn’s test were then performed to see if the Day 10 MICPIP values of the lineages evolved to tobramycin and ciprofloxacin were significantly different from the lineages evolved to LB . For the comparisons presented in Fig 8 , the raw MIC values for each lineage were first normalized by subtracting the average Day 1 MIC of each of their respective lineages . A two-sample t test and a Wilcoxon rank sum test were performed for the Day 15 MICTOB values of the “WT” and “PM” lineages evolved to tobramycin in each of the 4 pairs of isolates . Calculations were done in MATLAB R2016b , using the functions “anova1” for one-way ANOVA , “multcompare” for Tukey’s HSD test , “ttest2” for two-sample t test , and “ranksum” for the Wilcoxon rank sum test . The Kruskal-Wallis test was done with the “kruskal . test” command in R , and the Dunn’s test was done with the “posthoc . kruskal . dunn . test” command with the PMCMR R package [85] .
|
Bacteria readily adapt to their environments and can develop ways to survive and grow in the presence of antibiotics . While many studies have investigated how bacteria evolve to become resistant to single drugs , it is unclear how adaptation to other drugs and environments in the past affect the way bacteria adapt to new drugs and environments . In this study , we allowed bacteria in a laboratory setting to adapt to three different antibiotics . We first exposed wild-type susceptible bacteria to high concentrations of the three antibiotics individually and then exposed these populations to each of the other drugs . By tracking the levels of resistance to all three drugs in all of the treatments , we identified cases in which past adaptation to one treatment influenced subsequent evolutionary dynamics with regard to both phenotypes ( levels of resistance ) and genotypes ( genes that became mutated ) . Additionally , by allowing bacterial isolates originating from human patients to adapt to the three drugs , we recapitulated a subset of the adaptation history-dependent evolutionary dynamics . Overall , this study sheds light on how adaptation history in bacteria can potentiate or constrain the rise of multidrug resistance , depending on the particular order of drugs used in a sequential therapy .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"bacteriology",
"antimicrobials",
"organismal",
"evolution",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"drugs",
"microbiology",
"pseudomonas",
"aeruginosa",
"antibiotic",
"resistance",
"pharmaceutics",
"antibiotics",
"microbial",
"evolution",
"evolutionary",
"adaptation",
"pharmacology",
"bacteria",
"bacterial",
"pathogens",
"pseudomonas",
"antimicrobial",
"resistance",
"medical",
"microbiology",
"microbial",
"pathogens",
"drug",
"research",
"and",
"development",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"drug",
"therapy",
"evolutionary",
"biology",
"bacterial",
"evolution",
"evolutionary",
"processes",
"organisms"
] |
2017
|
History of antibiotic adaptation influences microbial evolutionary dynamics during subsequent treatment
|
Cerebral malaria ( CM ) is a complex parasitic disease caused by Plasmodium sp . Failure to establish an appropriate balance between pro- and anti-inflammatory immune responses is believed to contribute to the development of cerebral pathology . Using the blood-stage PbA ( Plasmodium berghei ANKA ) model of infection , we show here that administration of the pro-Th2 cytokine , IL-33 , prevents the development of experimental cerebral malaria ( ECM ) in C57BL/6 mice and reduces the production of inflammatory mediators IFN-γ , IL-12 and TNF-α . IL-33 drives the expansion of type-2 innate lymphoid cells ( ILC2 ) that produce Type-2 cytokines ( IL-4 , IL-5 and IL-13 ) , leading to the polarization of the anti-inflammatory M2 macrophages , which in turn expand Foxp3 regulatory T cells ( Tregs ) . PbA-infected mice adoptively transferred with ILC2 have elevated frequency of M2 and Tregs and are protected from ECM . Importantly , IL-33-treated mice deleted of Tregs ( DEREG mice ) are no longer able to resist ECM . Our data therefore provide evidence that IL-33 can prevent the development of ECM by orchestrating a protective immune response via ILC2 , M2 macrophages and Tregs .
Malaria remains a major health problem for humans infected with Plasmodium species . Cerebral malaria ( CM ) is a severe and potentially fatal neurological manifestation of infection and accounts for approximately one million deaths annually of children in sub-Saharan Africa alone [1] . CM is characterized by a strong Th1 immune response , with a robust and uncontrolled production of proinflammatory cytokines ( IFN-γ and TNF-α ) and chemokines ( IP-10/CXCL10 , KC/CXCL1 and MCP-1/CCL2 ) [2 , 3] that contribute to vascular leakage and sequestration of parasitized red blood cells ( pRBCs ) and leukocytes within the brain blood vessels [4 , 5] . In malaria , the balance between pro- and anti-inflammatory cytokines is critical in determining the outcome of infection , and recent evidences suggest that helminth co-infection may dampen immunopathological responses to malaria parasite by inducing a protective type-2 response [6 , 7] . Studies using murine models of malaria have established that genetic background of the host affects the development and outcome of ECM . Infection of C57BL/6 mice , which present a Th1-biased phenotype , with the rodent parasite Plasmodium berghei ANKA ( PbA ) induces a fatal cerebral disease characterized by neurological disorders including paralysia , convulsion and coma . In contrast , BALB/c mice , that present a Th2-biased phenotype , do not develop neurological complications and die at later stages from high parasitemia and anaemia [8 , 9] . IL-33 , the latest member of the IL-1 cytokine family [10] , plays an important role in Th2-associated immune responses [11 , 12] . IL-33 has been linked to a number of inflammatory disorders including allergic asthma , rheumatoid arthritis , allergic rhinitis and ulcerative colitis [13] . IL-33 binds to a heterodimer receptor composed of ST2 ( IL-33R ) and IL-1R accessory protein , leading to the production of IL-4 , IL-5 , IL-10 and IL-13 from mast cells , eosinophils , Th2 lymphocytes and the newly discovered population of type 2 innate lymphoid cells ( ILC2 ) [12] . In vitro , IL-33 has also been shown to synergize with IL-4 to drive the polarization of alternatively-activated macrophages ( M2 ) [14] , that secrete high levels of IL-10 and TGF-β . We hypothesized that IL-33 could divert the deleterious Th1-immune response during infection and therefore protects mice from ECM . Results reported here demonstrate that PbA-infected C57BL/6 mice treated with recombinant IL-33 presented no signs of neurological pathology associated with CM and had reduced production of pro-inflammatory cytokines and chemokines . This IL-33-protective effect was mediated by the activation of ILC2 that produced type 2 cytokines which in turn polarized anti-inflammatory M2 macrophages . Furthermore , M2 macrophages expanded Tregs , the depletion of which abrogated the protective effect of IL-33 . We therefore present a previously unrecognised role of IL-33 in ECM , and provide evidence that induction of type 2 immunity by an exogenous cytokine treatment was sufficient to down-regulate the inflammatory Th1 response and ECM induced by PbA .
To determine whether IL-33 could modulate malaria pathogenesis , we first infected C57BL/6 mice with Plasmodium berghei ANKA ( PbA ) parasites ( 104 parasitized red blood cells , pRBC ) and treated the mice with recombinant murine IL-33 ( 0 . 2 μg/mouse/day , intraperitoneally ) starting from day 0 . Body weight loss , parasitemia , clinical score and survival were monitored daily . PbA-infected control mice developed parasitemia , body weight loss and CM symptoms ( head deviation , ataxia and paraplegia ) from day 5 post infection , and all mice succumbed to CM by day 7–8 ( Fig . 1A-D ) . In contrast , mice treated with IL-33 displayed reduced body weight loss , clinical score and survived up to day 20 post-infection when they were euthanised due to development of hyperparasitemia ( up to 40% parasitemia ) ( Fig . 1D ) . This indicates that IL-33 administration protects mice from ECM but not from malaria-induced hyperparasitemia and death . Similar results were obtained when the IL-33 treatment began one day after infection ( day +1 ) . Similar results were also obtained with 100× higher PbA infective dose ( S1A-S1B Fig . ) . Adherence of parasited red blood cells ( pRBC ) to the vascular endothelium of organs plays a key role in the pathogenesis of Plasmodium species allowing the parasite to escape clearance in the spleen [15] . In vivo imaging using luciferase-expressing PbA confirmed that the parasite biomass was significantly reduced in IL-33-treated mice indicating that the reduction in blood parasitemia was not due to an increase of parasite sequestration in the peripheric organs ( Fig . 1E , F ) . CM is associated with parasite sequestration into the brain microvasculature and cerebral hemorrhage that result from excessive systemic inflammation , which involves pro-inflammatory cytokine production leading to endothelial cell activation and vascular permeability [16] . Using the luciferase-expressing PbA , we found a strong increase in parasite biomass in the brain of PBS-treated mice on day 7 post-infection ( Fig . 1G , H ) . In contrast , IL-33-treated animals showed a significant reduction of luciferase activity in the brain , indicating diminished pRBC accumulation . Histopathological analysis of brains from PBS-treated mice showed microhemorrhages and cytoadhesion of erythrocytes and leucocytes to the brain vasculature on day 7 . Mice treated with IL-33 displayed markedly fewer hemorrhages and less vessel obstruction compared to the PBS-treated mice ( Fig . 1I , J ) . Finally , quantitative PCR analysis showed an upregulation of Icam-1 expression in the brain tissues of PbA-infected PBS-treated mice , that was absent in IL-33-treated mice ( Fig . 1K ) , consistent with the reduction of cytoadhesion in IL-33-treated mice compared to the controls . To investigate how IL-33 might interfer with the host immune response , we measured the levels of the key Th1 cytokines ( IFN-γ , IL-12 and TNF-α ) , key Th2 cytokines ( IL-4 , IL-5 and IL-13 ) , and the regulatory cytokine IL-10 in the serum at various time-points after infection with PbA . Levels of IFN-γ , IL-12 and TNF-α rose progressively to day 5 in the serum of PBS-treated mice . In contrast , the levels of these cytokines were significantly reduced in IL-33-treated mice ( Fig . 2A-C ) . The levels of serum IL-5 were low in the control mice but were strongly enhanced in IL-33-treated mice , ( Fig . 2D ) . The levels of serum IL-4 and IL-13 were barely detectable in all groups of mice . The concentrations of serum IL-10 also increased in PBS control mice but were reduced in IL-33-treated mice ( Fig . 2E ) . Early production of the pro-inflammatory chemokines IP-10/CXCL10 , KC/CXCL1 and MCP-1/CCL2 was also reduced in the serum of IL-33-treated mice compared to PBS-control mice ( Fig . 2F-H ) . We then assessed the expression of lineage-specific transcription factors in CD4+ T cells purified from the spleen of PbA-infected mice . The Th1-specific transcription factor Tbet peaked at day 3 of infection in PBS control mice but was reduced on day 3 and unchanged on day 5 in IL-33-treated mice compared to PBS-treated mice ( Fig . 2I ) . IL-33 did not affect Gata3 expression in purified CD4+ T cells ( Fig . 2J ) , suggesting that Th2 cells are unlikely involved in IL-33-mediated protection . T cell-derived Granzyme B ( GrmB ) is known to drive cytotoxic T cell-mediated cerebral pathology [17] . We therefore determined GrmB expression in splenic CD8+ and CD4+ T cells . Percentage and frequency of GrmB+CD8+ and GrmB+CD4+ T cells in PbA-infected mice was markedly increased compared to non-infected mice ( Fig . 2K , L ) . Granzyme B positive CD8+ and CD4+ T cells were significantly reduced in percentage and number in IL-33-treated mice compared to PBS-treated control mice ( Fig . 2K , L ) . Together , our results suggest that IL-33-mediated protection against ECM is likely associated with reduction in the early pro-inflammatory type-1 response . Recently , IL-33 has been found to directly induce ILC2 expansion and cytokine production in vitro and in vivo [18–20] . We therefore analysed the effect of IL-33 on ILC2 in the ECM model . The frequency and number of ILC2 in the spleen of non-infected ( NI ) or PbA-infected mice after treatment with PBS or IL-33 were analysed by FACS . A small but consistent percentage of lineage negative CD45+ST2+ICOS+ cells , corresponding to ILC2 [21] , was found in the spleen of NI mice in the absence of IL-33-treatment . However IL-33-treatment significantly increased the percentage and number of ILC2 in the spleen in NI and PbA-infected mice compared to PBS-treated mice ( Fig . 3A-C ) . These cells were negative for lineage markers ( CD4 , CD11b , CD11c , NK1 . 1 , CD3e , Ter119 , FcεRI , Siglec F , Gr1 , CD49b , CD5 , F4/80 ) and positive for the innate lymphoid cell markers CD127 , CD44 , Sca-1 , IL-1R1 and CD25 ( S1C-S1D Fig . ) . The proportion of Ki67+ cells among splenic ILC2 was higher in IL-33-treated mice , suggesting that at least some of these cells proliferated in situ ( Fig . 3D ) . Intracellular staining revealed that ILC2 from IL-33-treated mice , but not from PBS-treated mice , expressed substantial levels of IL-4 , IL-5 and IL-13 after ex vivo PMA-ionomycin stimulation ( Fig . 3E-F ) . We were unable to detect IL-4 , IL-5 or IL-13 production by CD4+ T cells or FcεR1+ cells . These data demonstrated that IL-33 not only induced the recruitment and proliferation of ILC2 but further activated these cells to produce Type-2 cytokines in vivo during PbA infection . To investigate the potential role of ILC2 in IL-33-driven protection from CM , we sorted ILC2 and adoptively transferred them into naïve WT mice . One day after ILC2 transfer , mice were infected with PbA and monitored daily for parasitemia , body weight loss and neurological symptoms . Two injections of IL-33 ( 0 . 2 μg/mouse , i . p . ) were given to the recipient mice 30 min and 24 h after cell transfer . An earlier study has shown that the provision of IL-33 boosts the survival and cytokine production of the transferred ILC2 cells [22] . A control group infected with PbA and similarly treated with IL-33 confirmed that IL-33 at this dose and schedule of treatment was suboptimal and not protective ( Fig . 3G , H ) . While the two control groups developed severe body weight loss and succumbed to CM by day 7 , all the recipient mice given ILC2 exhibited limited clinical disease and survived beyond day 14 ( Fig . 3G , H ) . The parasitemia of ILC2-transferred mice was reduced within the first week of infection compared to PbA control group ( Fig . 3I ) . Mice that received only 2 injections of IL-33 , exhibited a slight reduction of parasitemia but nevertheless succumbed to CM by day 7 . Histopathology analysis of the brain on day 7 revealed a significant reduction of microhemorrhages and vessel obstruction in the ILC2 recipients compared to the control mice ( Fig . 3J , K ) . The cells producing IL-5 and IL-13 in the ILC2 recipient mice are CD4− T cells and not CD4+ T cells ( S2A-S2C Fig . ) , indicating that they are unlikely to be Th2 cells . These results therefore demonstrate that ILC2 play an important role in the IL-33-mediated protection against ECM . We then investigated the mechanism by which ILC2 protects mice against ECM . Macrophages can be divided into specific subsets according to their polarization environment , phenotype , and function . M1 ( classically-activated macrophages ) typically produce pro-inflammatory cytokines , including TNF-α and IL-12 , whereas M2 ( alternatively-activated macrophages ) have been implicated in immune regulation , phagocytosis , and tissue remodeling [23 , 24] . As M2 macrophages are polarized by IL-4 and IL-13 , and that these cytokines are produced by ILC2 , we therefore assessed the profile of macrophage polarization in mice infected with PbA with or without IL-33 treatment . The percentage and number of CD11b+F4/80+CD11c− cells in the spleen of mice administered with IL-33 were significantly elevated compared to that treated with PBS ( Fig . 4A-C ) . Expansion of CD11b+F4/80+CD11c− cells was accompanied by increased expression of the key M2 marker ( CD206 , mannose receptor ) and the reduction of the M1 markers ( CD86 , MHC-II and CD40 ) on macrophages recovered from IL-33-treated mice compared to that of the PBS control mice ( Fig . 4D and S3A Fig . ) . The expression of other M2 markers ( Arginase-1 , Ym1/chitinase 3–like 3 and Fizz1/resistin-like α ) were also elevated in the spleen whereas the expression of the key M1 marker , Nos2 , was reduced in IL-33-treated mice compared to that of the PBS control mice ( Fig . 4E ) . Interestingly , mRNA expression of heme oxygenase-1 ( Hmox-1 ) , an enzyme that converts heme into carbon monoxide , was significantly increased in the spleen of IL-33-treated mice compared to PBS-control mice ( Fig . 4E ) . QPCR analysis on sorted CD11b+F4/80+CD11c− cells confirmed that splenic macrophages from IL-33-treated mice , but not from PBS-treated mice , exhibited an M2-polarization status ( S3B Fig . ) . Overall , our data indicate that IL-33 increases macrophage number in the spleen and promotes their polarization towards M2 phenotype . Since IL-33 alone is not sufficient to fully differenciate M2 macrophages [14] , we investigated the potential role of ILC2 in M2 polarization in vitro . Purified ILC2 were co-cultured in a transwell culture with bone marrow-derived macrophages ( BMDM ) in culture medium alone ( M0 ) or supplemented with IL-4 ( M2-polarizing conditions ) . After 24 h , BMDM were harvested and analysed by qPCR for M2 markers . Under the M0 conditions , BMDM alone did not expressed detectable M2 markers . However , when co-cultured with ILC2 , a low level of Arginase-1 , Ym1 and Fizz1 RNA became detectable ( Fig . 5A ) . Under the M2-polarizing conditions , these markers were clearly detected and markedly enhanced by the presence of ILC2 ( Fig . 5A ) . Functionally , the polarized M2 macrophages displayed enhanced capability to uptake dextran-FITC or pRBC compared to unpolarized macrophages ( S3C-S3D Fig . ) . Since ILC2 proliferation and cytokine production are IL-7- and IL-33-dependent [25] , we determined the effect of activated-ILC2 on BMDM . To avoid any direct effect of IL-33 on BMDM , we used ST2-deficient BMDM . When cultured alone , ST2-deficient BMDM did not express any M2 markers even when IL-33 or IL-33 + IL-7 were added to the culture . In the presence of ILC2 , the expression of Arginase-1 and Ym1 in the ST2-deficient BMDM was increased , and the expression of these markers were further enhanced by the addition of IL-33 alone or in combination with IL-7 ( Fig . 5B ) , suggesting that activated-ILC2 can produce cytokines involved in M2 polarization . The level of Fizz1 expression was high in the presence of ILC2 alone and was reduced in the presence of IL-33/IL-7 . The reason of this reduction is not clear but could be due to over-stimulation of Fizz1 expression . We then analysed the production of Th2 cytokines by ILC2 in vitro . ILC2 alone ( without BMDM ) were able to produce low levels of IL-4 , IL-5 and IL-13 in the culture supernatants . This production was markedly increased after stimulation by IL-33 and IL-7 ( Fig . 5C ) . Flow cytometry analysis of ILC2 confirmed that ILC2 in the culture produced IL-4 and IL-13 and the production was further enhanced by the presence of IL-33 + IL-7 ( Fig . 5D , E ) . The polarized M2 did not produce detectable amount of IL-33 . We next investigated the role of ILC2 in the polarization of M2 in vivo . Sorted ILC2 were adoptively transferred to naïve C57BL/6 mice which were infected with PbA and treated with suboptimal doses of IL-33 , as described in Fig . 3F . The protected ILC2 recipients had increased expression of Arginase-1 , Ym1 and Fizz1 in their spleen cells ( Fig . 5F ) , indicating that ILC2 are involved in M2 polarization in vivo . Together , our data showed that IL-33-induced ILC2 can effectively drive M2 polarization in vitro and in vivo . Previous studies have implicated Tregs to limit disease and immunopathology in the PbA-induced models of ECM [26–28] and IL-33 has been shown to induce Tregs in vivo [20 , 29–31] . We therefore explored the impact of IL-33 , ILC2 and M2 macrophages on Tregs in the ECM model . First , we noted that the level of Foxp3 message in sorted splenic CD4+ T cells was increased in IL-33-treated mice during PbA-infection compared to PBS-treated mice . The percentage and total number of Foxp3+ splenic CD4+ T cells in the mice infected with PbA were significantly increased by the treatment with IL-33 ( Fig . 6A-B ) . We then determined the effect of adoptively transferred ILC2 in the induction of Tregs in vivo ( as described in Fig . 3F ) . Suboptimal doses of IL-33 led to increased frequency of Foxp3+ cells among splenic CD4+ suggesting that IL-33 alone could induce Treg polarization which was however not sufficient to protect the mice from ECM ( see Survival curve and Clinical Score in Fig . 3F ) . However , the spleens of IL-33-treated PbA-infected mice given ILC2 contained significantly higher frequency of Foxp3+ among CD4+ T cells compared to those not given ILC2 ( Fig . 6C-D ) . To demonstrate a direct link between M2 and Tregs , we co-cultured purified CD4+CD25+ T cells with M2 in the presence of soluble anti-CD3 . M2 significantly expanded the Foxp3+ Tregs ( Fig . 6E ) . We also cultured CD4+CD25− T cells under the inducible Treg ( iTregs ) conditions ( plate-bound anti-CD3 , soluble anti-CD28 + TGF-β , anti-IL-4 and anti-IFN-γ ) in the presence or absence of M2 . M2 markedly enhanced the development of iTregs as determined by the frequency of Foxp3+CD4+ T cells ( Fig . 6F ) . We then investigated whether the Treg population was involved in IL-33-mediated protection from ECM . For this purpose , we used DEREG mice , in which administration of diphtheria toxin ( DT ) leads to specific-depletion of Tregs due to expression of DT receptor-enhanced Gfp under the control of the Foxp3 promoter [32] . DEREG mice were infected with PbA and treated daily with PBS or IL-33 from the start of infection . DT was administered intraperitoneally every second day from day 1 . Treg depletion in IL-33-treated DEREG mice was confirmed in the peripheral blood by flow cytometry ( Fig . 7A ) . As previously reported [33] , Treg depletion in PBS-treated DEREG mice has no effect on the parasitemia and survival . PbA-infected DEREG mice treated with PBS died on day 7 with severe ECM ( Fig . 7B , C ) . IL-33-treated infected DEREG mice did not develop CM and died at later stages from hyperparasitemia . In contrast , IL-33-treated PbA-infected DEREG mice that received DT developed cerebral disease and died by day 7 . IFN-γ and Granzyme B production by splenic CD8+ T cells , which were significantly reduced in the IL-33-treated mice , was partly reversed in IL-33-treated mice after DT administration ( Fig . 7D-E ) . In addition , serum levels of IFN-γ and IL-12 , which were markedly reduced in IL-33-treated mice , were also restored when Tregs were depleted ( Fig . 7F ) . Together these results indicate that IL-33 induces ILC2 which in turn polarize M2 macrophages . M2 can expand Tregs which mediate the suppression of the Th1 response , which is critical to ECM pathogenesis ( Fig . 8 ) .
Data reported here reveal a previously unrecognised role of IL-33 in the protection against cerebral malaria by reducing pro-inflammatory cytokines and chemokines production and inhibiting vascular sequestration of infected erythrocytes and inflammatory cells in the brain . Furthermore , we provide a plausible mechanistic pathway by which IL-33 induces the expansion of ILC2 which in turn promote the polarization of M2 macrophages and Tregs that are critical for the protection against ECM . ILC2 have emerged as key players in experimental and clinical diseases [34] . They expand strongly in vivo in response to IL-25 and IL-33 , and represent the predominant early source of IL-5 and IL-13 during allergic inflammation and worm infection [25 , 35] . In PbA-infected mice , exogenous IL-33 induces a robust expansion and mobilization of ILC2 which have the potential to produce IL-4 , IL-5 and IL-13 . Adoptive transfer of ILC2 markedly ameliorated ECM . IL-33 administration following ILC transfer was necessary to induce the protection , likely because ILC2 require IL-33 stimulation to expand and produce sufficient amount of type 2 cytokines [22] . However , the possibility that IL-33 may also act on other cell types that participate in the protection against ECM cannot be excluded . In vitro , we confirmed that ILC2 can polarize BMDM into M2 macrophages in a cell-cell contact independant manner . IL-33 increased production of IL-4 and IL-13 by ILC2 , which can synergize to polarize M2 macrophages [14 , 36] . This is supported by our data showing that adoptively transferred ILC2 can collaborate with IL-33 to polarize M2 in vivo . M2 macrophages exhibit potent anti-inflammatory properties and play important roles in parasite clearance , tissue repair and remodeling [37] . One of the proposed mechanism for the immunomodulatory role of M2 macrophages is the competition between Arginase-1 ( expressed by M2 ) and iNOS ( expressed by M1 ) for the subtrate L-Arginine [37] . Another mechanism is the production of carbon monoxide ( CO ) by heme-oxygenase-1 ( HO-1 ) , an enzyme which has been shown to be preferentially expressed in CD206+ M2 macrophages [38] . HO-1 catalyzes the degradation of heme into biliverdin , iron and CO [39] . Here , we found that IL-33-treated mice expressed higher levels of HO-1 in the spleen . Free heme release during Plasmodium spp . infection contributes to blood brain barrier disruption and ECM pathogenesis [40] . CO production by HO-1 has been shown to suppress PbA-induced ECM by inhibiting blood brain barrier disruption , reducing adhesion molecule expression in the brain microvasculature and CD8+ T cell sequestration in the brain [40] . The role of IL-33 in the induction of CO via HO-1 merits further investigation . Tregs expansion by IL-2/anti-IL-2 complexes in vivo has been implicated to protect mice against T cell-mediated immune pathology in PbA-induced ECM [28] , although direct evidence for a role of Tregs in ECM remains elusive . We and others have shown earlier that IL-33 administration leads to Treg induction [20 , 29–31 , 41] . Here , we provide data supporting that IL-33-mediated induction of Tregs in PbA-infected mice involves the activity of ILC2 and M2 macrophages . We also show that M2 expand natural Tregs and inducible Tregs in vitro . Importantly , Tregs depletion abrogated the protective effect of IL-33 in ECM by reducing the Th1 cell response . These results therefore demonstrate a cascade of events leading to the protection of ECM by IL-33 ( Fig . 8 ) . The detailed mechanism by which Tregs suppress effector T cells , the major immunopathological mediators of ECM , remains to be explored . It is important to note that IL-33-treated mice produced minimal amount of IL-10 during PbA-infection ( Fig . 2E ) . Furthermore , administration of anti-IL-10 monoclonal antibody in IL-33-treated mice did not affect the protection conferred by IL-33 ( S4A-S4B Fig . ) . Therefore , it is unlikely that M2- or Tregs-produced-IL-10 participates to ECM protection by IL-33 . This is consistent with an earlier report which shows CTLA4-dependent but IL-10-independent protection against ECM [28] . IL-33-treated mice , though consistently showed significant reduction in parasitemia at the early stage of infection ( day 5–7 ) compared to untreated mice , were unable to clear the parasite and eventually died at later stage from hyperparasitemia . IL-33-mediated protection was achieved when the cytokine was given relatively early after infection and delaying treatment by 48 h failed to control the disease . This observation suggests a fine temporal interplay between the protective T cell response against the parasite and the anti-inflammatory response during PbA infection . Spleen is a key site for removal of pRBC during malaria through production of reactive oxygen species and phagocytosis by activated macrophages [42] . Moreover , it has been shown that Flt3L-induced CD11bint F4/80+ red pulp macrophages , which ressemble the macrophages induced by IL-33 , displayed higher phagocytic activity and contributed to parasite clearance in PbA-infected mice [43] . Here , we observed that polarized M2 macrophages showed enhanced capacity in dextran-FITC or pRBC uptake compared to unpolarized macrophages ( S3C-S3D Fig . ) . Therefore , the initial reduced parasitemia observed in our model could be explained by local activation and proliferation of red pulp macrophages , which might contribute to the parasite killing . Although it is possible that the low parasitemia contributed to IL-33-mediated protection by reducing PbA antigenic stimulation , it is unlikely an influential mechanism since we found that IL-33 treatment was still equally protective in mice infected with a 100× higher dose of pRBC which displayed a parasitemia > 5% ( S1A-S1B Fig . ) . IL-33 can directly stimulate eosinophil differentiation and survival [44] . Eosinophil granules contain cytotoxic , highly basic proteins , including the eosinophilic cationic protein that has been shown to inhibit P . falciparum in culture [45] . In our model , although the percentage and number of splenic eosinophils were augmented by 6 folds in IL-33-treated mice compared to untreated mice , eosinophil depletion using anti-Siglec-F antibody did not affect IL-33-mediated protection during PbA infection ( S4C-S4G Fig . ) suggesting that eosinophils are unlikely an influential mechanism in IL-33-mediated protection against ECM . Increased number of CD4+CD25+Foxp3+ Tregs have been observed in humans infected with Plasmodium falciparum [27 , 46 , 47] . It would be of considerable interest to investigate if the observation reported here is also applicable to clinical cerebral malaria .
Female C57BL/6 mice ( 8–10 weeks old ) were obtained from Charles River UK Ltd . ST2−/− female mice ( on the C57BL/6 genetic background ) were originally provided by Dr . Andrew McKenzie ( Medical Research Council Laboratory of Molecular Biology , Cambridge , U . K . ) and bred in-house in a pathogen-free facility at University of Glasgow . DEREG mice ( on the C57BL/6 genetic background , originally provided by Dr . Tim Sparwasser , Hannover Medical School , Germany ) were bred at Transgenose Institute animal facility ( UPS44 CNRS , Orleans , France ) . Mice under procedure were kept in polyethylene boxes with free access to food and water , and subjected to 12 h light-dark cycles . All experiments were performed in accordance with the UK Home Office guidelines and within the terms of the Project License ( PPL 70/7293 ) granted for this work under the Animals ( Scientific Procedures ) Act 1986 . All efforts were made to minimize the number of animals used and their suffering . C57BL/6 red blood cells infected with Plasmodium berghei ANKA parasites expressing a green fluorescent protein ( PbA GFPcon 259cl2 , MRA-865 , deposited by CJ Janse and AP Waters ) were stored in liquid nitrogen and thawed and passed into wild-type mice that served as parasite donor . All mice , unless otherwise stated , were inoculated intravenously ( i . v . ) into the tail vein with 1×104 parasitized red blood cells ( pRBC ) . Parasitemia was monitored daily ( from day 4 ) by flow cytometry using the FL3 channel ( PbA-GFP ) and TER-119 APC ( erythrocytes ) in a BD FACScalibur cytometer ( BD Biosciences ) . Clinical score was assessed using the following clinical scale: 1 = no signs; 2 = ruffled fur and/or abnormal posture; 3 = lethargy; 4 = reduced responsiveness to stimulation and/or ataxia and/or respiratory distress/hyperventilation; and 5 = prostration and/or paralysis and/or convulsions . All animal that reachs stage 4 developped ECM . Recombinant murine IL-33 ( Biolegend ) was injected intraperitoneally ( 0 . 2 μg/mouse/200 μl ) daily , routinely from the beginning of infection ( day 0 ) . As a control for IL-33 effects , non-infected ( NI ) mice also received IL-33 for 5 consecutive days . For some experiments , mice were administered with anti-Siglec F ( MAB17061 , R&D Systems , 50 μg daily ) or anti-IL-10 ( MAB417 , R&D Systems , 40 μg daily ) or appropriate isotype control Abs . Mice infected with the transgenic PbA strain that constitutively expresses luciferase ( PbA-luc , gift of Dr . AP Waters , Glasgow , UK ) were imaged using an IVIS Imaging 100 system ( Xenogen Corp . ) Mice were injected intraperitoneally with D-luciferin ( PerkinElmer , 150 mg/kg in DPBS ) and anesthetized in 5% isoflurane/1L O2 . min−1 atmosphere . The animals were then placed in the imaging chamber of the IVIS and anesthesia was maintained using 2% isoflurane/0 . 2L O2 per mouse min−1 atmosphere . Bioluminescence ( photons per second per square centimeter per steridian ) was monitored over a 20 min period in previously defined regions of interest ( ROI ) . Exposure times varied between 0 . 5 and 1 min , depending on signal intensity . To standardize imaging and to allow comparison between mice , the images presented in the figures were taken once luminescence plateaued . For brain bioluminescence imaging , mice were sacrificed on day 7 , perfused with 20 ml ice-cold PBS and the whole brains were excised and imaged ex vivo as described previously [48] . To enhance the signal and avoid desiccation , 100 μl D-luciferin ( 150 μg/ml ) was pipetted onto the surface of each brain 5–10 min prior to imaging . After intracardiac perfusion with 20 ml ice-cold PBS the brain was removed , fixed with 4% neutral phosphate-buffered formalin ( Merck ) and embedded in paraffin . The tissue were cut into 4 μm sections and stained with hematoxylin-eosin ( H&E ) following standard procedures . Brain microvascular obstruction in coronal brain sections was scored by two independent observers blinded to the experimental groups using a Nikon Eclipse E400 microscope at ×400 . For each brain , fields containing vessels were scored using a semi-quantitative scale ( 0–5 ) according to the severity of obstruction and the presence of microhaemorrhages: 0 , no obstruction; 1 , only small vessels obstructed; 2 , presence of leukocytes attached to the endothelium; 3 , partial obstruction , presence of leukocytes and RBC; 4 , total obstruction , without haemorrhages; 5 , total obstruction with haemorhages . Data are presented as the average score for each brain . Brains and spleens were excised at indicated time-points and preserved in RNAlater ( Qiagen ) . CD4+ T cells were purified from total splenocytes at indicated time-point by negative selection ( AutoMACS , Miltenyi Biotec ) with 85–90% purity . After homogenization in TRIzol ( Sigma-Aldrich ) , total RNA was extracted with an RNeasy Mini kit ( Qiagen ) . cDNA was synthesized using M-MLV Reverse Transcriptase ( Promega ) . The quantitative RT-PCR ( qPCR ) assays were performed using TaqMan Real-Time PCR Master Mix in an ABI PRISM 7500 Fast Sequence Detection System ( Applied Biosystems ) . Relative expression levels were calculated as ΔCt values by normalizing Ct values of target genes to Ct values of hypoxanthine phosphoribosyl transferase-1 ( Hprt1 ) . Data are represented as relative % of Hprt1 expression . All primers were purchased from Applied Biosystems ( TaqMan Gene Expression Assay ) . Cells were first blocked with FcγR blocker and stained with fluorochrome labeled Abs or their corresponding isotype controls . Abs were purchased from BD Bioscience , Biolegend or eBioscience . The following Abs were used: anti-ST2 ( DJ8 ) , anti-CD45 ( 30-F11 ) , anti-ICOS ( C398 . 4A ) , anti-CD11b ( M1/70 ) , anti-F4/80 ( BM8 ) , anti-CD11c ( N418 ) , anti-CD40 ( 1C10 ) , anti-CD206 ( C068C2 ) , anti-CD86 ( GL1 ) , anti-MHC-II ( M5/114 . 15 . 2 ) , anti-IL-4 ( 11B11 ) , anti-IL-5 ( TRFK5 ) , anti-IL-13 ( eBio13A ) , anti-Granzyme B ( NGZB ) . For intracellular cytokine staining , cells were incubated for 4 h with phorbol-12-myristate-13-acetate ( 50 ng/ml; Sigma-Aldrich ) , ionomycin ( 750 ng/ml; Sigma-Aldrich ) and GolgiStop ( 1 μl/ml; BD Biosciences ) . After surface staining , cells were fixed and permeabilized with BD Fixation/permeabilization kit ( BD Biosciences ) and stained for intracellular cytokines . For all experiments , cells were stained with a Live/Dead Fixable dye ( Molecular Probes ) to allow gating on viable cells . Data were acquired using a Beckman Coulter CyAn ADP ( Beckman Coulter , USA ) . Gating strategy and analysis were performed using the FlowJo software ( treeStar Software , USA ) and shown in S1C Fig . To induce ILC2 in vivo , naive C57BL/6 mice were inoculated intranasally with 1 μg recombinant IL-33 ( BioLegend ) on five consecutive days . Lung tissue was digested with Liberase TL ( Roche , 0 . 2 mg/ml ) and DNAse I ( Sigma , 0 . 5 mg/ml ) for 45 min at 37°C under rotation . Total lung cells were stained with lineage cocktail Abs ( anti-CD3ε , anti-CD11b , anti-CD11c , anti-NK1 . 1 , anti-siglec F , anti-FcεRI , anti-B220 ) , anti-CD45 , anti-ST2 and anti-ICOS Abs for 30 min at 4°C . ILC2 were sorted by FACSAria ( BD Biosciences ) ( purity >98% ) . ILC2 were defined as CD45+ICOS+ST2+ lymphoid cells negative for lineage markers as described previously [18] . For adoptive transfer , 2×106 freshly purified ILC2 were injected i . v . to naive C57BL/6 mice , which were infected i . v . with 104 pRBC 24 h later . Mice were then treated with IL-33 ( 0 . 2 μg , i . p . ) 30 min and 24 h after cell transfer . Bone marrow cells were harvested from femur bones of C57BL/6 WT or ST2-deficient mice and cultured in petri dishes in complete medium [RPMI-1640 supplemented with 10% ( vol/vol ) FCS ( LONZA ) , 2 mM L-glutamine , 100 U/ml penicillin , 100 μg/ml streptomycin] containing 25% L929 cell-conditioning medium as a source of macrophage colony-stimulating factor ( M-CSF ) to differentiate into bone marrow-derived macrophages ( BMDM ) . BMDM were harvested on day 6 and co-cultured ( 106 cells ) in the lower chamber of a 24-well transwell plate ( 0 . 4 μM porous membrane , Corning ) under M0 polarizing conditions ( complete medium only ) or M2 polarizing conditions ( + 10 ng/ml IL-4 ) . In some experiments , 2×105 freshly sorted ILC2 were added in the upper chamber of the transwell . Cells from the lower chamber were collected for RNA analysis after 24 h co-culture . For phagocytosis assay , polarized M0 or M2 ( from BMDM ) were plated at 106 cells/ml and incubated overnight in complete medium . FITC-labeled Dextran ( Sigma , 1 mg/ml ) or PbA-GFP-parasitized RBC ( ratio macrophage:pRBC , 1:100 ) were then added and the cells were incubated at 4°C ( controls ) or 37°C for 30 min . Cells were harvested , washed and analyzed by FACS ( Beckman Coulter CyAn ADP ) . At least 20 , 000 events were collected and data were analyzed by FlowJo software , and changes were presented as percentage of FITC+ or GFP+ cells . CD4+CD25+ T cells were purified ( AutoMACS , Miltenyi ) from the spleen and lymph nodes of naïve C57BL/6 mice and cultured ( 5×105 cells/ml ) with equal number of BMDM-derived M2 for 2 days . Foxp3 expression was determined by FACS gated on live CD4+ cells . In some experiments , CD4+CD25− T cells from naïve C57BL/6 mice were cultured ( 5×105 cells/ml ) for 2 days under iTreg polarizing conditions ( 3 μg plate-bound anti-CD3 , 1 . 5 μg soluble anti-CD28 + 10 ng/ml TGF-β , 10 μg/ml anti-IL-4 and anti-IFN-γ ) in the presence or absence of equal number of M2 . Foxp3 expression was determined by FACS gated on live CD4+ cells . ELISA for serum IL-13 ( Ebioscience ) , IL-4 , IL-5 , CXCL1 , CXCL10 , CCL2 ( all from R&D Systems ) were performed following the manufacturer’s instructions . Sensitivity of the assays was between 20 and 40 pg/ml . Concentrations of serum IFN-γ , IL-4 , IL-5 , IL-10 , IL-12 and TNF-α were determined using a multiplex mouse cytokine assay ( Invitrogen ) according to the manufacturer’s instructions using a Luminex 200 reader ( Luminex Corp . ) . Comparisons between 2 groups were performed using a 2-tailed unpaired Student’s t test . Multiple groups were compared using a 2-way ANOVA followed by a Bonferroni’s post-test . Values for all measurements are expressed as mean ± SEM . P<0 . 05 was considered statistically significant . Data are representative of at least 3 separate experiments unless otherwise stated in the legend . Statistical analysis were performed using GraphPad Prism 5 . 0 . ( GraphPad Software ) .
|
Cerebral malaria ( CM ) caused by the parasite Plasmodium sp . is a fatal disease , especially in children . Currently there is no effective treatment . We report here our investigation on the role of a recently discovered cytokine IL-33 , in treating experimental cerebral malaria ( ECM ) in the susceptible C57BL/6 mice . IL-33 protects the mice against ECM . The protection is accompanied by a reduction of Th1 response and the enhancement of type 2 cytokine response . We also found that IL-33 mediates its protective effect by inducing a population of type 2 innate lymphoid cells ( ILC2 ) , which then polarize macrophages to alternatively-activated phenotypes ( M2 ) . M2 in turn expand regulatory T cells ( Tregs ) which suppress the deleterious Th1 response . Our report therefore reveals hitherto unrecognised mechanisms of the regulation of ECM and provide a novel function of IL-33 .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
IL-33-Mediated Protection against Experimental Cerebral Malaria Is Linked to Induction of Type 2 Innate Lymphoid Cells, M2 Macrophages and Regulatory T Cells
|
Hookworm infections are significant public health issues in South-East Asia . In women of reproductive age , chronic hookworm infections cause iron deficiency anaemia , which , upon pregnancy , can lead to intrauterine growth restriction and low birth weight . Low birth weight is an important risk factor for neonatal and infant mortality and morbidity . We investigated the association between neonatal birth weight and a 4-monthly deworming and weekly iron-folic acid supplementation program given to women of reproductive age in north-west Vietnam . The program was made available to all women of reproductive age ( estimated 51 , 623 ) in two districts in Yen Bai Province for 20 months prior to commencement of birth weight data collection . Data were obtained for births at the district hospitals of the two intervention districts as well as from two control districts where women did not have access to the intervention , but had similar maternal and child health indicators and socio-economic backgrounds . The primary outcome was low birth weight . The birth weights of 463 infants born in district hospitals in the intervention ( 168 ) and control districts ( 295 ) were recorded . Twenty-six months after the program was started , the prevalence of low birth weight was 3% in intervention districts compared to 7 . 4% in control districts ( adjusted odds ratio 0 . 29 , 95% confidence interval 0 . 10 to 0 . 81 , p = 0 . 017 ) . The mean birth weight was 124 g ( CI 68 - 255 g , p<0 . 001 ) greater in the intervention districts compared to control districts . The findings of this study suggest that providing women with regular deworming and weekly iron-folic acid supplements before pregnancy is associated with a reduced prevalence of low birth weight in rural Vietnam . The impact of this health system-integrated intervention on birth outcomes should be further evaluated through a more extensive randomised-controlled trial .
Low birth weight is widely recognised as a risk factor for neonatal mortality and morbidity , as well as reduced cognitive function and the development of chronic diseases in later life [1]–[3] . Iron deficiency anaemia during pregnancy is an important cause of restricted foetal growth leading to low birth weight and preterm delivery , and also maternal illness and death [4] , [5] . It is estimated that more than one third of women in the world are anaemic [6] , and iron deficiency is the most common cause of anaemia in the majority of settings [7] . In addition , many of these women live in rural communities of developing countries where intestinal parasitic infections are endemic . Hookworm infections contribute to anaemia severity and persistence by causing chronic blood loss [8] , [9] . It has been hypothesized that the anti-helminthic drugs mebendazole and albendazole may have a positive impact on birth outcomes if administered during pregnancy , but conclusive evidence is still lacking . Preventive chemotherapy through mass deworming is recommended when the prevalence of infection with any soil-transmitted helminth exceeds 20% [10] . However , very few countries have promoted routine anti-helminthic treatment in women of reproductive age [11] . Previous studies have reported benefits to maternal and infant health through antenatal supplementation with iron-folic acid supplements and multiple micronutrients [12]–[14] , although poor compliance and variable supply have limited the impact of this approach [15] . Preventative weekly iron-folic acid supplementation for women of reproductive age given before pregnancy is effective in improving iron stores and women are less likely to develop iron deficiency anaemia during pregnancy if iron stores are replete at the time of conception [16] . The WHO has recently recommended that weekly iron-folic acid supplementation be made available for women of reproductive age in areas where the prevalence of anaemia in women of reproductive age is above 20% [17] . In Vietnam , the prevalence of anaemia has previously been reportedly as high as 65% in pregnant and 54% in non-pregnant women , and in 2003 it was estimated that 21 . 8 million people had hookworm infections [18]–[20] . In November 2005 we conducted a survey of anaemia , iron deficiency and hookworm infection in women of reproductive age in Yen Bai province , northern Vietnam . The results showed prevalences to be 38% , 23% and 78% respectively [21] . Anaemia was associated with iron deficiency and meat consumption , however there was no association between hookworm infection and either anaemia or iron deficiency [21] . In response , a pilot weekly iron-folic acid supplementation and deworming program for women of reproductive age was introduced in two districts in May 2006 . By September 2007 impact and compliance surveys identified that: 90% of women were taking the weekly iron-folic acid supplements; mean haemoglobin had risen from 122 g/L to 131 g/L; mean serum ferritin levels had risen from 28 . 1 µg/L to 44 . 7 µg/L; anaemia , iron deficiency and hookworm infection prevalence had dropped to 20% , 6% and 26% respectively [22] , [23] . We hypothesized that the consequent improvement in women's nutritional status would translate to improvements in birth weight of their babies compared to babies born to women in adjacent districts where the intervention was not available .
The project was approved by the Human Research Ethics Committee of the National Institute of Malariology , Parasitology and Entomology ( Hanoi , Vietnam ) , the Walter and Eliza Hall Institute of Medical Research ( Melbourne , Australia , Project No . 03/07 ) and Melbourne Health ( Melbourne , Australia ) . The birth weight survey was locally approved by the Yen Bai Ministry of Health . Extensive consultation was undertaken between the project team and community leaders , as well as liaison with village , district and provincial health staff . Village health workers provided participants with information regarding the intervention and signed informed consent was documented . For the birth weight survey , the mother's informed consent was obtained verbally before data collection , as approved by the Yen Bai Ministry of Health Research Committee and in accordance with the Vietnamese Ministry of Health protocols for surveys . Oral consent was documented by the presence of a witness . The pilot intervention was conducted in Tran Yen and Yen Binh districts , with Luc Yen and Nghia Lo as control districts , all in Yen Bai province , a remote , mountainous province in Vietnam with low population density ( 104 people/km2 ) [24] and poor road and transport infrastructure . Intervention districts were selected based on advice from provincial authorities that they were representative of most other districts in the province in terms of population density and socioeconomic factors . All non-pregnant women between 16 and 45 years in the two districts ( estimated 51 , 623 ) were targeted . Pregnant women were identified by asking women whether they were pregnant and the timing of their last menstrual period , and were not given deworming treatment if they were or thought they may be pregnant [25] . This protocol was approved by the provincial health authorities . In May 2006 the distribution of iron ( FeSO4 ) /folic acid ( 60/0 . 4 mg ) and 4-monthly albendazole ( 400 mg ) to women was introduced . A detailed account of the protocol has been previously reported [25] . Briefly , iron-folic acid supplements were distributed to village health workers monthly ( 4 tablets , i . e . weekly consumption for one month ) through the administrative strata of the Department of Preventive Medicine . The village health workers then distributed the supplements to individual women either through organized community meetings or home delivery . Provincial authorities chose to distribute 4-monthly deworming treatment through the commune health centres as they felt that this level of the health system was more appropriate for a mass chemotherapy intervention . The intervention was preceded by training of local health workers and delivered with educational activities , distribution of promotional materials to women and community educational meetings [25] . The intervention was not introduced or publicised in other districts of Yen Bai prior to the completion of data collection for this study .
Four hundred sixty-three infants were born in the 4 district hospitals during the period of observation ( January 2008–Jun 2008 ) , 168 in intervention districts ( 47 . 5% females ) and 295 ( 47 . 2% females ) in control districts . Demographic data for the participants by study arm including maternal age at delivery , parity and socio-economic and educational background are shown in table 1: socio-economic background was similar in the two groups , while women in the intervention districts had higher maternal age but lower level of education . The prevalence of low birth weight was 3% and 7 . 4% in the intervention and control districts respectively ( Table 2 ) . This equates to a 4 . 4% absolute and 59% relative lower prevalence of low birth weight ( number needed to treat = 23 , 8 . 84–59 . 11 , i . e . one less low birth weight infant for every 23 delivering mothers who had access to deworming and iron-folic acid supplementation prior to pregnancy ) . The odds ratio of low birth weight comparing infants born to mothers from intervention versus control areas , controlling for the effect of potential confounders and incorporating the possible clustering of inhabitants of a district , was 0 . 29 ( 95% CI 0 . 10 to 0 . 81 , p = 0 . 017; clustering-adjusted and covariate-unadjusted p = 0 . 040; clustering-unadjusted and covariate-unadjusted p = 0 . 077 ) . Mean birth weight was 124 g higher in intervention districts ( 3135 vs 3011 g , CI for mean birth weight increase 68 to 255 , p<0 . 001 ) ( Table 2 ) . Mean birth weight positively correlated with socio-economic background , level of education , maternal age and parity ( data not shown ) .
We present here results that show that the prevalence of low birth weight in infants born to mothers in rural Vietnam who had access to a pre-pregnancy program of 4-monthly albendazole treatment and weekly iron-folic acid was 3 . 0% compared to 7 . 4% for those born in neighbouring districts where women did not have access to this intervention . The latter is comparable to the Vietnam national average for low birth weight of 7 . 0% [28] . In addition , we observed a significantly higher mean birth weight in the intervention group . While there are numerous studies assessing the benefits of ante-natal iron supplementation for pregnant women , birth outcome and infant health [14] , [29] , [30] , there are few reported long-term studies of the impact of pre-pregnancy weekly iron-folic acid and deworming programs for women of reproductive age on birth outcomes and infant health [31] . Berger et al ( 2005 ) reported a prevalence of 2 . 9% low birth weight infants in a group of women who were participants in a pre-pregnancy weekly iron-folic acid supplementation program who subsequently became pregnant . The prevalence of low birth weight in the daily supplementation arm was 9 . 3% , however there was no standard treatment control group in the study [32] . Our study has limitations . The study design is not that of a randomized controlled trial , which would allow a more conclusive interpretation of the results . We have tried to adjust for potential confounders but other variables may exist that would bias the results . However , our study was conducted during the implementation of a large scale anaemia and iron deficiency prevention program , and therefore provides information about the impact of the intervention when implemented under field conditions through routine health services . We are not sure about the exact supplements women took during pregnancy and with what frequency . When questioned , women were unsure of the number and source of free iron-folic acid supplements provided by health workers during pregnancy . Those who bought antenatal nutrition supplements privately did so on the recommendation of a doctor or relation/friend and again were unable to clearly state what the supplements were or how much iron they contained . We therefore cannot exclude a possible bias in the results due to differences in antenatal iron supplementation patterns between the intervention and control districts . The intervention in Tran Yen/Yen Binh did not have separate arms for iron-folic acid or deworming alone so the relative contribution of each cannot be ascertained . A previous pregnancy supplementation and deworming study suggested that haematinics and anthelminthics had an additive effect on stabilizing haemoglobin during pregnancy when given to 125 Sierra Leone women , with haematinics having the greater effect [33] . We find it plausible in our case that iron-folic acid and deworming acted additively or even synergistically , targeting the problem of maternal anaemia at different levels [6] . Regular hookworm control is likely to have complemented iron-folic acid supplementation by reducing iron loss due to chronic hookworm infection . Moreover , the pre-pregnancy population-based approach has been previously shown to result in a gradual and stable improvement in iron status prior to conception [22] . Another significant advantage of regular deworming for non-pregnant women in Vietnam is that Ministry of Health regulations proscribe administering deworming medication during pregnancy . Another limitation was that we were not able to sample the entire pool of deliveries of mothers from intervention and control districts , as the cost and logistics in such a remote area were beyond the resources available to this study . District hospitals collect about a third of routine deliveries; although this is not a fully comprehensive sample , we believe it is representative enough across the study districts . Furthermore , if selection bias did exist , we assume it would be similar in intervention vs control districts . It is also important to acknowledge that this intervention would increase the workload of community health workers , which is currently a topic of debate in the international development community . This may challenge the long term sustainability of the implementation and needs to be taken into consideration by health authorities in planning for the intervention . The data presented here is the result of one of few studies of the impact of pre-pregnancy iron-folic acid supplementation and deworming for non-pregnant women on infant birth weight . Whilst it has been recently suggested that there is a need for stronger , more robust data to support long-term intermittent iron-folic acid supplementation in women of reproductive age [34] , there is growing evidence that these programs are not only important [35] but also feasible and implementable in resource constrained settings [23] , [25] , [31] . The results of our study suggest that the pre-pregnancy combination of deworming and weekly iron-folic acid supplementation for women of reproductive age in northern Vietnam is associated with a reduced incidence of low birth weight and higher mean birth weight . Such a program could potentially represent a high-impact and easily implementable intervention to apply in settings with a high prevalence of hookworm infection and anaemia , for the health of both women and newborns .
|
Low birth weight is an important risk factor for neonatal and infant morbidity and mortality and may impact on growth and development . Maternal iron deficiency anaemia contributes to intrauterine growth restriction and low birth weight . Hookworm infections and an iron-depleted diet may lead to iron deficiency anaemia , and both are common in many developing countries . A pilot program of deworming and weekly iron-folic acid supplementation for non-pregnant women aiming to prevent iron deficiency was implemented in northern Vietnam . We compared the birth weight of babies born to women who had had access to the intervention to babies born in districts where the intervention had not been implemented . The mean birth weight of the intervention districts' babies was 124 g more than the control districts' babies; the prevalence of low birth weight was also reduced . These results suggest that providing women with deworming and weekly iron-folic acid supplements before pregnancy is associated with increased birth weight in rural Vietnam . This intervention was provided as a health system integrated program which could be replicated in other at-risk rural areas . If so it could increase the impact of prenatal and antenatal programs , improving the health of both women and newborns .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine"
] |
2012
|
Increased Birth Weight Associated with Regular Pre-Pregnancy Deworming and Weekly Iron-Folic Acid Supplementation for Vietnamese Women
|
To develop a valid and reliable quantitative measure of leprosy Type 1 reactions . A scale was developed from previous scales which had not been validated . The face and content validity were assessed following consultation with recognised experts in the field . The construct validity was determined by applying the scale to patients in Bangladesh and Brazil who had been diagnosed with leprosy Type 1 reaction . An expert categorized each patient's reaction as mild or moderate or severe . Another worker applied the scale . This was done independently . In a subsequent stage of the study the agreement between two observers was assessed . The scale had good internal consistency demonstrated by a Cronbach's alpha >0 . 8 . Removal of three items from the original scale resulted in better discrimination between disease severity categories . Cut off points for Type 1 reaction severities were determined using Receiver Operating Characteristic curves . A mild Type 1 reaction is characterized using the final scale by a score of 4 or less . A moderate reaction is a score of between 4 . 5 and 8 . 5 . A severe reaction is a score of 9 or more . We have developed a valid and reliable tool for quantifying leprosy Type 1 reaction severity and believe this will be a useful tool in research of this condition , in observational and intervention studies , and in the comparison of clinical and laboratory parameters .
Leprosy is a chronic granulomatous disease caused by Mycobacterium leprae . More than 254 000 new cases were reported to the World Health Organization in 2007 [1] . The disease predominantly affects the skin and nerves . The nerve involvement associated with the disease may lead to permanent deformity and disability . A spectrum of disease phenotypes is recognised and these are determined by the host response to the organism [2] . The tuberculoid pole of the spectrum is characterised by strong host cell mediated immunity to the organism , whereas patients with lepromatous leprosy have a predominantly humoral immune response [3] . The borderline states of the disease are immunologically unstable . Leprosy may be exacerbated by immunological complications–Type 1 ( reversal ) reactions and erythema nodosum leprosum ( Type 2 reactions ) . Type 1 reactions occur predominantly in individuals with the borderline forms of leprosy . They are characterised by inflammation of the skin , nerves or both . Type 1 reactions may occur before , during or after the successful completion of multi-drug therapy . Type 1 reactions affecting the peripheral nerves may result in decreased sensory and motor function and lead to disability . 20–30% of individuals diagnosed with leprosy will have a Type 1 reaction [4] , [5] . Type 1 reactions are usually treated with oral corticosteroids but approximately 40% of individuals do not experience complete recovery of clinically detectable nerve function impairment ( NFI ) [6] . Clinical trials with appropriate outcome measures are needed to determine the most effective treatment regimens [7] . It has proved difficult to compare the small number of studies because of the different outcome measures used . There are also difficulties in comparing the severity of Type 1 reactions between different cohorts and even between different arms of clinical trials . A tool which enables clinicians to accurately assess the severity of leprosy Type 1 reactions would be useful in defining outcomes for clinical trials . It would facilitate the even distribution of patients with similar disease severity between the arms of clinical trials . A measure of reaction severity could also be used in treatment guidelines to indicate the need for therapy . A quantitative measure of reaction severity may be a useful prognostic tool . A scale devised as part of the ILEP Nerve Function Impairment and Reaction ( INFIR ) Cohort study examined 21 parameters for the basis of a severity scale of both Types of reactions and retrospectively assessed the performance of this scale [8] . There was good agreement between items in the scale . A different scale ( with 24 parameters ) was used by Marlowe et al in a different INFIR study of azathioprine and prednisolone in Type 1 reactions but it was not validated [9] . An “indice névritique”–a composite scale using various assessments of nerves including electrophysiological studies–was developed by Naafs and colleagues but has not been validated [10] , [11] . Using the INFIR scales as a starting point we decided to develop and validate a scale for Type 1 reactions and nerve function impairment in leprosy .
A questionnaire was sent to eight leprologists who were not involved in the development of the current scale . The questionnaire used open questions to ascertain the signs they believed to be important in Type 1 reaction , which signs indicated a more severe reaction and how they categorised Type 1 reaction severity . The severity scale for leprosy Type 1 reactions was developed by modifying the two previous scales used in the INFIR studies . The scale we developed and tested has 24 parameters grouped into three parts ( see Appendix S1 ) : Section A contained six parameters which scored between 0 and 3 depending on the assessment of their severity by the examiner using the scale . Section B is an assessment of sensory function of each of the trigeminal , ulnar , median and posterior tibial nerves . Cotton wool is used to assess the trigeminal nerve . Graded Semmes-Weinstein monofilaments ( SWM ) are used for the ulnar , median and posterior tibial nerves . The ulnar and median nerves are examined using a 2 and 10g monofilament at three sites on the palmar aspect of the hand for each nerve ( ulnar and median ) and the posterior tibial nerves are assessed using 10 and 300g at four sites on the sole of the foot ( Fig . 1 ) . A score from 0 to 6 was assigned depending on the ability of the patient to successfully recognise the weighted monofilaments and the number of sites in which they were felt . For example , on the hand if a person could feel the 2g monofilament at the three sites innervated by the ulnar nerve then a score of zero was recorded . If the 2g was felt at two sites and the 10g at the third site a score of one was recorded . If however the 10g monofilament was not felt at one site then a score of 4 was recorded even if the patient was able to feel the 2g monofilament at the other two sites . Section C measures motor function of ten nerves ( facial , ulnar , median , radial , posterior tibial ) by voluntary muscle testing ( VMT ) using the MRC grading system [12] . Normal muscle power ( MRC Grade 5 ) scores zero on the scale . Grade 4 scores 1 and grade 3 scores two . An MRC grade of less than three scores three on the severity scale . The sum of the total for each section gives the overall severity scale score which ranges from 0–96 , the lower the score the less severe the reaction . The assessment of the severity scale was performed at the specialist leprosy referral centres of DBLM Hospital , Nilphamari , Bangladesh and Oswaldo Cruz Institute , Rio de Janeiro , Brazil between June 2006 and November 2007 . Ethical approval was granted for the external validation of the scale and the assessment of inter-observer agreement by the Ethics committee of the London School of Hygiene and Tropical Medicine , the Bangladesh Medical Research Council and the Institutional Review Board of the Oswaldo Cruz Institute . Patients attending the centres with evidence of a Type 1 reaction or nerve function impairment of less than 6 months duration were eligible . Eligible individuals were invited to participate by the attending physician . Written informed consent was obtained from individuals who participated in the external validation of the scale and also from those enrolled in the study of inter-observer agreement . Individuals were examined independently by a worker who was trained to use the scale and experienced leprologists ( >20 years experience ) who categorized the reaction as mild or moderate or severe . Neither assessor ( nor the patient ) was aware of the result of the others examination . All of the demographic and clinical data were recorded on a standard form . The Ridley-Jopling classification was used to classify the type of leprosy each patient had [2] . Inter-observer agreement was tested at the two centres in a subsequent stage of the study using the same eligibility criteria . Two assessors independently used the scale to assess individuals diagnosed as having Type 1 reactions . The scale was applied in the same way as in the validation part of the study . The time interval between the two assessments was kept as short as was practicable . Four pairs of assessors were used . The results were entered into an Access database . The data were analysed using the Statistical Package for the Social Sciences ( SPSS version 14 . SPSS Inc , Illinois , Chicago ) . The item to total score correlation was examined using Spearman rank correlation . The internal consistency or reliability was assessed using Cronbach's alpha . An alpha between 0 . 7 and 0 . 9 is considered acceptable [13] . The contribution of each item in the scale was assessed by calculating Cronbach's alpha for the scale if that item were removed . The ability of the scale to discriminate between different clinical severity categories was determined using analysis of variance . The threshold for accepting statistical significance was p<0 . 05 . Inter-observer reliability was evaluated using Intra-Class Correlation of the total score of each examiner using a two-way analysis of variation ( 5% level of significance ) and the strength of agreement criteria of Landis and Koch [14] . A Bland Altman plot of the difference between pairs of observations and the mean of those pairs was used to highlight any potential systematic differences between raters . Receiver Operating Characteristic ( ROC ) curves were used to determine cut off points for mild , moderate and severe reactions by calculating the sensitivity and specificity of the scale scores for mild and moderate groups and moderate and severe groups respectively .
The questionnaire sent to eight leprologists was returned by seven . The features of Type 1 reaction that were considered important indicators of severity were extent and degree of inflammation of skin lesions , the presence of peripheral oedema , nerve tenderness and nerve function impairment . These parameters are all part of the clinical severity scale we have developed and thus gives our scale face validity . 81 individuals were recruited ( 56 from Bangladesh and 25 from Brazil ) . 64 ( 79% ) were male and 17 ( 21% ) female . The clinical features are summarised in Table 1 . The range of the item to total score correlation was −0 . 09 to +0 . 73 . Nerve pain and nerve tenderness appeared to show no correlation with the total score . The internal consistency of the scale was assessed using Cronbach's alpha . The Cronbach's alpha was 0 . 819 . Removal of the following individual items resulted in an increase in the alpha: the degree of inflammation of skin lesions , the number of raised inflamed lesions , nerve pain , nerve tenderness , fever , function of right trigeminal nerve , function of the left trigeminal nerve , motor function of the right and left radial nerves ( Table 2 ) . This indicates that removal of one or more of these items might improve the ability of the remaining items to measure the severity of Type 1 reactions . Principal component analysis ( PCA ) identified a general factor to which all but nerve pain , nerve tenderness and the number of inflamed lesions contributed accounting for 23 . 5% of total variance . The important variables in the second factor accounting for 11 . 6% of the total variance were those related to the eye , namely , trigeminal nerve sensation and facial nerve motor function . The third factor which accounted for 10 . 7% contrasted individuals with skin signs and no NFI with those who only had NFI . The severity of the Type 1 reaction was categorized as mild in 19 ( 23 . 5% ) , moderate in 40 ( 49 . 4% ) and severe in 12 ( 14 . 8% ) . The severity was not recorded in 10 cases . The median scores for each category of reaction severity are shown in the box plots in Fig . 2 with the inter-quartile range ( IQR ) . The median scores for each category were: mild = 5 . 0 ( IQR = 11 ) , moderate = 10 . 5 ( IQR = 13 ) and severe = 18 . 0 ( IQR = 29 ) . The differences between the mild and moderate group and the moderate and severe groups did not reach statistical significance ( p = 0 . 053 and 0 . 052 respectively ) . The performance of the scale was not materially affected by excluding the seven individuals who did not have skin involvement . Thirty nine individuals ( 27 from Bangladesh and 12 from Brazil ) were recruited to the second stage of the study to assess inter-observer agreement . The details of these patients are presented in Table 1 . The Intra-Class Correlation coefficient based on a two-way analysis of variance with a random effects model is 0 . 994 . The strength of agreement is very good [14] . A Bland and Altman plot [15] ( Fig . 3 ) of the difference between the scores for pairs of observers plotted against the mean of the scores shows good agreement between observers with 95% of differences less than two standard deviations from the mean . The scale was adjusted and the analysis repeated in the light of the data obtained ( see Appendix S2 ) . The items nerve pain , nerve tenderness and fever were removed . The rationale for removing these items was that nerve pain and nerve tenderness performed least well of all the items in the scale ( in terms of Cronbach's alpha ) . Fever was removed because occurred in only four of the 120 participants in the study as a whole . We felt it was important to retain the cutaneous signs and trigeminal and radial nerve function parameters as these are important clinical features of Type 1 reactions . The scores for the sensory testing ( using SWM and cotton wool ) were reduced by 50% to make the maximum score possible for each sensory nerve three . This is the maximum score possible for each of the motor and cutaneous items . These adjustments result in the final scale which consists of 21 items and has a range of 0–63 . The maximum score possible for sections A , B and C are 9 , 24 and 30 respectively . For this adjusted version of the scale Cronbach's alpha remained satisfactory at 0 . 833 . The median scores for each severity group were: mild = 5 . 0 , moderate = 7 . 5 and severe = 15 . 25 . The differences between the mild and moderate groups ( p = 0 . 038 ) and the moderate and severe groups ( p = 0 . 048 ) reached statistical significance . The ROC curve for the final scale scores was plotted for individuals identified as mild or moderate by the expert raters and for those categorized as moderate or severe ( Fig . 4 ) . This facilitates the determination of cut off scores for each category [13] . Using the ROC curves in conjunction with a consideration of the clinical meaning of a given score we determined the following cut off points . A mild Type 1 reaction is characterized using the final scale by a score of 4 or less . A moderate reaction is a score of between 4 . 5 and 8 . 5 . A severe reaction is a score of 9 or more . The area under the curve for mild and moderate categories is 0 . 701 for the final scale ( 0 . 688 for the original scale ) . The area under the curve for the moderate and severe categories is 0 . 734 for the final scale ( 0 . 731 for the original scale ) . These values indicate that the final scale is a fair discriminator between the severity categories traditionally used by clinicians .
In many branches of medicine a single test or diagnostic criterion is either not available or insufficient to adequately measure or describe a clinical syndrome . This has led to difficulties in measuring the severity and prognosis of conditions . The response by researchers has been to develop composite measurement scales . Psychologists have for many years been concerned with accurately measuring and predicting behaviour and there is a large literature on how to develop and test such measures [13] , [16] . The use of unpublished scales to measure outcome has been shown to be a significant source of bias in psychiatry [17] . The lack of clear descriptions of scales and familiarity with them make clinical research difficult to interpret . We have developed and prospectively validated a reliable 21 item severity scale to measure leprosy Type 1 reactions . This scale requires the examiner to be proficient in recognising the cutaneous signs of Type 1 reaction , the assessment of VMT and the use of SWM . These skills are not widely practised in many leprosy endemic countries and we anticipate that the main use of this tool , at least initially , will be in the context of research and referral settings . We believe the scale is easy to use and requires little additional training or equipment for workers based in referral centres . Using a standard assessment form the additional time required to use the scale is minimal . Type 1 reactions are a significant cause of nerve function impairment and this is the major concern of the physician managing a patient with this condition . The scale we have developed reflects the importance of NFI in the severity of Type 1 reactions . VMT and SWM in the assessment of NFI have been shown to be reliable [18] . Monofilaments have been shown to be concordant with other sensory function tests [19] . These factors undoubtedly contribute to the robustness of the current scale but careful training and assessment of examiners is required [20] . The use of two monofilaments on the hands ( 2g and 10g ) and feet ( 10g and 300g ) simplifies the system used in the INFIR Cohort Study . However this also results in a higher sensory threshold before an individual's NFI impacts on their Type 1 reaction severity scale score . The INFIR Cohort study also used a single monofilament test site for the purely sensory radial cutaneous and sural nerves [4] . These two nerves are not commonly tested in routine clinical practice and are not included in the severity scale . The radial cutaneous and sural nerves may be assessed using various forms of quantitative sensory testing before new impairment identified by monofilaments is demonstrable . Recently published data analysing 188 individuals from the INFIR Cohort who did not present with reaction or nerve involvement has shown that impairment identified using monofilaments occurred in the radial cutaneous nerve in 7% of individuals and in the sural nerve in 6 . 1% [21] . However the definition of impairment in the radial cutaneous nerve was the inability to feel monofilaments less than 10g or in the sural nerve less than 300g [4] . The lack of a gold standard measure of Type 1 reactions has resulted in us having to compare the scale with the variable and somewhat vague clinical categories of severity as mild , moderate or severe . This has undoubtedly led to a degree of heterogeneity of Type 1 reaction severity within these categories but despite this the scale has performed well . The final scale has a high degree of inter-observer reliability . We were unable to test intra-observer reliability because of the effect of treatment on the signs of reaction . It would be unethical to withhold treatment . The assessment of intra-observer variation is desirable but not absolutely necessary in scales with a high level of inter-observer reliability [13] . The assessment of intra-observer variation has not been possible in the development of valid scales in other fields such as neurology [22] . In its present form we have found the adjusted scale to be valid and sensitive . Neurological parameters are well represented and reflect the importance of nerve function impairment . The addition of weighting of the different components of the scale would add to its complexity . A consideration we have not addressed is the performance of the scale in individuals who have nerve damage of greater than 6 months duration . The treatment of nerve damage present for this length of time with corticosteroids is not associated with significant clinical benefit compared to placebo [23] . Nerve damage greater than six months duration should not be included in the severity score . The issue of longstanding NFI can be problematic as patients who are presenting for the first time may be unsure as to the duration of the NFI and may have some acute NFI in a nerve which already has some pre-existing permanent impairment . Longstanding nerve damage in an individual who experiences a Type 1 reaction would lead to a higher score than an individual with an identical reaction but who has no pre-existing nerve damage . The severity of the Type 1 reaction in the two individuals is presumably the same . However it could be argued that individuals who already have some degree of permanent nerve damage have less neurological reserve and are thus more at risk from even a mild reaction . This however needs to be formally tested . The scale is currently being used as an additional measure in a clinical trial of methylprednisolone in Type 1 reactions . In this cohort the performance of the scale over time and its ability to reflect change will be assessed . This is the first prospective validation of a severity scale for leprosy Type 1 reactions . We believe it will prove a useful tool in more accurately assessing Type 1 reactions particularly in clinical trials where the ability to accurately compare the severity of Type 1 reactions in different patients is vital .
|
Leprosy is caused by a bacterium and is curable with a combination of antibiotics known as multi-drug therapy which patients take for six or 12 months . However , a significant proportion of leprosy patients experience inflammation in their skin and/or nerves which may occur even after successful completion of multi-drug therapy . These episodes of inflammation are called leprosy Type 1 reactions . Type 1 reactions are an important complication of leprosy because they may result in nerve damage which leads to disability and deformity . Type 1 reactions require treatment with immunosuppressive agents such as corticosteroids . The severity of Type 1 reactions varies with time , treatment and between individuals . We have developed a clinical severity scale to measure the severity of Type 1 reactions . The scale has three sections . The first measures the involvement of the skin using the number of affected skin lesions , the degree of inflammation of those lesions and the presence of swelling of the hands , feet or face . The second section is a measurement of the sensory function of the nerves supplying the eyes , hands and feet by assessing a patient's ability to feel graded nylon fibres . The third section uses a standard measure of muscle power to assess motor function of the nerves of the face , hands and feet . The clinical severity scale we have developed will facilitate the study of Type 1 reactions and enable direct comparison between different studies . This will improve the management of this disabling complication of leprosy .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"infectious",
"diseases/skin",
"infections"
] |
2008
|
Development and Validation of a Severity Scale for Leprosy Type 1 Reactions
|
Kallmann syndrome ( KS ) associates congenital hypogonadism due to gonadotropin-releasing hormone ( GnRH ) deficiency and anosmia . The genetics of KS involves various modes of transmission , including oligogenic inheritance . Here , we report that Nrp1sema/sema mutant mice that lack a functional semaphorin-binding domain in neuropilin-1 , an obligatory coreceptor of semaphorin-3A , have a KS–like phenotype . Pathohistological analysis of these mice indeed showed abnormal development of the peripheral olfactory system and defective embryonic migration of the neuroendocrine GnRH cells to the basal forebrain , which results in increased mortality of newborn mice and reduced fertility in adults . We thus screened 386 KS patients for the presence of mutations in SEMA3A ( by Sanger sequencing of all 17 coding exons and flanking splice sites ) and identified nonsynonymous mutations in 24 patients , specifically , a frameshifting small deletion ( D538fsX31 ) and seven different missense mutations ( R66W , N153S , I400V , V435I , T688A , R730Q , R733H ) . All the mutations were found in heterozygous state . Seven mutations resulted in impaired secretion of semaphorin-3A by transfected COS-7 cells ( D538fsX31 , R66W , V435I ) or reduced signaling activity of the secreted protein in the GN11 cell line derived from embryonic GnRH cells ( N153S , I400V , T688A , R733H ) , which strongly suggests that these mutations have a pathogenic effect . Notably , mutations in other KS genes had already been identified , in heterozygous state , in five of these patients . Our findings indicate that semaphorin-3A signaling insufficiency contributes to the pathogenesis of KS and further substantiate the oligogenic pattern of inheritance in this developmental disorder .
Kallmann syndrome ( KS , MIM 147950 , 244200 , 308700 , 610628 , 612370 , 612702 ) is an inherited neurodevelopmental disorder defined as the association of hypogonadotropic hypogonadism , due to gonadotropin-releasing hormone ( GnRH ) deficiency , and the inability to smell ( anosmia or hyposmia ) , related to abnormal development of the peripheral olfactory system ( olfactory nerves and olfactory bulbs ) . The genetics of KS involves various modes of transmission , specifically , autosomal recessive , autosomal dominant with incomplete penetrance , X-chromosome linked , and also oligogenic inheritance [1] , [2] . Pathohistological studies of fetuses with olfactory bulb agenesis have shown that the reproductive phenotype of KS results from a pathological sequence in embryonic life , whereby premature interruption of the olfactory , vomeronasal and terminal nerve fibers in the frontonasal region disrupts the migration of neuroendocrine GnRH cells , which normally migrate from the nose to the brain along these nerve fibers [3] , [4] . What causes the primary failure of these fibers to establish proper contact with the forebrain is , however , still unknown . Since KS is genetically heterogeneous , identification of the various genes involved and the study of appropriate animal models are expected to provide valuable clues . Barely 30% of the KS patients have mutations in any of the eight genes known so far , specifically , KAL1 ( ID 3730 ) [5]–[7] , FGFR1 ( ID 2260 ) [8] , FGF8 ( ID 2253 ) [9] , PROKR2 ( ID 128674 ) , PROK2 ( ID 60675 ) [10] , WDR11 ( ID 55717 ) [11] , HS6ST1 ( ID 9394 ) [12] , CHD7 ( ID 55636 ) [13] , [14] , and current efforts thus concentrate on the identification of other genes that contribute to this disorder . One strategy is based on close pathohistological examination of targeted mutant mice that may reproduce the human KS phenotype . Here , we show that Nrp1sema/sema mutant mice , which are defective for the semaphorin-binding domain of the membrane coreceptor neuropilin-1 , have a KS-like phenotype , and we provide genetic evidence that insufficient semaphorin-3A signaling can contribute to the KS phenotype in man .
In the mouse , GnRH cells begin to leave the epithelium of the medial olfactory pit around embryonic day 11 . 5 ( E11 . 5 ) . They migrate in the frontonasal region in close association with growing fibers of the vomeronasal and terminal nerves , then penetrate into the rostral forebrain together with the central processes of these nerves , and continue their migration towards the hypothalamic region along a branch of the vomeronasal nerve that projects to the basal forebrain or along fibers of the terminal nerve itself [15]–[17] ( Figure 1A ) . Proper navigation of growing axons depends on guidance cues , which include semaphorins , a large and diverse family of secreted and membrane-associated proteins [18] . Among these , there is semaphorin-3A ( Sema3A ) , a secreted protein with repulsive effects on primary olfactory axons expressing the coreceptor neuropilin-1 ( Nrp1 ) [19]–[21] . The role of semaphorins in the navigation of vomeronasal/terminal axons and embryonic GnRH cells is still unclear , but previous studies in rodents have shown that migrating GnRH cells are morphologically associated with Nrp1-immunoreactive axons and are themselves immunoreactive [22] , [23] . Indeed , we were able to confirm these findings in E14 . 5 mouse embryos , and extend them to a 9-week old human fetus ( Figure 1B–1D ) , using specific antibodies to Nrp1 ( Figure S1 ) in immunohistofluorescence experiments . Notably , the caudal branch of the vomeronasal nerve that accompanies GnRH cells in their intracerebral path was also Nrp1-immunoreactive in the mouse embryos ( Figure 1C ) . These observations suggested that semaphorin signaling through Nrp1 imparts guidance information to axons of the vomeronasal neurons and migrating GnRH cells . We thus analyzed Nrp1sema/sema mutant mice that harbor inactivating aminoacid substitutions in the semaphorin-binding domain of Nrp1 . Unlike Nrp1−/− knockout mice , which die around E12 . 5 [24] , these mice survive until birth [25] . In Nrp1sema/sema newborn mice ( n = 4 ) , many axons of olfactory receptor neurons were stuck at the dorsal aspect of the cribriform plate and did not project to the olfactory bulb glomeruli ( Figure 2A ) . Olfactory cues are thought to play an important role in suckling behavior [26] . Analysis of six litters at postnatal day 1 ( P1 ) indeed showed that 7 out of 8 Nrp1sema/sema pups had little or no milk in their stomachs , whereas most Nrp1+/+ and Nrp1sema/+ littermates ( 18 out of 21 ) had full stomachs . These findings account for the decreased survival rate of homozygous , but not heterozygous , mutant pups [25] , and strongly suggest that the sense of smell is affected in Nrp1sema/sema mice . Most importantly , DiI axonal labeling at E14 . 5 showed abnormal projection of the vomeronasal nerve to the ventral forebrain in the homozygous mutant embryos ( n = 4 ) ( Figure 2B ) . Since this projection forms the axonal scaffold for the intracerebral migration of GnRH cells [17] , [27] , we analyzed the distribution of these cells in E14 . 5 and newborn mice . At E14 . 5 , a significant accumulation of GnRH cells in the nasal compartment and concomitant decreased cell number within the brain already indicated abnormal cell migration in the mutants ( n = 4 ) ( Figure 2E ) . In addition , while GnRH cells normally turn ventrally towards the basal forebrain , in Nrp1sema/sema embryos , many GnRH cells were found to migrate dorsally and medially towards the cortex and the thalamus , respectively , along aberrantly projecting axonal fibers ( Figure 2C , Figure S2 ) . Incidentally , conditional mutant mice that lack Nrp1 only in GnRH cells ( GnRH::cre; Nrp1loxP/loxP mice ) displayed a normal distribution of these cells between the nose and the brain at E14 . 5 as well as a normal number of these cells in the adult brain ( Figure S3 and data not shown ) , thus confirming that the defective migration we found in Nrp1sema/sema embryos is not a cell-autonomous trait . The migration defect was still conspicuous at birth ( Figure 2D ) , a time when neuroendocrine GnRH cells have completed their migration in normal mice [3] . The ventral forebrain region of Nrp1sema/sema newborn mice ( n = 4 ) indeed contained 38% fewer GnRH cells , which were dispersed , while there was a 36% increase in the number of GnRH cells detected in the rostral forebrain compared with Nrp1+/+ littermates ( n = 5 , p<0 . 01 for both comparisons ) ( Figure 2E ) . This GnRH-cell migration defect in Nrp1sema/sema animals resulted in decreased GnRH immunoreactivity in the median eminence of the hypothalamus ( Figure 2D ) , which is the projection field of neuroendocrine GnRH cells . Of the Nrp1sema/sema newborn mice , only four males and two females survived into adulthood . Both females had delayed pubertal activation , specifically , the first ovulation occurred more than 10 days later than in Nrp1sema/+ heterozygous littermates , and monitoring of the ovarian cycle from P60 showed that one female stayed in the diestrous stage ( a stage with low gonadotropin outputs ) throughout the 3-week study period , while the other female had disrupted ovarian cyclicity ( data not shown ) . Male reproductive capacity was assessed by breeding the young adult ( P90 ) Nrp1sema/sema males with confirmed wild-type dams , and monitoring the occurrence of litters over 10–13 months . While Nrp1sema/+ males ( n = 4 ) produced about one litter per month , as did Nrp1+/+ males , the fertility index ( number of litters per month ) was markedly reduced in the Nrp1sema/sema males , which only gave birth to 2 to 4 litters ( fertility index: 0 . 29±0 . 04 vs . 1 . 08±0 . 12 in Nrp1sema/+; Student's t-test , p<0 . 001 ) . Moreover , neuroanatomical analysis of Nrp1sema/sema adult brains showed significantly reduced GnRH cell populations in the preoptic and hypothalamic regions ( 384±67 GnRH cells , n = 4 ) compared to Nrp1sema/+ littermates ( 767±49 GnRH cells , n = 4; Student's t-test , p<0 . 001 ) , whereas Nrp1sema/+ mice did not differ from Nrp1+/+ mice ( 701±11 GnRH cells , n = 4; Student's t-test , p>0 . 05 ) . Therefore , the GnRH cell migration defect found in Nrp1sema/sema mouse embryos was not corrected during later development , and caused subfertility in adult homozygous mutants . The KS-like phenotype of Nrp1sema/sema mice , and that , even more pronounced , of Sema3a−/− mice [22] , prompted us to ask whether insufficient Sema3A signaling through Nrp1 might also be involved in the human disorder . We sought mutations , by Sanger sequencing , in the 17 coding exons of SEMA3A ( ID 10371 ) and flanking splice sites , in 386 unrelated KS patients ( 297 males and 89 females ) . All of them had confirmed hypogonadotropic hypogonadism and anosmia or hyposmia , and some already harbored a mutation in one of the five KS genes we had previously analyzed , specifically , in KAL1 ( 13 patients ) , FGFR1 ( 30 patients ) , FGF8 ( 3 patients ) , PROKR2 ( 30 patients ) , or PROK2 ( 12 patients ) . Nonsynonymous mutations in SEMA3A were found in 24 patients ( 20 males and 4 females ) , all in heterozygous state ( Table 1 ) . They consist of a frameshifting deletion of 14 nucleotides ( c . del1613_1626; p . D538fsX31 ) , and seven different missense mutations ( p . R66W , p . N153S , p . I400V , p . V435I , p . T688A , p . R730Q , p . R733H ) that affect evolutionarily conserved aminoacid residues located in different domains of the protein ( Figure 3 ) . In addition , the p . R730Q and p . R733H mutations , which both remove basic residues in the C-terminal basic motif of Sema3A , are predicted to affect in vivo proteolytic processing by furin-like endoproteases at residue R734 [28] . Notably , all the missense mutations , but not the frameshifting mutation , have been reported in the Exome Variant Server database , with allele frequencies in the European American population below 0 . 03% except for p . N153S ( 0 . 4% ) and p . V435I ( 1 . 3% ) . Three of these mutations ( p . R66W , p . V435I , p . R730Q ) were also detected in our sample of 386 unrelated Caucasian controls ( see Table 1 ) . We thus studied the effects of the eight mutations on the signaling activity of Sema3A using the GN11 cell line , derived from murine embryonic GnRH cells [29]–[31] , and conditioned media from transfected COS-7 cells producing Sema3A either from the wild-type SEMA3A cDNA or from cDNAs harboring the mutations . We found that the conditioned medium from COS-7 cells transfected with the wild-type SEMA3A cDNA was as potent at inducing phosphorylation of FAK ( focal adhesion kinase ) and ERK1/2 ( extracellular signal-regulated kinases 1 and 2 ) in GN11 cells as the purified recombinant human Sema3A ( 100 µg/L ) . By contrast , Sema3As harboring the N153S , I400V , T688A , or R733H missense mutations were ineffective , despite normal production and secretion of the proteins by COS-7 cells , shown by western blot analysis of the conditioned media . The R66W and V435I mutant proteins were not detected in the conditioned medium , which indicates defective secretion . Likewise , the c . del1613_1626 ( p . D538fsX31 ) frameshifting mutation resulted in the absence of protein secretion , as expected ( Figure 4 ) . From these results , we were able to conclude that all the mutations , except p . R730Q , are loss-of-function mutations that affect the secretion or signaling activity of Sema3A , which strongly argues in favor of their pathogenic effect in the KS patients . In addition , the p . R730Q mutation may still have a pathogenic effect not detected in our experimental system , especially since this mutation is expected to impair proteolytic processing of Sema3A in vivo , as mentioned previously . Notably , the patients carrying the p . T688A and p . I400V mutations , and three patients carrying the p . V435I mutation also carry , in heterozygous state , p . Y217D , p . R268C ( two patients ) , p . H70fsX5 , and p . G687N pathogenic mutations in KAL1 , PROKR2 , PROK2 , and FGFR1 , respectively ( Table 1 ) , which further substantiates the digenic/oligogenic mode of inheritance of KS [1] , [2] . Based on the seemingly normal reproductive phenotype of Sema3a+/− heterozygous mice [21] , [22] , we suggest that the monoallelic mutations in SEMA3A are not sufficient to induce the abnormal phenotype in the patients , but contribute to the pathogenesis of KS through synergistic effects with mutant alleles of other disease-associated genes . Accordingly , the other KS patients who carry monoallelic mutations in SEMA3A are also expected to carry at least one pathogenic mutation in another gene ( see footnote ) . Although NRP1 ( ID 8829 ) might be viewed as one of the best candidates , we did not find a mutation within its 17 coding exons and flanking splice sites in any of these patients , nor did we in a group of 100 KS patients without SEMA3A mutations , which indicates that mutations in NRP1 , if any , are infrequent . It is also possible that some of the additional mutations affect other proteins involved in Sema3A-signaling , such as members of the plexin family of transmembrane receptors or neuropilin-2 [18] , [22] . A whole-exome sequencing strategy should prove useful to explore the spectrum of genes which , when mutated , can lead to a KS phenotype in conjunction with SEMA3A mutations . While this article was under review , Young et al . reported the coexistence of KS and a large deletion in SEMA3A , in heterozygous state , in two siblings and their clinically affected father ( Hum . Reprod . , 2012; 27:1460–1465 ) . Our findings do not support mere autosomal dominant Mendelian inheritance in this family , and suggest that another , as yet unidentified genetic hit combines with SEMA3A haploinsufficiency to produce the disease phenotype .
This study was approved by the national research ethics committee ( agence de biomédecine , Paris , France ) . All experiments on mice were carried out in accordance with Directive 86/609/EEC of the Council of the European Communities regarding the mammalian research and French bylaw . Nrp1sema/+ mice ( B6 . 129 ( C ) -Nrp1tm1Ddg/J ) [25] were purchased from the Jackson laboratory ( Maine , USA ) , maintained on a controlled 12 h∶12 h light cycle , provided with food and water ad libitum , and genotyped as described previously [25] . E14 . 5 ( plug day , E0 . 5 ) , P0 , and adult Nrp1+/+ , Nrp1sema/+ and Nrp1sema/sema mice were obtained and processed for immunohistofluorescence analyses as previously described [30] . In addition , homozygous Nrp1loxP/loxP mice ( B6 . 129 ( SJL ) -Nrp1tm2Ddg/J ) [25] from the Jackson laboratory were crossed with a transgenic mouse line expressing the cre recombinase under the control of the GnRH gene promoter ( GnRH::cre mice ) [32] , a gift from C . Dulac ( Harvard university , Cambridge , USA ) , to obtain GnRH::cre; Nrp1loxP/loxP mice that lack Nrp1 in GnRH cells only . Nrp1loxP/loxP and GnRH::cre; Nrp1loxP/loxP mice were used for immunohistofluorescence analyses at E14 . 5 and adult stages . The human fetus was obtained from a voluntary terminated pregnancy , with parent's written informed consent . Gestational age was established by crown-rump length measurement . The fetus was fixed in 4% paraformaldehyde in 0 . 1 M phosphate buffered saline ( PBS ) , pH 7 . 4 , for three weeks at 4°C , and then immersed in 0 . 1 M PBS containing 30% sucrose for two days at 4°C . The head was embedded in OCT embedding medium ( Tissue-Tek ) , frozen , and sagittal cryosections ( 20 µm thick ) were cut and processed for immunohistofluorescence . Immunohistofluorescence experiments were carried out as described previously [30] . Primary antibodies were: rabbit anti-GnRH ( dilution 1∶3000 ) , a gift from G . Tramu ( University of Bordeaux , France ) ; rabbit anti-peripherin ( dilution 1∶1000 ) , AB1530 ( Millipore ) ; goat anti-neuropilin1 ( dilution 1∶400 ) , AF566 ( R & D systems ) ; goat anti-olfactory marker protein ( dilution 1∶6000 ) , a gift from F . L . Margolis ( University of Maryland , Baltimore , USA ) . Vomeronasal nerve fibers were traced anterogradely with the lipophilic fluorescent dye DiI ( 1 , 1′-dioctadecyl-3 , 3 , 3′ , 3′-tetramethylindocarbocyanine perchlorate , Molecular Probes ) as previously described [17] . After diffusion of the tracer , serial sagittal sections ( 100 µm thick ) were cut through the forebrain , and analyzed using a LSM 710 confocal microscope ( Zeiss ) and the ImageJ analysis software ( NIH , Bethesda , USA ) . COS-7 cells and GN11 cells were grown in monolayers in 5% CO2 at 37°C , in Dulbecco's modified Eagle's medium ( Life Technologies , Inc . ) containing 1 mM sodium pyruvate , 2 mM glutamine , 50 mM glucose , and supplemented with 10% fetal bovine serum ( Invitrogen ) , 100 µg/ml streptomycin and 100 U/ml penicillin . A cDNA containing the entire coding region of the human SEMA3A ( GenBank NM_006080 ) was inserted into a pRK5 plasmid expression vector . Recombinant plasmids containing SEMA3A cDNAs harboring each of the eight mutations identified in the KS patients were then engineered using the QuickChange mutagenesis protocol ( Stratagene ) . COS-7 cells were transiently transfected using a fast-forward protocol ( Lipofectamine 2000 , Invitrogen ) [30] . Conditioned medium was collected 48 h after transfection , tested for the presence of Sema3A by western blot analysis using an anti-Sema3A antibody ( Santa Cruz , sc-10720 , dilution 1∶100 ) , and then processed for signaling activity experiments in the GN11 cell line . Briefly , subconfluent GN11 cells were grown overnight in serum-free medium , and then stimulated for 20 min with human recombinant Sema3A ( R&D systems ) at 100 µg/L , or with the concentrated conditioned media from transfected COS-7 cells . Western blot experiments [30] were carried out on cell lysates using antibodies to P-ERK ( #9101L ) and ERK ( #9102L ) from Cell Signaling ( dilution 1∶1000 ) , or P-FAK ( sc56901 ) and FAK ( sc81493 ) from Santa Cruz ( dilution 1∶500 ) . Informed consent was obtained from all individuals analyzed . Genomic DNAs were prepared from white blood cells using a standard procedure . Each of the SEMA3A and NRP1 coding exons and flanking splice sites was PCR-amplified from genomic DNA using a specific primer pair ( see Tables S1 and S2 for primer sequences ) , and sequenced using either PCR oligonucleotide as sequencing primer . The mutations were confirmed by sequencing two independent PCR products on both DNA strands . Exons 2 , 5 , 11 , 14 , and 17 of SEMA3A , which harbor the mutations identified in some patients , were analyzed by denaturing high performance liquid chromatography ( DHPLC ) scanning on an automated HPLC instrument ( Wave technology ) in 386 unrelated Caucasian controls , followed by Sanger sequencing of the exon in case of abnormal DHPLC profile .
|
Kallmann syndrome is a hereditary developmental disease that affects both the hormonal reproductive axis and the sense of smell . There is a developmental link between the reproductive and olfactory disorders: neuroendocrine cells producing the gonadotropin-releasing hormone that is deficient in the patients normally migrate from the nose to the forebrain along olfactory nerve fibers during embryonic life , and they fail to do so in the patients . Affected individuals usually do not undergo spontaneous puberty . Hormone replacement therapy is the treatment to initiate virilization in males or breast development in females and later to develop fertility in both sexes . This is a genetically heterogeneous disease . Mutations in any of eight causative genes identified so far have been found in approximately 30% of the affected individuals , thus indicating that other genes remain to be discovered . We report on the identification , in 6% of the KS patients , of various loss-of-function mutations in the gene coding for semaphorin-3A , a secreted protein involved in the navigation of olfactory nerve fibers during embryogenesis . The fact that many of these mutations were also detected in clinically unaffected individuals indicates that they must combine with other genetic defects to produce the disease phenotype .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"endocrinology",
"genetics",
"molecular",
"genetics",
"biology",
"human",
"genetics",
"neuroendocrinology",
"genetics",
"of",
"disease",
"diabetes",
"and",
"endocrinology",
"genetics",
"and",
"genomics"
] |
2012
|
SEMA3A, a Gene Involved in Axonal Pathfinding, Is Mutated in Patients with Kallmann Syndrome
|
The methylation of cytosines in CpG dinucleotides is essential for cellular differentiation and the progression of many cancers , and it plays an important role in gametic imprinting . To assess variation and inheritance of genome-wide patterns of DNA methylation simultaneously in humans , we applied reduced representation bisulfite sequencing ( RRBS ) to somatic DNA from six members of a three-generation family . We observed that 8 . 1% of heterozygous SNPs are associated with differential methylation in cis , which provides a robust signature for Mendelian transmission and relatedness . The vast majority of differential methylation between homologous chromosomes ( >92% ) occurs on a particular haplotype as opposed to being associated with the gender of the parent of origin , indicating that genotype affects DNA methylation of far more loci than does gametic imprinting . We found that 75% of genotype-dependent differential methylation events in the family are also seen in unrelated individuals and that overall genotype can explain 80% of the variation in DNA methylation . These events are under-represented in CpG islands , enriched in intergenic regions , and located in regions of low evolutionary conservation . Even though they are generally not in functionally constrained regions , 22% ( twice as many as expected by chance ) of genes harboring genotype-dependent DNA methylation exhibited allele-specific gene expression as measured by RNA-seq of a lymphoblastoid cell line , indicating that some of these events are associated with gene expression differences . Overall , our results demonstrate that the influence of genotype on patterns of DNA methylation is widespread in the genome and greatly exceeds the influence of imprinting on genome-wide methylation patterns .
Methylation of the 5 carbon of a large number of cytosines in the genome is necessary in mammalian development [1] . Aberrant patterns of DNA methylation have been reported in a wide variety of human diseases , including cancer [2] , [3] , psychiatric disorders [4] , autoimmune diseases [5] and diabetes [6] . Some of these patterns are indicative of underlying functional changes that have occurred during disease progression and shed light on genes involved in pathogenesis . One of the most fascinating discoveries surrounding DNA methylation is the observation that two homologous chromosomes can be differentially methylated . Differential methylation of homologous chromosomes can be the result of epigenetic phenomena such as gametic imprinting [7] , [8] or X chromosome inactivation [9] , [10] . DNA sequence , or genotype , may also play a role in establishing differential methylation , as a few well-established cases have been identified in which a locus' DNA methylation state clearly depends on an individual's DNA sequence [11] , [12] . Recent advances in DNA sequencing technology have opened the door to exploring differential methylation on homologous chromosomes with high accuracy and detail . It is now possible to examine the prevalence of genetic versus epigenetic causes of differential methylation with unprecedented precision and thoroughness . To distinguish the impact of gametic imprinting vs . genotype on DNA methylation , the inheritance patterns of alleles along with corresponding methylation levels should be observed . Recent studies have suggested that the majority of differential DNA methylation on homologous chromosomes is sequence-dependent and not the result of gametic imprinting , as the same allele has the same influence on DNA methylation in unrelated individuals [13]–[15] . However , to differentiate definitively between genetic inheritance and imprinting , analysis of DNA methylation in primary tissues from a family is necessary . Analysis of a family allows for the determination of a SNP's parental origin along with inheritance patterns of DNA methylation levels and therefore permits the direct examination of genetic and epigenetic mechanisms of differential methylation . By analyzing DNA methylation in a family , the impact of alleles verses the impact of a chromosome's parental origin on the inheritance of methylation can be clearly resolved . Surveying DNA methylation in multiple related and unrelated individuals also enables quantitative estimation of the effect of genotypic variation on DNA methylation levels . Examination of DNA methylation patterns in twins has indicated that related individuals exhibit more similar DNA methylation than unrelated individuals , implying that genetics plays a role in establishing DNA methylation patterns . Baranzini et al . [16] observed a striking similarity in DNA methylation patterns between monozygotic twins that was not seen between unrelated individuals . In addition , Kaminsky et al . [17] found that monozygotic twins share similar DNA methylation patterns compared to dizygotic twins; however , they suggest that this similarity may be due to epigenetic events in the zygote as opposed to genetic relatedness . Considering the promise of DNA methylation as a disease biomarker [18] , it is important to determine the influence that genome sequence has on this epigenetic mark . To determine the genetic contribution to DNA methylation , we quantified DNA methylation in non-immortalized peripheral blood leukocytes from a three-generation family . We used reduced representation bisulfite sequencing ( RRBS ) [19] to achieve single molecule resolution of DNA methylation for a large number of CpG dinucleotides distributed throughout the human genome . We were able to find SNPs by sequencing and directly observing DNA methylation on the same homologous chromosome . Our results show that differential methylation of homologous chromosomes is prevalent , transmitted through families in a genotype-dependent manner , and linked with allele-specific gene expression .
To discover differential methylation on homologous chromosomes , we used reduced representation bisulfite sequencing ( RRBS ) [19] , which can be used to measure the DNA methylation state in a subset of the genome in many samples . Because RRBS uses bisulfite treatment , it detects both 5-methylcytosine and 5-hydroxymethylcytosine [20] . Thus , the allelic differences identified by the method can be differences in 5-methylcytosine or 5-hydroxymethylcytosine . However , alleles that differ in the ratio between levels of 5-methylcytosine and 5-hydroxymethylcytosine are not detectable . We refer to the combination of these marks as DNA methylation . We used the Illumina Genome Analyzer IIx ( GAIIx ) to sequence 36 base pair ends of bisulfite-treated MspI digestion fragments ranging in length from 40 to 120 bp . For each sample , we assayed approximately 1 million CpG dinucleotides with at least 10 sequencing reads , and calculated the percent of reads that are methylated at each CpG . This type of deep sequencing also identifies DNA sequence variants in the fragments for which we are measuring methylation ( see Materials and Methods ) . Detecting these SNPs in the bisulfite sequence reads allows direct determination of whether a particular allele is found in cis with methylated or unmethylated CpGs ( Figure 1A ) . Bisulfite treatment obscures C to T SNPs in forward strand reads and G to A SNPs in reverse strand reads , because a C to T mismatch in forward strand reads could represent an unmethylated cytosine and cannot be unambiguously called a SNP . We do not analyze possible C to T SNPs in our sequence reads because they cannot be identified . These SNPs account for approximately 30% of human SNPs; thus , most SNPs remain observable . We first measured DNA methylation on different alleles in the human embryonic stem cell line H1 [21] . The majority of alleles show no cis association with DNA methylation . At a 5% false discovery rate ( FDR ) , 5 . 4% ( 1 , 340 of 24 , 979 ) of autosomal heterozygous SNPs are associated with allele-specific methylation ( ASM ) . In total , 1 , 340 SNPs are associated with ASM at 1 , 574 CpGs . Because a SNP can be linked to multiple CpGs and multiple SNPs can be linked to the same CpG , we were able to identify 1 , 937 ASM events in this cell line ( all ASM data can be found in Dataset S1; Table S1 provides a summary of ASM events in each sample , including the number of heterozygous SNPs called ) . Of the 1 , 937 ASM events , 573 ( 29 . 6% ) involve a SNP that mutates a CpG . These cases represent a trivial mechanism for generating ASM because there is no longer a CpG to methylate when the CpG-disrupting allele is present; however , they are functional variants in that they result in a change in the methylation status of a locus . The average difference in percent methylation between alleles is 59 . 8% for ASM events that do not involve CpG disrupting SNPs ( Figure 1B ) . Less than 8% of ASM events exhibit greater than a 95% difference in percent methylation , which suggests that most changes in genotype induce more subtle changes in the frequency of DNA methylation than complete reversal of CpG methylation . These quantitative differences in percent methylation between alleles are highly reproducible . In two separate growths of the H1 cell line , the correlation of the difference in percent methylation between the reference and variant allele for all ASM events in both replicates is 0 . 98 ( Figure 1C ) . These results show that measurements of the quantitative differences in percent methylation associated with SNPs are highly reproducible . One of the two X chromosomes in women are randomly silenced in each cell during development , and DNA methylation at many sites has been shown to exist specifically on alleles from the inactive X [10] . To assess our ability to observe inactivation of the X chromosome and thus validate the ASM approach taken here , we analyzed ASM in four clonal cell lines derived from single cells of the EBV-transformed lymphoblastoid line GM12878 . Gene expression patterns indicate that two of these cell lines silence the paternal X chromosome , while the other two cell lines silence the maternal X chromosome ( Table S2; K . S . K . and H . F . W . , manuscript in preparation ) . We observed reproducible patterns of ASM on the X chromosome ( Figure S1 ) . Clones with an inactivate X chromosome from the same parental lineage exhibited the same allelic bias for every ASM event on the X chromosome ( e . g . , either the variant allele was more methylated in each clonal line or the reference allele was more methylated in each clonal line ) . When comparing any maternal X-inactive clone with any paternal X-inactive clone , an average of 59 . 7% of ASM events on the X chromosome switch the allele that is most often methylated , which is consistent with the pattern of inactivation . Of the genes nearby the ASM events that are consistent with inactivation , six were previously assayed for allelic expression [22] and all six were identified as being inactivated . Because inactivation is not complete across the chromosome , we do not expect all ASM to switch the most often methylated allele when the inactive X chromosome is switched . These results demonstrate that our method is sensitive to X inactivation in single cell clones; however , many ASM events on the X chromosome are driven by allele status rather than by parental origin . To determine the prevalence of ASM throughout the human genome and to deconvolute whether it is dependent on DNA sequence variants or parental origin imprinting , we applied RRBS to DNA extracted from leukocytes in fresh blood from six members of a three-generation family as well as two unrelated individuals ( Figure 2A ) . We collected data for more than 950 , 000 CpGs that had at least ten sequencing reads in each family member . Overall , 859 , 531 CpGs had at least 10 sequencing reads in all samples and 668 , 545 had at least 20 sequencing reads in all samples . Leukocytes from fresh blood are comprised of a heterogeneous population of cells , consisting primarily of neutrophils , lymphocytes and monocytes . It has been shown that the relative proportion of each cell type does not considerably affect DNA methylation levels [23] . Consistent with this observation , we found that the patterns of DNA methylation are strikingly similar between the family members . The average correlation in percent methylation of all autosomal CpGs assayed between any two family members is 0 . 985 . All of the samples assayed were highly similar , indicating that there were no systematic biases in DNA isolation , library preparation or DNA sequencing . Analysis of the CpGs that vary within the eight individuals reveals that DNA methylation patterns recapitulate relatedness . We clustered individuals based on the methylation status of the top 237 most varying autosomal CpGs across all samples . The resulting tree , based on hierarchical clustering , is shown in Figure 2B . The two unrelated individuals are separated from the family members and serve as an out-group . Genetically unrelated pairs of family members ( e . g . , maternal grandmother and father , and father and mother in the middle generation ) are also separated in the tree . The closest relationships in DNA methylation patterns exist between the grandmothers and their children and the methylation patterns in the grandchildren are mixtures of their parents' methylation patterns . These results indicate that DNA methylation levels can capture the relatedness of individuals , suggesting that there is a strong genetic component to DNA methylation levels . We identified an average of 1 , 702 ASM events for each member of the family . When individuals are clustered on CpGs that exhibit ASM in at least one individual ( Figure 2C ) , the same relationships are observed as those shown in Figure 2B . To look at features of CpGs exhibiting ASM , we focused on 2 , 391 autosomal ASM events that we identified in at least two of the six family members . 42 . 6% ( 1 , 018 ) of these ASM events are SNPs that mutate a CpG . The high prevalence of CpG-disrupting SNPs is not due to higher sequence coverage ( Table S3 ) . Our analysis showed that SNPs that disrupt CpGs tend to lie in intergenic regions that are not CpG islands . Only 21 . 9% of CpGs that are disrupted by SNPs are found in CpG islands , while 65% of all CpGs queried by RRBS reside in CpG islands ( P<2 . 2×10−16; Fisher's exact test ) . CpGs that are disrupted by SNPs are over-represented in intergenic regions , as 44 . 1% are at least 2 kilobase pairs ( kb ) away from the nearest gene , compared to 22% of all CpGs assayed by RRBS that are at least 2 kb away from the nearest gene ( P<2 . 2×10−16; Fisher's exact test ) . SNPs that mutate a CpG are often linked to the methylation status of other CpGs in the immediate vicinity . More than 45% of CpG-mutating ASM SNPs exhibit ASM associated with at least one other CpG within 36 base pairs , which represents a 5 . 5-fold increase in the chance that a CpG with a CpG-disrupting SNP in the same read shows allele specificity ( P<2 . 2×10–16; Fisher's exact test ) . These data demonstrate that , while CpG-disrupting SNPs alter DNA methylation at their particular position due to a DNA sequence difference , they also influence the methylation of nearby CpGs . This is further evidence that there is a strong genetic effect on the regulation of DNA methylation . ASM events that do not involve a mutated CpG are under-represented in CpG islands and are most often located in intergenic regions , as were CpGs disrupted by SNPs . Only 49 . 4% of CpGs that exhibit ASM are present in CpG islands , even though 65% of CpGs with nearby SNPs assayed by RRBS are in islands ( Figure 3A; P<2 . 2×10−16: Fisher's exact test ) . Figure 3B shows the distribution of gene features for CpGs that exhibit ASM . The majority of ASM CpGs are present in intergenic regions , even though most CpGs with nearby SNPs assayed by RRBS are present in promoters , first introns or first exons . There is a 3 . 5-fold increase in the fraction of CpGs that reside in intergenic regions ( P<2 . 2×10−16; Fisher's exact test ) . The overall trends of where ASM events are located in the genome are not the result of differences in coverage between these categories of SNPs ( Table S3 ) . The tendency for ASM CpGs to be located outside of CpG islands and further away from genes may indicate that ASM events occur in regions under less selective pressure . Because ASM tends to occur in regions without clear regulatory function , we looked at evolutionary constraint at these sites by analyzing mammalian alignments . ASM events involving SNPs that disrupt CpGs and SNPs that do not disrupt CpGs were found more often outside of CpG islands and in intergenic regions . We therefore assayed both sets and determined the level of evolutionary conservation in each group . We found significantly less conservation in CpG-disrupting ASM events ( P = 0 . 008; Wilcoxon test ) and non-CpG-disrupting ASM events ( P = 0 . 002; Wilcoxon test ) compared to all assayed regions that contained SNPs . To determine if this reduction in conservation could be explained by the overabundance of ASM events in intergenic regions , we repeated the conservation analysis using only SNPs in intergenic regions . We found that even within intergenic regions , SNPs associated with ASM were significantly less conserved than all intergenic SNPs assayed ( Figure 3C ) . This was true for both CpG-disrupting ASM events ( P = 0 . 009; Wilcoxon test ) and non-CpG-disrupting ASM events ( P = 0 . 007; Wilcoxon test ) . These data are consistent with the hypothesis that there is evolutionary constraint on DNA methylation levels , as genetic variants that affect DNA methylation tend to lie in regions under less selective pressure . By carrying out RRBS in a family , we were able to follow SNPs through each generation and determine whether ASM is associated with the identity of the SNP or with parental origin . To distinguish these modes of transmission , we first had to identify SNPs with unambiguous inheritance in the pedigree that switched parental origin ( e . g . , those SNPs that have a maternal origin in one generation and a paternal origin in the next ) . We then identified those SNPs that were associated with ASM in both parent and child . We identified 432 autosomal ASM events in which the variant allele switched parental origin in the family . We then eliminated the 341 of these ASM events that were due to SNPs that mutated a CpG , as these are necessarily associated with the identity of the SNP . Only 7 ( 7 . 7% ) of the remaining 91 ASM events displayed a methylation pattern that followed parental origin . The methylation patterns of the other 84 ASM events depended on the DNA sequence and an example is shown in Figure 4A . These results indicate that , in most cases , an allele's sequence plays a larger role in determining DNA methylation than an allele's parental origin . While the majority of ASM events depend on the underlying allelic sequence , we found seven autosomal ASM events that are consistent with a genetic imprinting mode of transmission . These events occur in five distinct loci , most of which are intergenic . The introns of OVOS1 and TRAPPC9 both contain parent-of-origin ASM events , while the other three loci are at least 20 kb from the nearest gene . Figure 4B shows ASM of TRAPPC9 ( also know as NIBP ) , which is involved in neuronal NF-κB signaling [24] and is thought to have a maternal effect on height [25] . The TRAPPC9 SNP mutates a CpG , which exhibits ASM that follows the sequence of the allele as expected . However , the CpG located 32 bp away is unmethylated on the variant allele in the mother ( which came from her father ) and is methylated on the variant allele in both sons . The RRBS data suggests that parent-of-origin ASM may extend further in this region , as CpGs extending 65 bp upstream and 125 bp downstream of the SNP are all close to 50% methylated in every member of the family . The transcription start site of KCNK9 , a known maternally imprinted gene [26] , resides downstream of TRAPPC9 more than 350 kb away from the SNP . It is possible that methylation of a TRAPPC9 intron is involved in silencing KCNK9 . The TRAPPC9 region represents a candidate for maternal imprinting , which provides supporting evidence for the locus' role in maternal effect on height . Because genotype-dependent differences in DNA methylation were far more prevalent than parental origin-dependent differences , we built a general model of DNA methylation and genotype that was able to incorporate information from homozygous individuals . We built a linear model of the relationship between DNA methylation and genotype that assumes that a CpG's methylation level in a heterozygous individual is halfway between the level of methylation in an individual homozygous for one allele and an individual homozygous for the other allele . When the model is constructed for the top 2 , 900 most variable CpGs in the family that have detectable SNPs nearby , 602 ( 20 . 8% ) showed significant genotype contributions and 370 ( 12 . 8% ) were still significant after multiple hypothesis correction . Even with a small sample size of six , genetic association with DNA methylation is detectable for a large number of loci . These results , together with ASM data , suggest that DNA sequence plays a major role in determining inter-individual differences in DNA methylation levels . Some of the genetically linked ASM events found in the family may be dependent on genetic background and specific to the particular family studied , while other ASM events may be common to many individuals . To determine the persistence of ASM in different genetic backgrounds , we analyzed two unrelated individuals as well as the lymphoblastoid cell line GM12878 . We could test 40 of the 84 sequence-dependent autosomal regions that we observed in the family because at least one unrelated individual was heterozygous at the same SNP as that observed in the family . 30 of those 40 regions ( 75% ) were found to be allele-specific in the unrelated individuals as well . In every case , the more often methylated allele in the family was also more often methylated in the unrelated individuals . We also found that the CpGs in the TRAPPC9 locus discussed above were also methylated between 39% and 57% in the unrelated individuals , which is consistent with parental origin imprinting at this locus . These results indicate that many ASM events that we observed in the family are present in other genetic backgrounds . For a broader view of the influence of genotype on DNA methylation , we used the linear model of DNA methylation and genotype discussed above to predicted DNA methylation levels in the two unrelated individuals . In both individuals , the model was able to predict methylation levels with high accuracy based on each individual's genotype ( R2 = 0 . 82 and 0 . 83 ) . The ability of the linear model , which was trained on the family , to predict DNA methylation in unrelated individuals accurately shows that the relationship between DNA sequence and methylation levels is consistent among individuals . While the influence of genotype could be observed across individuals in the same tissue , we sought to determine whether the impact of DNA sequence on DNA methylation could be observed in different tissues from the same individual . We analyzed DNA extracted from kidney and skeletal muscle of a third unrelated individual . There were 1 , 521 ASM events in the kidney sample and 1 , 332 ASM events observed in the skeletal muscle sample . Of the 984 ASM events observed in the kidney with sufficient coverage in both samples , 834 ( 84 . 8% ) were also allele-specific in skeletal muscle . For all 834 overlapping ASM events , the more often methylated allele was identical in the kidney and skeletal muscle samples . If CpG disrupting SNPs are removed from the analysis , there were 931 ASM events in the kidney sample and 790 events observed in the skeletal muscle sample . Of the 614 ASM events observed in the kidney with sufficient coverage in both samples , 445 ( 72 . 5% ) were also allele-specific in skeletal muscle . These results indicate that allelic differences in DNA methylation can be observed across different tissues in the same individual at a high frequency . To determine whether ASM events are indicative of functional gene regulatory differences , we compared ASM with allele-specific gene expression ( ASE; T . E . R . et al . , manuscript in preparation ) in the original lymphoblastoid line GM12878 . Of the autosomal genes that have an ASM event within 5 kb of the transcription start site , we found that 21 . 7% ( 18 out of 83 ) of the autosomal genes with sufficient RNA-seq read depth ( greater than 25 reads that cover SNPs ) show ASE ( Table S4 ) . This represents more than a two-fold enrichment ( P = 4 . 2×10−4; hypergeometric test ) over an overall rate of 9 . 2% ( 594 out of 6464 ) of ASE for all autosomal genes with sufficient read depth . Data for the gene encoding acid alpha-glucosidase ( GAA ) are summarized in Figure 5 . GAA , which is involved in the degradation of glycogen to glucose and is associated with Pompe's disease [27] , showed 3% methylation on the maternally inherited copy of chromosome 17 and 89% methylation on the paternally inherited copy; RNA-seq showed that 76 . 7% of the transcripts came from the maternal copy of chromosome 17 . TRAPPC9 and OVOS1 did not harbor SNPs with sufficient read depth in the RNA-seq data to determine allele-specificity . These results indicate that genes exhibiting ASM are enriched for gene expression differences between alleles . We sought to validate the next generation sequencing results by performing Sanger sequencing on a small number of loci . To investigate ASM , we performed bisulfite PCR , cloning and Sanger sequencing of four loci from family member C ( Figure 6 ) . All four loci exhibited ASM in the Sanger sequencing data . We also observed that the allele-specificity of DNA methylation extended beyond the region assayed by RRBS . In two cases , ( chr21:4415835 and chr8:141109575 ) allele-specificity was found across the entire region . In the other two examples , half of the assayed region exhibited allele-specificity , while the other half of the region was similar between the two alleles . These results indicate that ASM identified in the RRBS data is also observed with Sanger sequencing and that patterns of ASM are complex . To look further at extended ASM , we analyzed SNPs within 200 bp of each other and found that adjacent SNPs were often concurrently associated with ASM . For example , in family member A 655 ASM SNPs have another SNP detected within 200 bp and 455 ( 71 . 4% ) of these SNPs also exhibit ASM . On average across the family members , 76 . 3% of SNPs adjacent to ASM SNPs were also associated with ASM themselves . These results show that ASM can be seen across extended regions . To validate the next generation ASE findings , we cloned and Sanger sequenced genomic DNA and cDNA from GM12878 for six loci . Using the ratio of alleles observed in the genomic DNA as a background , we found that five ( GAA , KCNQ10T1 , HLA-DPB2 , LOC654433 and LOC253039 ) loci exhibited significant allele biased expression with the bias coming from the expected parental chromosome . ZNF132 showed a bias in the expected direction ( 69% from the allele predicted to be higher expressed ) ; however , the number of reads ( 16 ) , was too low to call the bias significant . Both ASE and ASM validation results show that the allele-specificity observed with our next generation sequencing approaches are replicated with an independent technique .
Using short-read , ultra high-throughput DNA sequencing , we identified reproducible quantitative differences in DNA methylation between alleles . When we observe allele-specific DNA methylation , the differences in methylation levels between homologous chromosomes are rarely completely reversed . The abundance of subtle differences indicates that some alleles influence the propensity of DNA methylation within a cell population , but do not completely exclude or cause methylation . ASM events were found outside of CpG islands and were highly enriched in intergenic regions . Because ASM tends to occur in regions without clear regulatory function , we looked at evolutionary constraint at these sites . ASM tends to be found in genomic locations with significantly low levels of evolutionary conservation . This result is consistent with the hypothesis that there is selective pressure to maintain DNA methylation , as we found that most alleles that affect DNA methylation exhibit less evolutionary constraint . Selection may be acting on the underlying events that lead to DNA methylation or directly on the ability to methylate a particular sequence , as is the case with SNPs that mutate CpGs . We found that DNA methylation levels could distinguish family members from unrelated individuals and capture family relationships . By analyzing six members of a three-generation family , we were able to follow the patterns of inheritance of both DNA methylation and the SNPs associated with methylation levels directly . When the parental origin of a SNP switched between generations , the vast majority of ASM events were genotype-dependent and followed a particular sequence variant rather than parental origin . These results show that DNA sequence plays a larger role in establishing DNA methylation patterns than do parental origins . We found that the majority of the ASM events seen in the family could also be observed in other individuals , indicating that the ASM events observed are not specific to the family described here and that the alleles have the same influence on DNA methylation in different genetic backgrounds . The strong association between genotype and DNA methylation indicates that genetics plays a prominent role in the establishment of DNA methylation patterns . Our data supports a non-Lamarckian model of evolution , where genetic variants , as opposed to environment , shape epigenetics [28] . These genetic variants may not lead directly to phenotypic differences , but may cause phenotypic variability through changes in epigenetic states . While ASM events could be observed across individuals in the same cell type , we also observed a concordance of ASM between tissue types . The fact that we see a strong overlap between ASM in different tissues indicates that these allelic differences are most likely due either to shared gene regulatory events that occur early in development or an inherent property of DNA sequence that directly affects the propensity of DNA methylation . The prevalence of ASM in different tissues brings up the possibility that methylation at these loci are directly inherited with the haplotype through the germline . While the prevailing model of DNA methylation would suggest that methylation patterns are erased during gamete formation and just after fertilization [29] , the possibility exists that DNA methylation is being constantly maintained for these loci in an allele-specific manner . Our data are consistent with both the re-establishment of allelic methylation during development and the direct transmission of DNA methylation in the germline , and cannot distinguish between these modes of transmission . There may be alleles that influence the conversion of 5-methylcytosine to 5-hydroxymethylcytosine by the TET family of enzymes . Because the use of bisulfite sequencing does not distinguish between modifications , these alleles would not be detectable in our data . It will be interesting to determine if newly described genome scale methods [30]–[32] will be able to identify allelic differences in 5-methylcytosine and 5-hydroxymethylcytosine separately .
We extracted high molecular weight genomic DNA from 8 ml of blood from each individual . The buffy coats from each sample were isolated by centrifugation . Buffy coat was gently mixed and incubated for 30 minutes with lysis buffer at room temperature ( 0 . 32 M Sucrose , 10 mM Tris-HCl pH 7 . 5 , 5 mM MgCl2 and 1% Triton X-100 ) , then centrifuged at 4°C at 2500 rpm for 20 minutes . The supernatant was discarded , and the pellet was vortexed with 20 ml lysis buffer , then centrifuged at 4°C at 2 , 500 rpm for 20 minutes . The supernatant was discarded and the pellet was incubated in 5 ml guanidine isothiocyanate buffer on a shaker for 25 minutes ( 5 M Guanidine thiocyanate , 25 mM sodium acetate , 0 . 84% beta-mercaptoethanol ) . 5 ml 4°C isopropanol was added and the sample was inverted gently until precipitate appeared . Samples were then incubated at −20°C for at least one hour , and centrifuged at 4°C at 2 , 500 rpm for 20 minutes . The supernatant was then discarded and 500 µl T10E . 2 ( 10 mM Tris-HCl pH 7 . 4 , 2 mM EDTA pH 8 . 0 ) , 50 µl 3 M sodium acetate pH 5 . 2 , and 1 ml 100% ethanol were added and mixed and the DNA was precipitated by centrifuging at 12 , 000 rpm for 30 seconds . The supernatant was discarded and the pellet was washed with 500 µl 70% ethanol and centrifuged at 12 , 000 rpm for 30 seconds . The supernatant was discarded and the DNA was allowed to air-dry for 10 minutes . The genomic DNA pellet was resuspended in 1 ml T10E . 2 buffer by gentle vortexing . For GM12878 and H1 hESC , cells were grown according to ENCODE standards [33] . To generate the clonal cell lines , GM12878 was cultured according to the ENCODE protocol . At the time of culturing , the original GM12878 cell line was heavily skewed ( 92% ) towards the paternally inherited inactive X chromosome [34] . To obtain pure populations of cells with an inactivated maternal X chromosome ( Ximat ) or an inactivated paternal X chromosome ( Xipat ) , single cell clones of GM12878 were obtained by serial dilution . Each selected clone was tested for complete nonrandom inactivation by a PCR-based SNaPshot expression assay [22] using heterozygous SNPs in monoallelically expressed X-linked genes ( Table S2; K . S . K . and H . F . W . , manuscript in preparation ) . Four clones ( two with Ximat and two with Xipat ) were chosen for further study . For all cell lines , DNA was extracted from cell pellets using a DNeasy kit ( Qiagen ) according to the manufacturer's instructions . Kidney and skeletal muscle genomic DNA from the same donor was purchased from BioChain ( Hayward , CA ) . Genomic DNA was quantified using fluorescent DNA binding dye and a fluorometer ( Invitrogen Quant-iT dsDNA High Sensitivity Kit and Qubit Fluorometer ) . Reduced representation bisulfite sequencing ( RRBS ) was performed as described [35] . Briefly , 1 µg genomic DNA was digested with the methylation insensitive restriction enzyme MspI ( NEB ) . Ends of each restriction fragment were filled in and a 3′ adenosine was added with Klenow Fragment ( 3′→5′ exo-minus; NEB ) . Methylated paired-end Illumina adapters were ligated to the ends of the DNA fragments using T4 DNA Ligase ( NEB ) . Fragments between 105 bp and 185 bp were purified by agarose gel extraction . The purified fragments were treated with sodium bisulfite and then amplified by PCR with long-range PCR conditions and Platinum Taq Polymerase ( Invitrogen ) . The final PCR products were sequenced on Illumina GAIIx machines . All of the sequence data that is presented is of high quality with average quality scores of more than 25 for each cycle . Sequence data for the cell lines as well as the kidney and skeletal muscle tissue as available through the UCSC genome browser's [33] ENCODE DNA methylation track: HAIB Methyl RRBS . Sequence data for the primary blood leukocytes can be found under Gene Expression Omnibus ( GEO ) submission GSE30253 . All analysis was performed using the February 2009 ( GRCh37/hg19 ) build of the human genome . To determine instances of ASM , we first converted all cytosines in sequence reads to thymidines . We did the same to the reference genome sequence and then aligned the converted sequence reads to the converted reference sequence using Bowtie [36] to look for unique alignments . We required the read be uniquely aligned to only the best position in the converted reference , and that the reference position be the best and unique alignment to itself across the reference genome . We then identified the CpG positions in the unconverted reference sequence , and calculated the fraction of original sequence reads that have a C in that position . We performed further analysis only on CpGs with at least 14× read depth coverage . Bowtie also identifies mismatches between the reads and the reference , so we used it to look for common mismatches in the alignments to identify SNPs . We required at least 7 instances of the same mismatch and also required that the mismatch be found in at least 10% of the total reads that cover that position . We did not use an existing software tool to identify SNPs because the bisulfite reads do not align to a standard genome and all of the reads that cover a potential SNP start at one particular position due to the restriction digest . When analyzing the distribution of SNPs , we found that ASM SNPs that involve a reference G are the most prevalent , which is due to CpG-disrupting SNPs ( Table S5 ) . The next most prevalent SNPs involve a reference T , which is most likely due to extra Ts in the reference genome , since both Cs and Ts are represented as Ts in our bisulfite reference genome . For each called SNP , we determined whether the SNP was heterozygous by requiring that at least 7 sequence reads contain the reference allele . Then for each SNP-CpG pair that is present in the same 36 bp read , we calculated the amount of methylation on the variant allele and the reference allele . To test association of the SNP and the methylation status of the CpG , we performed a Fisher's Exact Test and calculated q-values [37] to assess false discovery rates for each SNP-CpG pair . To look at whether ASM SNPs were from dubious SNPs at the lower quality ends of sequence reads we determined the median distance between SNPs and CpGs for SNP-CpG pairs that exhibit ASM and those that do not . We found that the median distance was 8 bp in each class . For Sanger sequencing validation , bisulfite PCR , cloning and Sanger sequencing was performed as previously described [38] . For allele-specific methylation validation a Fisher's exact test was used to determine significance with a p-value cutoff of 0 . 05 . For allele-specific expression we compared the proportion of reads coming from each allele in the genomic DNA clones to the proportion of reads coming from each allele in the cDNA clones . A binomial test was used to determine significant allele-specific expression with a p-value cutoff of 0 . 05 . CpG island locations [39] , phyloP [40] scores ( Mammalian Cons – phyloP46wayPlacental ) and RefSeq [41] exon locations were downloaded from the UCSC genome browser [42] . Average linkage hierarchical clustering based on Euclidean distances was performed on the top 237 most varying CpGs as well as all CpGs that exhibited ASM across the 8 buffy coat samples using the software package Cluster 3 . 0 [43] . TreeView [44] was used to visualize the data . A linear model relating genotype and methylation levels was constructed in R . A CpG-SNP pair was analyzed only if the CpG had a standard deviation in percent methylation of at least 25 across family members and the SNP exhibited at least two different genotypes . Genotype values were assigned as follows: 0 = homozygous reference allele , 0 . 5 = heterozygous , and 1 = homozygous variant allele . We required at least 25 reads covering each SNP to assure that there were no errors in genotype calls . For each CpG-SNP pair , the percent methylation was regressed on the genotype value . ANOVA was then performed to calculate a p-value for each model . We then calculated q-values to assess false discovery rates and considered any model with a false discovery rate less than 5% to be significant . RNA-seq was performed as previously described [45] and the analysis is discussed in greater detail elsewhere ( T . E . R . , et al . manuscript in preparation ) . To discover ASE , we first created and aligned RNA-seq reads to RefSeq transcripts assembled from a modified genome sequence that contains both haplotypes of GM12878 . We obtained SNP and haplotype information from the 1000 Genomes Project [46] . After alignment , the number of reads aligned to each copy of each transcript was calculated . Then a binomial test was used to determine significance and false discovery rates [47] were calculated in R . Transcripts with a false discovery rate less than 5% were considered to harbor allele-specific expression . We used a cutoff of 25 reads , so that at least an 80% bias towards one allele could be detected as significant .
|
DNA methylation is a dynamic epigenetic mark that is essential for mammalian organismal development . DNA methylation levels can be influenced by environment , a chromosome's parental origin , and genome sequence . In this study , we evaluated the impact that DNA sequence has on DNA methylation by analyzing methylation levels in a three-generation family as well as unrelated individuals . By following DNA methylation patterns through the family along with nearby SNPs , we found that allelic differences between chromosomes play a much larger role in determining DNA methylation than the parental origin of the chromosome , indicating that DNA sequence has a larger impact on DNA methylation than gametic imprinting . We also found that allelic differences in DNA methylation found in the family can also be observed in unrelated individuals . In fact , the majority of variation in DNA methylation can be explained by genotype . Our results emphasize the importance of genome sequence in setting patterns of DNA methylation and indicate that genotype will need to be taken into account when assessing DNA methylation in the context of disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genomics",
"functional",
"genomics",
"x",
"chromosome",
"inactivation",
"genomic",
"imprinting",
"gene",
"expression",
"genetics",
"epigenetics",
"biology",
"dna",
"modification",
"genetics",
"and",
"genomics"
] |
2011
|
Analysis of DNA Methylation in a Three-Generation Family Reveals Widespread Genetic Influence on Epigenetic Regulation
|
Nonallelic homologous recombination ( NAHR ) between highly similar duplicated sequences generates chromosomal deletions , duplications and inversions , which can cause diverse genetic disorders . Little is known about interindividual variation in NAHR rates and the factors that influence this . We estimated the rate of deletion at the CMT1A-REP NAHR hotspot in sperm DNA from 34 male donors , including 16 monozygotic ( MZ ) co-twins ( 8 twin pairs ) aged 24 to 67 years old . The average NAHR rate was 3 . 5×10−5 with a seven-fold variation across individuals . Despite good statistical power to detect even a subtle correlation , we observed no relationship between age of unrelated individuals and the rate of NAHR in their sperm , likely reflecting the meiotic-specific origin of these events . We then estimated the heritability of deletion rate by calculating the intraclass correlation ( ICC ) within MZ co-twins , revealing a significant correlation between MZ co-twins ( ICC = 0 . 784 , p = 0 . 0039 ) , with MZ co-twins being significantly more correlated than unrelated pairs . We showed that this heritability cannot be explained by variation in PRDM9 , a known regulator of NAHR , or variation within the NAHR hotspot itself . We also did not detect any correlation between Body Mass Index ( BMI ) , smoking status or alcohol intake and rate of NAHR . Our results suggest that other , as yet unidentified , genetic or environmental factors play a significant role in the regulation of NAHR and are responsible for the extensive variation in the population for the probability of fathering a child with a genomic disorder resulting from a pathogenic deletion .
Homologous recombination ( HR ) , leading to crossing over and exchange between homologous DNA sequences , occurs during meiosis and ensures that each gamete contains a unique mixture of maternal and paternal DNA . Occasionally HR occurs ectopically between highly similar duplicated sequences or paralogous genomic segments , such as segmental duplications , in a process known as non-allelic homologous recombination ( NAHR ) . NAHR between directly oriented duplicated sequences on the same chromosome gives rise to a chromosomal deletion , and , if it occurs in a inter-molecular fashion , can generate a reciprocal duplication on the other chromosome , whereas NAHR between duplicated sequences in an inverted orientation leads to inversions . The breakpoints of NAHR rearrangements cluster in defined hot spots within segmental duplications that reflect hotspots of HR activity [1] . Genomic rearrangements resulting from NAHR can be manifested as genomic disorders , predominantly due to the altered copy number of dosage-sensitive genes [2] , or non-pathogenic structural variation [3] . Several NAHR hotspots have been identified due to their association with specific genomic disorders , with reciprocal deletion and duplication events being associated with different disorders at some loci [4]– . For example NAHR between two copies of the CMT1A-REP segmental duplication on 17p12 leads to deletion of a 1 . 4 Mb region including the PMP22 gene resulting in hereditary neuropathy with liability to pressure palsies ( HNPP ) , with reciprocal duplication of the same region resulting in Charcot-Marie-Tooth disease type 1A ( CMT1A ) [8] . Until recently the rates of rearrangement at any particular locus were estimated from the frequency of the resultant dominant disease phenotype in the population . It is now possible to estimate the frequency of recombination in males through direct analysis of sperm by PCR amplification of breakpoint products [9] . Direct analysis of rates of rearrangement in germline DNA , at four NAHR hotspots , revealed variation in rate both between individuals and loci [9] , however very little is known about the extent of the interindividual variation or the factors which influence this . Both genetic and non-genetic factors have been suggested to play a role in influencing mutation rates . It has previously been shown that variation in cis can influence the rate of chromosomal translocations [10] , [11] . Several properties of duplicated sequences have been shown to be major determinants of the rate of nonallelic homologous recombination , with rate increasing with length and sequence similarity and decreasing with distance between repeats [12] , [13] . Recently PRDM9 has been identified as a genome wide trans regulator of meiotic recombination in humans and mice [14]–[18] , and variation within this gene has been shown to significantly alter rates of meiotic recombination and instability , including CMT1A-REP rearrangements [15] . Recombination rate is also correlated with density of the recombination hotspot motif [12] , [17] to which PRDM9 binds [14] . Evolutionary and epidemiological studies have suggested that mutation rates are higher in the paternal germline and increase with paternal , but not maternal , age . Thus age is an important potential confounder to consider in any investigation of genetic and environmental influences on mutation rate . These studies are consistent with the observation that the male germline entails greater numbers of mitotic replications than the female germline , and that the number of paternal mitotic replications increases with age , whereas the number of maternal mitotic replications does not [19] . However , most studies on this topic have focused on base substitutions and not structural variants . Two previous studies could not detect an age dependent effect for de novo deletions and duplications flanked by duplicated sequences [20] , [21] , although they observed conflicting results for events not flanked by duplications [20] , [21] . Moreover , there was little evidence of a genome-wide parent of origin bias for NAHR deletions and duplications from either study [20] , [21] . However , a significant paternal origin bias has been observed at the CMT1A-REPs specifically [22] . The presence or absence of a paternal age effect for a given mutational process may depend on whether the underlying mutation mechanism is dependent on , or potentiated by , mitotic replication . It has previously been shown that NAHR at four different NAHR hotspots , including the CMT1A-REPs , is specific to meiosis [9] . To further investigate the role of age , genetic variation and environmental factors on interindividual variation in rates of NAHR we directly estimated the rate of deletion at the CMT1A-REPs in sperm samples from 34 UK males from an extensively phenotyped twins cohort , for which we had determined their PRDM9 genotype . These samples included 8 pairs of MZ twins , which allowed us to demonstrate a significant role for shared genetic variation and/or shared environment in determining rate of NAHR in sperm .
We directly estimated the rate of NAHR generated deletions at the CMT1A-REP hotspot in sperm DNA from 34 UK males , including 16 monozygotic ( MZ ) co-twins ( 8 twin pairs ) , six dizygotic ( DZ ) co-twins ( 3 twin pairs ) and 12 unrelated individuals , using the assay described previously [9] ( results shown in Table S1 ) . The CMT1A-REP deletion was chosen for analysis in this study as it has the highest rate of NAHR among loci that can be robustly assayed , and therefore the lowest measurement error , as well as the ease of resequencing the NAHR hotspot in different individuals . The average rate of deletion in our cohort of 3 . 54×10−5 ( +/−3 . 04×10−6 s . e . m ) , with a range of 9 . 82×10−6 to 6 . 96×10−5 , is consistent with the rate ( 4 . 20×10−5 ) reported previously [9] . We analysed deletion rate as a function of the age of the individual at the time the sperm sample was produced ( 24 to 67 years ) ( Figure 1 ) . No age dependent increase in deletion rate was observed among males , with linear regression analysis of deletion rate on donor age showing no correlation ( R2 = 5×10−4 , p = 0 . 9019 ) . It has been suggested that the paternal mutation rate increases with age due to an increase in the number of germ-cell divisions , and therefore genome replications , to complete spermatogenesis . The number of genome replications occurring during spermatogenesis in a man of a certain age can be estimated as follows: 30 prior to puberty , 23 spermatogonial stem cells replications per year after puberty , followed by 5 during sperm maturation [19] . Assuming puberty occurs at age 15 a sperm produced by a man of age 24 will have undergone 242 genome replications , while a sperm produced by a man of age 67 will have undergone 1231 . This equates to a 5 . 1 fold linear increase in the number of genome replications between ages 24 and 67 ( the ages of the youngest and oldest donors in our study ) . To test whether we have the statistical power to detect such a fold change with our data we performed simulations to assess the power of our study design to detect a paternal age effect should it exist ( as described in Material and Methods ) . Our simulations showed that we had 100% power to detect a fold change of 5 . 1 between our youngest and oldest donors , and had 95% power to detect a much smaller fold linear change of 1 . 7 ( Figure S1 ) . We also used simulations to exclude a non-linear paternal age affect similar to that observed between maternal age and prevalence of Down syndrome [23] . Thus we can confidently exclude that even a weak linear or non-linear paternal age effect exists for NAHR at this locus . The cohort of 34 males studied includes 16 MZ co-twins ( 8 twin-pairs ) . To estimate the heritability of deletion rate we calculated the intraclass correlation ( ICC ) within MZ co-twins and unrelated MZ pairs . The estimate of deletion rate for each MZ co-twin plotted against one another is shown in Figure 2 . The analysis reveals a significant correlation between MZ co-twins ( ICC = 0 . 784 , 95% CI 0 . 292–0 . 952 , p = 0 . 0040 ) . In order to test whether the MZ twins are more highly correlated than unrelated pairs we randomly sampled 8 pairs of unrelated individuals from the MZ co-twins , calculated the ICC and p-value for the generated sample set , and repeated this 10 , 000 times . Out of the 10 , 000 simulated unrelated sample sets only 64 ( 0 . 64% ) had an ICC greater than that observed in the MZ co-twins and a p-value less than that observed in co-twins . The rate of NAHR in MZ co-twins is therefore significantly more similar than in unrelated pairs of individuals . The observed heritability is due to the rate of deletion being determined by elements shared by MZ co-twins , namely genetics or shared environment . Comparing the level of heritability observed between MZ and DZ co-twins is often used to distinguish between the relative effects of these elements on inherited traits . MZ co-twins share 100% of their genetic variation , compared to 50% on average shared between DZ co-twins , while MZ and DZ twins are expected to have a similar level of shared environment . Due to the limited number of DZ co-twins available in this study ( 6 co-twins , 3 twin-pairs ) we are unable to use the direct comparison between MZ and DZ co-twins to distinguish between the relative effects of genetics and shared environment on deletion rate , despite repeat attempts to sample sperm from additional DZ co-twins . However in the vast majority of traits explored in this twin cohort any estimated effect of shared environment is usually much less than 5% of the variance . In an attempt to identify specific genetic or environmental factors involved in the regulation of deletion rate we investigated the effect of possible determinants of NAHR . Variation in the gene PRDM9 is a known genetic determinant of NAHR rate , with variation at the PRDM9 zinc finger having been shown to regulate recombination including NAHR deletion rate at CMT1A [15] . We cloned and sequenced the PRDM9 zinc finger alleles from each individual , and aligned these to published sequences [14] , [15] to classify the alleles as described previously . The majority of individuals in our cohort were genotyped as A/A homozygote ( 29/34 , 85% ) , with 4 A/B heterozygotes and one A/L20 heterozygote . A/A homozygotes exhibited high levels of variability in deletion rate , with a seven-fold range of 9 . 82×10−6 to 6 . 96×10−5 . The high level of variability observed within A/A homozygotes means that we can exclude variation at PRDM9 as being a major cause of the correlation observed between MZ twins . We classified the different minor PRDM9 alleles that we observed into those predicted to bind to the canonical recombination hotspot motif , and those predicted a bind a non-canonical motif . Only the L20 allele present in a single male is predicted to bind a non-canonical motif . The deletion rate in the A/L20 sample falls within the range observed in the samples homozygous for alleles recognising the canonical motif , although it is towards the lower end . We also found no significant differences in NAHR rate between the males homozygous and heterozygous for the A allele ( p = 0 . 1104 , Mann-Whitney test ) . The deletion rate for each twin arranged by PRDM9 motif binding classification is shown in Figure 3 . We also tested for association of the rate of deletion at the CMT1A-REPs with 4 SNPs at 3 loci ( rs1670533 , rs3796619 , rs17542943 , rs7863596 ) previously shown in genome-wide association studies to have modest effects on rates of allelic recombination [24] , [25] , in 13 unrelated semen donors for whom data were available from genome-wide SNP genotyping chips . Unsurprisingly , given the small sample size and known modest effect sizes we observed no significant ( p<0 . 05 ) association after Bonferoni correction for multiple testing . Sequence similarity between duplicated sequences has been proposed to be one of the primary determinants of NAHR rate , along with the length of the duplicated sequences , the distance between the duplicated sequences and the density of HR hotspot motifs [26] , [27] . We sequenced the proximal and distal CMT1A-REP NAHR hotspots in one of each of the eight MZ co-twins . We tested for association between the average NAHR rate in these 8 co-twins and sequence similarity between proximal and distal hotspots but we observed no significant association ( p = 0 . 27 , linear regression ) , suggesting that variation in local sequence similarity is not a primary determinant of NAHR at this hotspot . In addition , we did not observe any size variation in the PCR amplifications of the proximal and distal segmental duplications , suggesting that in the males studied here there are no sizeable structural variants within the CMT1A-REP but outside the NAHR hotspot . The heritability of CMT1A deletion rate could be due to elements in the shared environments of MZ co-twins that effect recombination rate . Studies have shown that several environmental mutagens affect mutation rate and could have direct effects on NAHR rate . Analysis of human populations exposed to radiation has revealed that germline mutation rates are increased by ionizing radiation [28] , while studies in animals have revealed that several environmental pollutants increase mutation rate in the male germline [29]–[32] . Several lifestyle choices have also been identified as having potential mutagenic effects through their association with cancer [33] , while sperm of smokers have been shown to have higher frequencies of DNA damage [34] . We were able to obtain information on BMI , smoking status and alcohol intake for a subset of our twin samples . We did not detect any correlation between deletion rate and BMI ( R2 = 0 . 06 , p = 0 . 2357 , linear model ) in an analysis of 27 twin samples . Reliable information on alcohol intake and smoking status at the time of sampling was only available for ten samples . No correlation was detected between deletion rate and alcohol intake ( R2 = 0 . 03 , p = 0 . 6385 , linear model ) or smoking status , irrespective of whether ex-smokers are classified as smokers ( p = 0 . 9143 , Mann-Whitney ) or non-smokers ( p = 0 . 7111 , Mann-Whitney ) . However due to the small sample size for which smoking status and alcohol intake data was available we are unlikely to have the power to detect any subtle correlations between these risk factors and deletion rate .
In this study we have demonstrated that there is no appreciable paternal age effect on NAHR recombination rate at the CMT1A-REP locus . This raises the question of why a paternal age effect is observed for some mutation processes and not others . The absence of an age effect in this process is likely due to its meiotic-specific nature . Paternal age has been most strongly linked to the rate of base substitution [35] , which may occur due to errors in DNA replication . As we have described the number of DNA replications to complete spermatogenesis increases with age . Previous studies detected no paternal age effect on translocation [36] or allelic recombination [37]–[39] rates , which are both meiotic in nature . The number of meiotic divisions involved in spermatogenesis is fixed at two and does not increase with age . These observations suggest that germline mutation events that are observed to exhibit an increase in frequency with paternal age are predominantly mitotic in nature , while those that fail to exhibit such an effect are meiotic . These observations are supported by the observation that patients with pathogenic de novo deletions and duplications mediated by NAHR do not have significantly older parents than matched controls [20] . It is also worth noting that paternal age effects could also be affected by the positive selection and clonal expansion of cells in the testes , but this is typically linked to a very small number of specific activating mutations [40] . We have detected a significant correlation in NAHR deletion rate at the CMT1A locus between MZ co-twins , with MZ co-twins being significantly more correlated than unrelated MZ pairs . MZ co-twins are essentially genetically identical , are the same age and share many aspects of their environment . Any of these shared effects could be determinants of heritable traits . In our analysis we have excluded age , variation at the PRDM9 zinc finger and variation in sequence similarity between paralogous hotspot sequences as being the cause of the observed heritability of NAHR rate . Therefore we conclude that the observed heritability is due to the effects of genetics , shared environment or a combination of these factors . The mutation process we are analysing occurs during meiosis in the male germline . As the only two meiotic divisions in spermatogenesis occur immediately before sperm formation the deletion events we are analysing will have occurred only a few weeks prior to the sample being provided . It therefore seems likely that among different types of shared environmental factors , those that relate to lifestyle choices or habits in adulthood may play more of a role than shared childhood environment . Although it is also possible that shared environment in childhood could induce epigenetic effects that extend into adulthood . We did not detect any correlation between BMI , smoking status or alcohol intake with deletion rate in analysis of a subset of our twin samples , however for smoking status and alcohol intake the small number of these samples mean we only had the power to detect a substantial effect . Further investigation is required into the possible environmental determinants of deletion rate . Due to the previously reported links between smoking and genetic aberrations in sperm [34] we see particular value in extending this study to investigate the effects of smoking on NAHR in a larger number of samples . Our results suggest that a large fraction of the heritability in deletion rate between MZ co-twins may be due to shared genetic variation . Although PRDM9 has been identified as a significant regulator of NAHR , variation at the PRDM9 zinc finger cannot explain the heritability of recombination rate observed in this study . There are likely to be additional genetic factors , with either genome-wide or locus-specific effects , which determine the rate of deletion at the CMT1A locus . Several genome-wide association studies have set out to identify genetic determinants of allelic recombination through pedigree analysis of crossovers [24] , [25] , [41] . Several loci have been identified that are associated with allelic recombination in males including a two single-nucleotide haplotype in the RNF212 gene [24] , [25] , [41] , in addition to SNPs at 7q36 . 1 ( nearest gene NUB1 ) and 9q31 . 3 ( nearest gene UGCG ) [24] . Rnf212 , Ugcg and Nub1 expression is induced at meiosis with a peak at diplotene stage of prophase 1 , which corresponds to chromosome chiasmata resolution and supports their potential roles during meiosis [24] . It should be noted that different loci are associated with recombination rate in males and females; the RNF212 haplotype that is associated with high recombination rate in males is associated with low recombination rate in females [24] , [41] . It is possible that these genome wide determinants of allelic recombination also affect NAHR rate , although their modest effect sizes on allelic homologous recombination means that they are highly unlikely to explain the seven-fold range in deletion rates observed in our study . It is possible that locus-specific regulators of NAHR rate exist . This hypothesis is supported by observations that polymorphisms within the regions flanking the segmental duplications involved in Williams-Beuren Syndrome increase the rate of NAHR [10] . Locus specific regulators of mutation rate have also been identified at the NID1 meiotic recombination hotspot [42] , minisatellite MS32 [43] and within the palindromic repeats involved in t ( 11;22 ) translocation [11] , [44] . Locus specific regulators of NAHR rate at CMT1A-REPs may exist , however none have been identified to date . Repeat length , sequence similarity and distance between repeats have also been shown to influence NAHR [12] , [13] , along with concentration of a hotspot motif [12] , [17] . We did not detect any correlation between sequence similarity at the proximal and distal CMT1A-REP NAHR hotspots and NAHR rate , and we did not observe any variation in repeat length . It seems likely that variation in distance between repeats and concentration of the hotspot motif is more likely to effect variation between loci than between individuals at this particular locus . We expect that the phenomena of NAHR rate heritability and absence of paternal age effect that we observed in this study are generalisable to other loci subject to meiotic-specific NAHR . The CMT1A-REP NAHR hotspot is not an outlier with respect to the amount of variation in NAHR rates between individuals compared to other NAHR hotspots [9] . Indeed , if the primary determinants of this variation in rate are shared environmental factors or trans-acting genetic variation we may expect rates of NAHR to vary in a correlated fashion across many NAHR hotspots . Whereas if the primary determinants of rare variation are locus-specific in action ( e . g . cis-acting genetic variation ) rates of NAHR across loci in different individuals would be uncorrelated . We would also expect the same factors to influence rates of NAHR-generated duplications , as duplications and deletions are reciprocal products of the same event . These findings necessitate further investigation into the genetic control of NAHR rate . An important next step would be to determine whether genetic control is genome-wide or locus-specific through the analysis of multiple hotspots within each individual . Such an analysis would be limited however by the number of sperm available from each individual and the labour intensive nature of this technique . Given the pathogenic nature of the deletion we studied , these observations have striking clinical implications . They suggest that either there are potentially modifiable environmental factors that alter the probability of having a child with a genomic disorder , and/or different males vary markedly from the population average in their risk of having a child with a genomic disorder by virtue of variation in their genome . The population impact of these mutagenic factors will depend on whether they influence NAHR rate variation locus-specifically or genome-wide , and , with approximately half of pathogenic NAHR events coming from the maternal germline , whether they are specific to the male germline . While we estimated the average rate of NAHR-mediated deletion at the CMT1A-REP to be 1 in ∼30 , 000 sperm , the cumulative impact of pathogenic NAHR events genome-wide , and including both paternal and maternal germlines , is likely to be observed in more than 1 in 3 , 000 births [45] .
All samples and information were collected with written and signed informed consent . The study was approved by the St Thomas' Hospital Research Ethics Committee . Semen samples were obtained from 34 volunteers included in the TwinsUK adult twin registry , based at St Thomas' Hospital , King's College , London ( www . twinsuk . ac . uk ) . The ages of these individuals at sampling ranged from 24 to 67 years . DNA samples were randomized and relabeled to enable estimation of deletion rate blinded to sample relatedness . DNA was extracted from semen samples using a protocol adapted from the QIAamp Tissue Protocol using the QIAamp DNA Blood Maxi Kit ( Qiagen ) . 1 ml of semen was transferred to a 50 ml Falcon tube and 20 ml Buffer 1 ( 150 mM NaCl , 10 mM EDTA ( pH 8 . 0 ) ) added , before vortexing for 10 seconds and centrifuging at 4000 rpm for 10 minutes . The supernatant was discarded into Virkon disinfectant ( Day Impex Ltd ) and the pellet resuspended in 3 ml Buffer 2 ( 100 mM Tris·Cl ( pH 8 . 0 ) , 10 mM EDTA , 500 mM NaCl , 1% SDS , 2% β-mercaptoethanol ) . 400 µl Proteinase K was added to the solution and incubated at 55°C with gentle rocking or occasional inversion . After 2 hours an additional 100 µl Proteinase K was added and incubated for a further 2 hours at 55°C as before . 6 ml Buffer AL ( Qiagen ) was added to the solution and incubated for 10 minutes at 70°C , before adding 5 ml of Ethanol , followed by mixing by inverting the tube 10 times , then shaking . The solution was transferred onto the QIAamp maxi column ( Qiagen ) placed in a 50 ml centrifuge tube , taking care not to moisten the rim , the cap closed and centrifuged at 3000 rpm for 3 minutes . The filtrate was discarded and the QIAamp maxi column placed back in the 50 ml centrifuge tube . 5 ml of buffer AW1 was added to the QIAamp Maxi column , the cap closed , and centrifuged for 3350 g for 1 minute . 5 ml of buffer AW2 was added to the QIAamp Maxi column , the cap closed , and centrifuged at 3350 g for 15 minutes . The filtrate was discarded and the QIAamp column placed in a clean 50 ml centrifuge tube . 550 µl distilled water , equilibrated to room temperature , was pipetted onto the membrane of the QIAamp Maxi column , the cap closed and incubated at room temperature for 5 minutes before centrifuging at 3350 g for 2 minutes . This elution step was repeated to give a total elution volume of approximately 1 ml . The sequence encoding the PRDM9 zinc finger was amplified in 25 µl reactions using primers PRDM9_F3 and PRDM9_R1 [14] with final concentrations of 0 . 5 mM for each primer , 4 . 5 mM MgCl2 , 0 . 05 U/ml of Taq polymerase/Pfu polymerase mix ( 10 units Taq:1 unit Pfu ) , and 1× PCR buffer system as described in [46] to amplify 25 ng of input DNA . Thermal cycling conditions were: 96°C for 20 seconds for one cycle , followed by 96°C for 10 seconds , 60°C for 20 seconds and 68°C for 2 min , for 30 cycles . Following gel electrophoresis bands were excised , DNA extracted using the QIAquick Gel Extraction kit ( Qiagen ) and cloned using the TOPO TA Cloning Kit for Sequencing ( Invitrogen ) . Twelve colonies were picked from each transformation , cultured overnight and sequenced with primers 214F ( TGATTGTTTCTTCATTTGATCTTCA ) , 731F ( TGGAGAGTGTGGACAAGGTTT ) , 1742R ( AGCAGAGGCTTGACCTATCG ) and 1992R ( GTCATGAAAGTGGCGGATTT ) using 4∶1 Big Dye Terminator:dGTP Chemistry ( Applied BioSystems ) . PCR efficiency was estimated for each sample by carrying out PCR reactions with approximately single molecule inputs of DNA . Each DNA sample was diluted to 20 pg/µl with 1 ng/µl herring sperm DNA . The CMT1A proximal repeat was amplified using primers CMT1A_PF1 and CMT1A_PR1 in the primary PCR and primers CMT1A_PF2 and CMT1A_PR2 in the secondary PCR [9] . PCR reactions were carried out as described previously to amplify CMT1A-REP deletion products with an input of 5 pg sample per reaction . A subset of positive amplification products were confirmed by reamplifying a subset of plates and showing 100% concordance with wells containing positive products , along with sequencing products from a subset of positive wells to confirm that all were consistent with the CMT1A proximal segmental duplication ( data not shown ) . Poisson analysis was used to calculate the number of amplifiable molecules in each reaction using the equation –N ln[ ( N – R ) /N] , where N is the number of reactions performed and R is the number of positive reactions observed , and the mass of one haploid genome calculated for each sample . Amplification of CMT1A-REP deletion products was carried out as described in [9] , but with 2 µl of diluted PCR product used as a template in the secondary PCR . Poisson correction of positive results and calculation of confidence limits was also carried out as described previously [9] . The number of input molecules in each well was estimated from the DNA input and the mass of one amplifiable haploid molecule , estimated by limiting dilution PCR for each sample . A subset of positive deletion products were confirmed by reamplifying a subset of plates and showing 100% concordance with wells containing positive products , along with sequencing products from a subset of positive wells ( data not shown ) . Primers ( CCATGATCACCCTCATGTCA and CATGCAAACGAAAATGAAGC ) were designed to amplify the hotspot for NAHR across the proximal and distal repeats . PCR was carried out using REdAccuTaq LA DNA polymerase from Sigma Aldritch under the following conditions in a 50 µl volume , with approximately 25–50 ng of template DNA: 96° 30 secs , 94° 15 secs , 60° 30 secs −1° per cycle , 68° 3 min , go to step 2 6 times , 94° 15 secs , 57° 30 secs , 68° 3 min , go to step 6 29 times , 68° 10 mins , 4° hold . After amplification , PCR products were ethanol precipitated and resuspended in 5–10 µl of double distilled water . Ligation and transformation into Pgem-TEasy ( Promega ) was carried out according to the manufacturers instructions . 48 clones were prepped for each individual , and 24 were sequenced . Sequencing was carried out using standard dye-terminators and the external PCR primers . Internal sequencing primers were designed on each strand at +500 bp , +1000 bp and +1500 bp , respectively , so each clone was sequenced with a total of 8 primers . Sequence analysis was carried out in GAP4[47] . Clones were assembled with the proximal and distal reference sequences and trimmed to the length of the PCR product . Clones were separated into proximal or distal specific contigs based on a stable 5 bp insertion in the distal repeat . Each sequence was manually inspected and base calling errors caused by dye-blobs and PCR errors were edited . Finally , the consensus sequence for the proximal and distal repeats was exported and aligned in SeaView4[48] using ClustalW2 . Intraclass correlation was calculated in R [49] using the irr package ( http://CRAN . R-project . org/package=irr ) and 95% confidence intervals on the Poisson-corrected number of counts were calculated in R using the epitools package ( http://CRAN . R-project . org/package=epitools ) . SNP association analyses were performed including only one sample per family by linear regression using deletion rate as a quantitative trait against individual SNP genotypes . In addition , we also performed a linear regression haplotye-based association analysis . Both analyses were performed using PLINK [50] . We simulated a linear fold change in NAHR rate between ages 24 and 67 by setting the mean deletion rate in the simulated data to the observed mean deletion rate and estimating the expected rate for each sample , using their actual age , by assuming a linear increase from 24 to 67 years old , for different values of fold change from 1 . 1x to 10x . The measured rate of NAHR for each sample was simulated by random sampling from the Poisson distribution around the expected rate , and the correlation between age and simulated rates of NAHR was tested using linear regression , with a p-value threshold of 0 . 05 . The proportion of replicates for which a significant result was achieved was equated to the statistical power of the experiment for that fold-change . The observed increase in the prevalence of Down's syndrome with maternal age has been proposed to follow a logit logistic model [23] . We tested the power of our data to detect such a relationship between deletion rate and age by simulating a logit logistic curve from our data . The midpoint of the simulated deletion rates was set to the observed mean deletion rate in our data and a deletion rate for each sample generated by sampling from the Poisson distribution . A p-value for the correlation between the ages and simulated rates was calculated using Spearman's rank correlation test .
|
Many genetic disorders are caused by deletions of specific regions of DNA in sperm or egg cells that go on to produce a child . This can occur through ectopic homologous recombination between highly similar segments of DNA at different positions within the genome . Little is known about the differences in rates of deletion between individuals or the factors that influence this . We analysed the rate of deletion at one such section of DNA in sperm DNA from 34 male donors , including 16 monozygotic co-twins . We observed a seven-fold variation in deletion rate across individuals . Deletion rate is significantly correlated between monozygote co-twins , indicating that deletion rate is heritable . This heritability cannot be explained by age , any known genetic regulator of deletion rate , Body Mass Index , smoking status or alcohol intake . Our results suggest that other , as yet unidentified , genetic or environmental factors play a significant role in the regulation of deletion . These factors are responsible for the extensive variation in the population for the probability of fathering a child with a genomic disorder resulting from a pathogenic deletion .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetic",
"mutation",
"genetics",
"biology",
"human",
"genetics",
"genetics",
"of",
"disease"
] |
2014
|
The Rate of Nonallelic Homologous Recombination in Males Is Highly Variable, Correlated between Monozygotic Twins and Independent of Age
|
Interactions between individuals and the structure of their environment play a crucial role in shaping self-organized collective behaviors . Recent studies have shown that ants crossing asymmetrical bifurcations in a network of galleries tend to follow the branch that deviates the least from their incoming direction . At the collective level , the combination of this tendency and the pheromone-based recruitment results in a greater likelihood of selecting the shortest path between the colony's nest and a food source in a network containing asymmetrical bifurcations . It was not clear however what the origin of this behavioral bias is . Here we propose that it results from a simple interaction between the behavior of the ants and the geometry of the network , and that it does not require the ability to measure the angle of the bifurcation . We tested this hypothesis using groups of ant-like robots whose perceptual and cognitive abilities can be fully specified . We programmed them only to lay down and follow light trails , avoid obstacles and move according to a correlated random walk , but not to use more sophisticated orientation methods . We recorded the behavior of the robots in networks of galleries presenting either only symmetrical bifurcations or a combination of symmetrical and asymmetrical bifurcations . Individual robots displayed the same pattern of branch choice as individual ants when crossing a bifurcation , suggesting that ants do not actually measure the geometry of the bifurcations when travelling along a pheromone trail . Finally at the collective level , the group of robots was more likely to select one of the possible shorter paths between two designated areas when moving in an asymmetrical network , as observed in ants . This study reveals the importance of the shape of trail networks for foraging in ants and emphasizes the underestimated role of the geometrical properties of transportation networks in general .
Various ant species build networks of trails that link together nesting sites and exploited resources [1] . These networks are generally formed by one or several dendritic trees originating from the nest of the colony . They can stretch over large distances and display very intricate patterns . For instance , the harvester ant Messor barbarus forms trails that persist over several consecutive days and can extend up to 25 meters from the nest entrance [2] . The wood ant Formica aquilonia , whose body length is just 5–6 millimeters , can form networks where trails reach 200 meters in length , with up to nine successive branching points per trail [3] . As a last example , the trail system in a colony of leafcutter ants Atta colombica can cover an area larger than 1 hectare , with trails extending up to 250 meters from the nest [4] . One major challenge for ant workers is to orient themselves inside such labyrinths and in particular to keep track of the direction of their nest . To do so , they use at least four different , but non-exclusive , types of information . First , they can rely on visual information . Some species use forest canopy [5] or sun position [6] to estimate the direction toward their nest . Others memorize environmental landmarks along their path [7] . Second , they can also use proprioceptive information . Certain ant species approximate the direction toward their nest by summing their successive vectors of movements , measured as step numbers and body rotations [8]–[10] . Third , they can exploit social information , such as the food load of encountered workers . In ants carrying their food ( such as seeds or leaf fragments ) on surface trails , the proportion of laden ants is higher in the returning flow . Some ants use this difference to correctly reorient themselves on a trail [11] . The last type of information that ants can use to find the direction of their nest lies in the structure of the trail network itself . In several ant species , these networks display a particular property: the mean angle between trails as they branch out symmetrically from the nest lies around 60° , in the range 50°–100° depending on the species ( Leptogenys processionalis [12]; Atta sexdens , A . capiguara , A . laevigata and Messor Barbarus [13]; Monomorium pharaonis [14]; Formica aquilonia [3]; Linepithema humile , unpublished data ) . Therefore , an ant exiting the nest and moving toward the periphery of the network generally faces symmetrical bifurcations , i . e . the two trails that follow a bifurcation deviate by approximately 30° from the original direction of the ant . Conversely , an ant coming back to its nest faces asymmetrical bifurcations: the trail heading toward the nest after a bifurcation deviates less ( ∼30° ) from the ant's original direction than the other trail ( ∼120° ) which leads away from the nest . In this last situation , and in absence of any other information , ants preferentially follow the least deviating trail , as demonstrated in our recent study with the Argentine ant L . humile [15] , [16] . We also showed that this behavioral bias , associated with the pheromone recruitment of this ant species , led to a significant improvement of the colony's ability to select the shortest route between its nest and a newly discovered food source [16] and depends critically on the branching angle [17] . A question that remains to be elucidated is whether ants reaching a bifurcation actually use its geometry as an orientation cue to decide which trail to follow next , or whether their individual and collective behaviors are in fact the product of a passive interaction with the geometrical structure of the trail network . The answer to this question depends on , for the moment , rare behavioral observations whose conclusions differ according to the experimental procedure and species studied [14] , [15] . In order to gain new insight into the role of the trail geometry , we studied the behavior of robotics models of ants , whose perception abilities are known and whose behaviors can be specified . During the last fifteen years , the use of robots to investigate animal behavior has been increasingly popular ( see [18]–[20] for a review and examples ) and has led to the development of innovative control algorithms [21] , [22] . Several attempts have been made to produce ant-like robots that are able to lay and follow pheromone-like trails using heat trails [23] , chemical trails [24] , glow paint trails [25] , virtual trails [26] or light trails [27] , [28] . Such trail systems are a promising way of guiding and organizing the activities of robotics swarms in space , particularly in unknown environments . From a biological point of view , these robotic models also offer the possibility of investigating questions related to the influence of the perceptual/cognitive abilities of individual ants on the collective behavior of the colony . Here we present the results of an experiment where a group of ant-like robots had to establish a route between a starting area and a target area in a network of corridors , mimicking the experiments we performed with ants in our previous studies [15] , [16] . For technical convenience pheromone trails were replaced by light trails projected along the paths followed by the robots by a video projector ( as proposed in [27] , [29] and implemented in [28] ) . Robots can detect and follow these light trails thanks to two photoreceptors that mimic the antennae of the ants . The robots were tested in two types of networks , one type made only of symmetrical bifurcations and the other type containing asymmetrical bifurcations , as in natural ant networks . Their behavior was kept as minimal as possible to observe just the interaction between the displacement of the ants , their trail laying/following behavior and the structure of the environment . In particular and in contrast to previous simulation work [16] , [17] , they were not given the capability to measure the angle between the corridors when reaching a bifurcation and therefore they could not make a change of direction based on this information . A comparison between the behavior of the robots and the behavior of ants in our previous experiments demonstrates that simple individual behavioral rules are sufficient to explain the efficient pattern of network exploitation observed in ants . It also helps us to better understand how the physical structure of the environment can affect individual and collective activities in social insects .
The experimental setup was a scaled-up , simplified version of the setup used in [16] to study the behavior of Argentine ants . The behavior of the ant workers was tested in a maze of corridors carved in a PVC ( polyvinyl chloride ) board ( 5 mm wide , about 4–5 times the width of an ant ) . These corridors mimicked permanent trails that are found in ant species that remove vegetation and debris to form physical routes toward long-lasting food sources [1] , [30] . The experimental setup used with the robots was a network of corridors ( 9 cm width , 4 . 5 times the width of a robot ) built with white cardboard ( 5 mm thick , wall height of 2 . 5 cm ) . In ants , the network was made of four interconnected diamond-shaped loops connecting a starting area ( corresponding to the nest of ants ) on one end and a target area ( corresponding to a food source for instance ) on the other end . In robots , the network was made of only three interconnected diamond-shaped loops ( see Fig . 1 ) in order to keep its dimensions within the space allowed by the pheromone deposit device ( 140×105 cm ) while scaling up the length of the diamond-shaped loops by four ( robots move 4 times faster than the ants ) . The starting and target areas were hexagons of the same dimensions ( 22 . 5 cm diameter ) . In this network there were 7 ( vs 14 for the ants ) possible paths of different lengths ( shorter path: 86 cm; longer path: 178 cm ) that robots could use to go from the starting area to the target area , without using the same segment of the network twice ( a corridor between two bifurcations ) . Two network configurations were used . In configuration S ( for “symmetrical” ) , each diamond-shaped loop of the network was perfectly symmetrical . As a consequence , all bifurcations of the network were also symmetrical: whatever incoming branch was at a bifurcation , the two other branches deviated by an angle of 60° on the left or on the right . In configuration A ( for “asymmetrical” ) , each diamond-shaped loop of the network was flattened along one of its axes ( the same for each loop ) . As a consequence , the network bifurcations were not always symmetrical anymore: depending on the incoming branch at a bifurcation , the two other branches both deviated by an angle of 30° on the left or on the right ( symmetrical side of the bifurcation ) , or one branch deviated by an angle of 30° in one direction while the other branch deviated by an angle of 120° in the other direction ( asymmetrical side of the bifurcation ) . Except for this difference in the geometry of the bifurcations , configurations S and A were identical: they presented the same topology , had segments of the same length and had the same total length . 15 experimental replicates with 10 robots were performed with each network configuration . Each experimental replicate lasted 60 minutes . The micro-robots Alice ( see Fig . 2 ) were designed at the EPFL ( Lausanne , Switzerland [31] ) . They were very small robots ( 22 mm×21 mm×20 mm ) equipped with two watch motors with wheels and tires , with a maximum speed of 40 mm s−1 . Four infrared ( IR ) sensors and transmitters were used for detection of the starting and target areas , and for obstacle detection . The front left and front right sensors were oriented 45° toward the left and the right of the robots' moving direction respectively; the front and back sensors were oriented directly ahead and behind of the robots' moving direction respectively . Obstacles could be detected at a maximum distance of 3 centimeters [31] . An add-on module equipped with two photodiodes on each side of the robot and pointing upwards allowed the detection of light gradients . It also carried a red LED ( Light Emitting Diode ) to permit an easy and reliable tracking in conditions of changing background brightness . A NiMH rechargeable battery provided energy for about 3 . 5 hours in our experimental conditions . The robots had a microcontroller PIC16LF877 with 8K Flash EPROM memory , 368 bytes RAM and no built-in float operations . Programming was done with the IDE of the CCS-C compiler , and the compiled programs were downloaded in the Alice memory with the PIC-downloader software ( EHL elektronika ) . A firewire digital video camera Unibrain Fire-i400 ( resolution 640×480 pixels ) was hung about 1 . 5 m above the robots . It transmitted videos to a Dell Latitude D810 laptop computer via a 1394a PCMCIA card . Image acquisition on the computer was done with the open source CMU 1394 Digital Camera Driver ( Robotics Institute , Carnegie Mellon University ) and image treatment was done with the open source OpenCV library ( Intel Corporation ) . RGB ( Red Green Blue ) images were converted into HSV ( Hue Saturation Value ) space . The rest of the treatment was done on the H channel of the HSV space . This allowed the isolation of a given color in the images ( here the red LED carried by robots ) regardless of its intensity . White noise was removed thanks to a morphological opening ( erosion followed by dilatation ) with a 3×3 matrix [32] . The images were then converted to binary images by applying a min-max threshold to isolate the red portion of the H channel . The resulting blobs of pixels were fitted with an ellipse function whose center position provided the position of each robot . Positions were corrected with respect to camera lens distortion , position and angle using the Matlab Camera Calibration Toolbox ( Computer Vision Research Group , California Institute of Technology ) . Robot positions were used to produce an image ( 800×600 pixels ) where uniform light discs of fixed blue intensity ( Red = 0 , Green = 0 , Blue = 7 ) marked trail pheromone spots . Each disc was centered on the trajectory traced by a robot and did not overlap with the previous disc drawn along the same trajectory . Discs pertaining to different trajectories or that were not directly following each other on the same trajectory could overlap . In overlapping regions , pixel intensity corresponded to the sum over time of all the overlapping discs ( up to a maximum blue intensity of 255 ) . Finally , the light intensity ( I ) decreased following an exponential decay to simulate pheromone evaporation:where corresponded to the current time , to the period between two evaporation time-steps and to the characteristic evaporation time ( 1800 sec ) . To lower the processing charge , evaporation was triggered every 5 seconds only . The tracking and trail laying software performed all computations at an effective speed of about 5 images per second . Given the robot speed of two body lengths per second and the maze dimensions this fulfilled our needs . The final image was projected with a video-projector suspended about 3 m above the robots . Misalignment between the camera and the beamer was corrected using the Matlab Camera Calibration Toolbox . The projected image covered a surface of approximately 140×105 cm . The size of the blue disc after projection was fixed to 6 cm . This allowed the formation of light trails large enough for two robots to cross without being pushed outside the trail . These parameter values that produce consistent trail laying and trail following behaviors with these robots were established in a previous study [28] . The behavioral model was a generic model of trail laying and trail following behaviors in ants . Its purpose was to capture the essential features needed to achieve a path selection as it is observed in ant colonies [33] . In the absence of light pheromones , a robot ( laying a trail or not ) moved according to a correlated random walk , which is a random walk with a directional persistence , as is commonly found in insects [34] . This behavior is called “exploratory behavior” . If the robot detected an obstacle ( with its built-in infrared detectors [31] ) , it tried to avoid it by turning away from the obstacle . This behavior was called “avoidance behavior . ” If the robot detected a luminous trail with its photoreceptors , it tried to turn towards the brighter trail . This behavior was called “trail following behavior . ” Each of these behaviors triggered the computation of a movement vector . The exploratory vector was a unit vector that initially points straight ahead of the robot and is modified at random time intervals . The new direction was chosen by drawing a random angle from a uniform distribution ( using the Quick & Dirty algorithm [35] ) between −30° and +30° and adding it to the current direction . The time intervals between each direction change were drawn from a decreasing exponential distribution with characteristic time being 3 seconds ( i . e . , an exponent of −1/3 second−1 ) . Exponential random numbers were created from a uniform random number transformed to with an algorithm using only integers ( see Ahrens and Dieter [36] for the algorithm ) . The avoidance vector was the sum of four vectors ( , , , ) , each of them pointing in the opposite direction of one of the four proximity IR sensors of the robot ( ) . The intensity of each of the four vectors increased proportionally with the inverse of the distance between an obstacle and the corresponding sensor . Each sensor regularly and frequently emitted an IR signal that was reverberated by obstacles . The intensity of the reverberation perceived by the IR sensor was used as a proxy of the distance to the obstacle . This intensity diminished with the distance approximately following a sigmoid curve ( 0: the closest obstacle from the sensor is at least 3 cm away from it; 1: the obstacle is touching the sensor ) [37] . The trail following vector was the sum of two vectors pointing either to the right ( ) or to the left ( ) of the robots' current direction ( ) . The intensity of and was controlled by the light intensities perceived by the right and left photoreceptor ( 0: no light perceived; 1: photoreceptor maximally stimulated ) . The three vectors were summed together with different weights to obtain the direction as a unit vector: The robot then adjusted the direction and speed of the rotation of two independently driven wheels to achieve the new direction during the next step of its internal clock ( 50 ms ) . Finally , the starting and the target areas in the experimental setup described above were equipped with two infrared transmitters that continuously emitted a signal . This signal was different for each area and the robots could detect it with their IR sensors . Each time a robot entered either the starting or the target area , it switched off its red LED , becoming invisible to the tracking software . As a consequence , it also stopped laying a light trail . This prevented robots from marking these areas while continuing their exploratory and obstacle avoidance walks . The red LED was switched on again as soon as the robot left the starting or the target area . All data processing and statistical analysis were performed with R version 2 . 7 . 0 [38] . In order to investigate further the respective role of pheromone and network geometry on the overall foraging efficiency , we used a computer model of our system directly inspired from the one introduced in [16] for Argentine ants , but modified to account for the robots' specificities . In the starting and target areas , robots perform a random walk ( no pheromone ) with obstacle avoidance . As a consequence their probability per unit of time of leaving the starting area , , and the target area , , can be considered constant and equal ( both areas have the same shape and dimensions ) . Once a robot has entered a segment of the network , the time required to travel the segment is computed as follows: , with the length of the segment in centimeters and the speed of the robot ( 40 mm s−1 ) . At each symmetrical intersection , a robot has to choose between two segment and . The probability for an ant to choose the segment and to choose the segment at a symmetrical bifurcation are modeled as follows:with the intrinsic attractivity of segment and , and the quantity of pheromone on segment and , respectively , and the degree of nonlinearity of the choice . At an asymmetrical bifurcation , about 2/3 of the robots choose the segment deviating less from their incoming direction when the quantity of pheromone is equal on both segment . We computed the probability to select the segment and to select the segment at an asymmetrical bifurcation as follows: corresponds to the tendency of a robot to move forward and chose the segment deviating less from its incoming direction . It is positive if segment deviates by a 30° angle from the robot's incoming direction and negative if it deviates by a 120° angle . When is equal to 0 . 5 ( i . e . , ) , then is equal to , i . e . the robot's choice is influenced only by the geometry of the bifurcation because the two segments are equally marked with pheromone . Conversely , when one of the two segments becomes more marked with pheromone , then the robot's choice becomes influenced by the trail and we assume that the influence of the bifurcation geometry progressively decreases as the difference in pheromone concentration between the two segments increases . Therefore , when or tend to ( i . e . , when or ) , tends to 0 . Finally robots add a quantity of pheromone on each segment they visit . At each time step , the pheromone intensity ( ) decreased following an exponential decay:where corresponded to the current time , to the period between two time steps and to the characteristic evaporation time ( 1800 sec ) . A good match between the experimental data and the model is found for the following parameters values: ; ; ; in configuration S , or in configuration A; when pheromone deposition is allowed , or when it is not; .
Results related to the individual behavior of ants ( taken from [16] ) and robots at asymmetrical bifurcations are summarized in Fig . 4 . The figure shows how the proportion of individuals following a given branch is influenced by the angle this branch makes with the originating branch of the individual , in the absence of other information such as recruiting pheromone . When reaching an asymmetrical bifurcation ( configuration A , Fig . 4 ) , both ants and robots chose more often to enter the branch deviating by an angle of 30° ( 126 observations in ants and 107 observations in robots ) than the branch deviating by an angle of 120° ( 66 observations in ants , , df = 1 , p<0 . 001; 38 observations in robots , , df = 1 , p<0 . 001 ) . Additionally , the proportion of robots entering the most direct branch was not significantly different from the one observed in ants ( 107/145 = 74% for robots vs . 126/192 = 66% for ants , Fisher's exact test , p = 0 . 122 ) . As shown by these results , the choice behavior of the robots at an asymmetrical bifurcation is similar to the one of the ants . In this initial phase of the experiment , branches do not bear yet any pheromone marking but the robot's simple correlated random walk leads them to “choose” the branch that deviate less from their current trajectory . This shows that no complex orientation strategy is required to reproduce the individual choice behavior of the ants with the robots . Results from the collective path selection experiments are summarized in Fig . 5 . The typical time course of an experimental replicate is shown in Fig . 2c and Video S1 . As observed in ants ( see [16] ) , robots dispersed in the network during the first minutes of the experimental replicate , before limiting their displacement to a single path connecting the starting and the target areas . This path was the shortest possible path at the end of all 15 experimental replicates in both configurations A and S . The number of network segments used by the robots increased rapidly during the first 500 seconds of an experimental replicate ( see Fig . 5a ) , which corresponded to the initial dispersion of the individuals inside the maze . It reached a plateau value around which it oscillated during the rest of the experimental replicate . This plateau value was different between the two configurations , with a mean number of segments used at around 7 for configuration S and around 5 for configuration A . While the ant and robot experiments differed in population and maze size , the dynamics of the number of segments used in both cases were qualitatively similar ( see Fig . 5a vs . d ) and indicated a more important dispersion of the individuals in configuration S of the network . Although both ants and robots tend to find the shortest path in both configurations , there is more dispersion away from this path in configuration S . In order to determine if the robots preferentially used one particular path category , we computed the mean duration of the observed selection events for each path category , which is the mean time during which the robot colony preferentially used a path category before switching to another path category ( see Fig . 5c ) . This duration varied significantly among the different path categories ( see Fig . 5c , 2 way ANOVA , F = 29 . 27 , df = ( 4 , 94 ) , p<0 . 001 ) , with the shortest path category being selected for the longest time in both network configurations ( Tukey HSD , p<0 . 001 when comparing the 4-segment category with the other path categories; comparisons with 8- and 10-segment categories was not possible since they were selected respectively 1 and 0 times only during all the experimental replicates ) . Moreover , the mean duration of observed selection events was significantly longer in configuration A ( 2 way ANOVA , F = 10 . 31 , df = ( 1 , 94 ) , p = 0 . 002 ) , as was the mean duration of selection events for the shortest path category ( Tukey HSD , p<0 . 001 ) . These results are qualitatively similar to those observed in ants ( see Fig . 5c vs f ) that also preferentially used the shorter path category in both configurations of the network , and used it more consistently in configuration A than in configuration S . The previous observation was corroborated by the analysis of the number of switches between the different path categories during an experimental replicate . Robots that started using one path category could switch to another one several times during an experimental replicate , but the number of observed selection events was significantly smaller when the network was in configuration A than when it was in configuration S ( see Fig . 5b , W = 72 , p = 0 . 024 ) . A similar result was also observed in ants ( see Fig . 5b vs . e and [16] for its statistical analysis ) . We ran the computer model under four different conditions - configuration S with and without pheromone deposition , and configuration A with and without pheromone deposition - and we compared the ability of the robotic group to complete successful trips between the starting and target areas . For each of the four conditions , we ran 1000 simulation runs . The foraging efficiency of the robotic group under each condition is summarized in Fig . 6 . The foraging efficiency is expressed as the number of successful trips performed by the robots , i . e . the number of times a robot has returned to the starting area after visiting the target area . In absence of pheromone ( S-NP and A-NP in Fig . 6 ) , robots placed in a network with configuration A performed significantly better than those placed in a network with configuration S ( Wilcoxon rank sum test , W = 193972 , p<0 . 0001 ) . However the amplitude of the improvement was small: the robots in configuration A completed only 1 . 05 times more successful trips than those in configuration S ( measured as the ratio between the median number of successful trips in both conditions ) . The addition of pheromone in the model led to a significant increase in the number of successful trips when the robots were in configuration S ( W = 16538 . 5 , p<0 . 0001 ) and configuration A ( W = 171869 , p<0 . 0001 ) . In configuration S , robots completed 1 . 3 times more successful trips when pheromone was added to the model than when it was absent . This ratio grows to 1 . 8 in configuration A . Finally robots completed 1 . 46 times more successful trips in configuration A than in configuration S in the presence of pheromone . In conclusion our simulations show that the geometry of the network has an influence on the foraging efficiency of the robots , but this influence is small compared to the one of the pheromone ( compare 1 . 05 with 1 . 3 ) . When combined they result in a nonlinear increase in the foraging efficiency ( compare 1 . 05 and 1 . 3 with 1 . 8 ) .
In numerous ant species , pheromone trails play an essential role during foraging tasks by guiding workers toward previously discovered resources or helping them finding their way back to their nest [1] . In certain species these trails form an intricate network , thus challenging the navigation abilities of ants [3] . Recent studies have shown that the geometrical structure of the trail network directly affects the choice of which path to follow when an ant crosses a bifurcation , and thus modifies the foraging efficiency of the colony [14]–[16] . It was less clear however whether or not individual workers were actively considering the geometry of a bifurcation when choosing a path to follow , though this feature was used in previous simulation work [16] , [17] . Using a robotic model , we have shown that no representation or even simple detection of the presence of a bifurcation was necessary to explain the individual ant behavior . The robots were not explicitly programmed to identify the presence of a bifurcation or to estimate its geometrical configuration . Instead they were programmed only to move according to a correlated random walk and to avoid obstacles indifferently of their nature , be they gallery walls or other robots . Yet their behavior when crossing a symmetrical or an asymmetrical bifurcation was comparable to the behavior of Argentine ants in similar situations , suggesting that the individual decisions of Argentine ants at bifurcations are affected by the physical structure of the environment in a passive way ( i . e . , without the formation of a representation of the bifurcation prior to the decision ) . Considering the poor performance of the Argentine ants' visual system [39] and the high tempo of the workers along the trail ( up to 2 . 5 cm s−1 for an average body length of 3 mm , personal observation ) , it is unlikely that Argentine ants would have the time and capacity to evaluate the geometry of a bifurcation that they would cross in less than half of a second ( the length of a bifurcation in [15] , [16] is about 1 cm from the entrance to one of the two possible exits ) . Our results show that such a complex cognitive process is not necessary to explain the ants' behavior . At the collective level , the interaction between the pheromone-based recruitment process and the tendency to move into the least deviating branch of the bifurcation created a significant difference in the pattern of network use between symmetrical and asymmetrical networks . While the robots tended to more intensely use the shorter path between the starting and target areas in both configurations , robots collectively more consistently selected the shorter path and tended to spread less in the asymmetrical network . This result was qualitatively very similar to what was observed in ants , though a quantitative comparison was not possible because of the large-scale differences between the two systems ( differences in relative speed or quantity of pheromone deposited for instance ) . Experiments with ants were also performed with colonies of 500 workers [16] , while only 10 robots were used in each of our experiments . This resulted in a larger dispersion of the individuals in the ant experiments as shown by the greater number of selection events ( see Fig . 5e ) . This increased dispersion is probably caused by overcrowding on the trail that favors the use of alternative routes [28] , [40] . Argentine ants are also known to perform more U-turns with increasing deviations from their initial trajectories [15] , [16] . However this behavior does not seem to affect the collective ability of the colony to select the shortest path in the network , as shown by simulations in [16] . Our results support this observation as robots in our experiments were not explicitly programmed to perform U-turns ( though collisions with other robots can lead to such U-turns ) and yet their collective behavior was similar to that of ants . Note that pheromone marking is essential for the path selection to occur . Without pheromones , robots would simply diffuse in the network according to their correlated random walk and approximately reflective obstacle avoidance behavior . Assuming quasi-instantaneous direction changes ( relative to the moving speed of the robots , rotation time is negligible here ) , standard diffusion theory [41] predicts a completely homogeneous distribution of the robots in the network at stationary state ( reached in our system within 10 minutes , see Fig . 5a ) . Even moderate deviations from these assumptions could not lead to the preferential use of the shorter path by the robots . Finding the shortest path between two nodes in a network requires solving a series of binary choices at each bifurcation . Following the wrong path at one bifurcation can propagate over the following decisions because of the persistent nature of the attractive pheromone , therefore decreasing the chances of finding the best solution , or even locking the system in a loop . This study shows that the coupling of a particular geometrical configuration of trail networks and the forward oriented movement of ants reduces the chances of a bad choice and favors the selection of one of the shorter paths between the nest and the food source . It has an effect similar to the heuristic information in Ant Colony Optimization ( ACO ) algorithms [42]–[44] . Both provide a general axis for the information to propagate and therefore reduce the probability that ants ( virtual and natural ) get trapped in loops or less efficient solutions [44] , [45] . This last remark raises the question of the origin of the particular geometry of the trail networks built by several ant species . In their work about foraging trails in the ant L . processionalis , Ganeshaiah and Veena ( see [12] and references therein ) mention that a branching pattern is a good trade-off in minimizing both the total length of the network and the average distance between two endpoints ( where food can be localized for instance ) . They also note that bifurcation angles that minimize the resistance to the movement of the ants in such networks should be around 70°–80° , which is close to what has been found afterward in several ant species [3] , [13] , [14] . This last point suggests that the formation of a bifurcation may be strongly influenced by the movement of ants along a trail , and that the formation of specific geometrical configurations may not require complex cognitive abilities . One possible scenario to explain the emergence of these particular angle values could be the following . A first phase of random exploration around the nest or the endpoint of an existing trail would result in a random network of weak trails . Then the passage of ants along these trails combined with their forward oriented walk would reinforce bifurcation branches that deviate from the originating direction of the ants by no more than a threshold angle ( possibly 30°–40° from the originating direction of the ant , i . e . an angle of 60°–80° between the two branches ) . Largely deviating branches would be therefore abandoned little by little . Furthermore , at bifurcations where the branches would be very close to each other , the natural diffusion of the pheromone and its imperfect detection by ants would eventually lead to the fusion of the two branches into one trail only , thus preventing the maintenance of smaller angles between the two branches of a bifurcation . A recent model of trail formation introduced in [46] confirms part of this scenario . Finally , our findings emphasize the interplay between the behavior of a swarm system and the configuration of the environment into which the swarm system moves . While most studies of ant-made networks focus on the efficiency of their topological properties ( see for instance [3] , [47] , [48] ) , we show here that their geometrical configurations also affect the spatial distribution of individuals , and hence the foraging efficiency of the colony [16] . On a related note , Batty [49] suggested that the configuration of a building could explain why a human crowd would favor certain spaces and routes more than others . We also suspect that within an ant nest , local geometrical constraints might favor the formation of preferred paths channeling the motion of ant workers . Similarly , several swarm robotics studies have shown that the shape of interacting robots could be responsible for the emergence of collective patterns [50]–[52] . In all these cases , the physical configuration of the environment ( the structure of the network , the organization of the rooms or the shape of the other individuals ) directly influences the collective outcome and can potentially modify the pattern of interaction and information exchange between individuals . Understanding the constraints applied by the environment on the behavior of individuals should make it possible to use them appropriately to improve the design of crowded areas or to favor the emergence of certain desirable behaviors in a swarm of robots .
|
Most ant species form transportation networks , be they foraging trails linking food sources to the main colony or underground galleries connecting the different parts of the nest . As for human transportation networks ( roads , airlines , etc . ) , the design and the placement of the connecting points ( or nodes ) dramatically affects the movement of individuals and hence the exchanges of material and information . In a previous study , we have shown that the geometrical configuration of these nodes ( i . e . , the angles between the different exiting branches ) can affect the route followed by an ant in a network of galleries and , as a consequence , the efficiency of the pheromone-based recruitment toward a food source . Here we show that we can reproduce these results using ant-like robots with minimal perceptual and cognitive capabilities . We demonstrate that the simple interaction between the displacement of an ant and the geometrical configuration of the gallery network can greatly affect the foraging performances of the colony . This result increases our understanding of how workers move through structures built by ant colonies and more generally points toward possible improvements for the design of man-made transportation networks .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"biotechnology",
"bioengineering",
"animal",
"behavior",
"evolutionary",
"biology",
"biology",
"zoology",
"neuroscience",
"biomimetics"
] |
2013
|
Do Ants Need to Estimate the Geometrical Properties of Trail Bifurcations to Find an Efficient Route? A Swarm Robotics Test Bed
|
Oscillating levels of adrenal glucocorticoid hormones are essential for optimal gene expression , and for maintaining physiological and behavioural responsiveness to stress . The biological basis for these oscillations is not known , but a neuronal “pulse generator” within the hypothalamus has remained a popular hypothesis . We demonstrate that pulsatile hypothalamic activity is not required for generating ultradian glucocorticoid oscillations . We show that a constant level of corticotrophin-releasing hormone ( CRH ) can activate a dynamic pituitary-adrenal peripheral network to produce ultradian adrenocorticotrophic hormone and glucocorticoid oscillations with a physiological frequency . This oscillatory response to CRH is dose dependent and becomes disrupted for higher levels of CRH . These data suggest that glucocorticoid oscillations result from a sub-hypothalamic pituitary-adrenal system , which functions as a deterministic peripheral hormone oscillator with a characteristic ultradian frequency . This constitutes a novel mechanism by which the level , rather than the pattern , of CRH determines the dynamics of glucocorticoid hormone secretion .
A fundamental requisite for survival is the ability to respond and adapt to a changing environment . This ability to respond to change or “stress” becomes more complex in multicellular organisms , and mammals have developed a well-integrated organization of hormonal , neural , and immunological systems that protect them from internal and external threats to their homeostatic state [1]–[3] . One of the most important of these systems , and one that is critical for life , is the hypothalamic-pituitary-adrenal ( HPA ) axis . This neuroendocrine system regulates the secretion of vital adrenal glucocorticoid hormones ( cortisol in humans and corticosterone in rodents ) , which have major effects on brain and metabolic function and are essential for successful recovery and adaptation to stress [4] , [5] . Central regulation of glucocorticoid secretion is predominantly coordinated by the hypothalamic peptide corticotrophin-releasing hormone ( CRH ) [6] , [7] , the efficacy of which can be significantly potentiated by other hypothalamic factors , most notably vasopressin [8] . CRH is synthesized by parvocellular neurons in the paraventricular nucleus ( PVN ) of the hypothalamus [9] , and secreted into the hypothalamic-pituitary portal circulation from axons terminating in the external zone of the median eminence [10] . Activation of CRH receptors in corticotroph cells of the anterior pituitary results in adrenocorticotrophic hormone ( ACTH ) secretion into the general circulation , which in turn stimulates glucocorticoid-secreting cells within the adrenal cortex ( Figure 1A ) . Circulating glucocorticoids regulate gene expression in cells throughout the organism via activation of two widely expressed transcription factors—the glucocorticoid receptor ( GR ) and mineralocorticoid receptor ( MR ) [11]—as well as acting directly at the cell membrane to initiate more rapid non-genomic signalling processes [12]–[15] . Glucocorticoids also feed back on their own regulatory system to inhibit HPA activity [16] , [17] . Inhibition occurs at the level of the hippocampus and hypothalamus , and with particular sensitivity at the level of the anterior pituitary ( Figure 1A ) . In common with other neuroendocrine systems that signal through pituitary hormone secretion , the HPA axis is characterized by a dynamic ultradian rhythm , which is manifested by oscillating levels of ACTH [18] and glucocorticoid hormones ( Figure 1B ) both in the blood and in the brain [19] . At the cellular level , glucocorticoid oscillations induce “gene pulsing” through rapid , transient binding of glucocorticoid receptors to promoter elements of glucocorticoid-responsive genes [20] , [21] . This dynamic and versatile transcriptional system enables cells to rapidly detect and respond to changes in circulating glucocorticoid levels and provides a sensitive mechanism for the maintenance of homeostasis [22] . Indeed , when the glucocorticoid rhythm is pharmacologically replaced by constant levels of steroid , this results in abnormal gene expression [21] , and a desensitization of physiological and behavioural responses to stress [23] , [24] . The origin of glucocorticoid oscillations is not known . Since pulsatile patterns of CRH have been detected in cultured hypothalamic explants , median eminence , and portal blood [25]–[27] , this has led to speculation that oscillations in the pituitary-adrenal system are a consequence of a neuronal “pulse generator” within the hypothalamus . However , there is a lack of concordance between pulsatile patterns of hypothalamic factors and the ultradian ACTH and glucocorticoid oscillation . In the rat , for instance , CRH pulse frequency [26] is much higher ( ∼3 pulses/h ) than the near-hourly oscillation in ACTH [18] and glucocorticoids ( Figure 1B ) . This suggests that episodic secretion of hypothalamic hormones is not the primary controlling factor of the ultradian rhythm and implies that oscillatory mechanisms exist at a sub-hypothalamic level . This concept of a “peripheral oscillator” is in keeping with in vivo lesion studies demonstrating maintenance of ultradian pulsatility following hypothalamic-pituitary disconnection in the sheep [28] , and a loss of circadian but not ultradian glucocorticoid oscillation following suprachiasmatic nucleus lesions in the rat ( unpublished data ) . Since glucocorticoids rapidly inhibit CRH-induced ACTH secretion from the anterior pituitary [29]–[32] , we postulated that this fast inhibitory feedback process provides a potential mechanism within the pituitary-adrenal system for generating oscillatory dynamics . To explore this hypothesis further , and to determine qualitatively the dynamics that result from hormonal interactions between the anterior pituitary and adrenal cortex , we previously developed a mathematical model based on differential equations that incorporates rapid glucocorticoid inhibition of ACTH secretion [33] . Numerical analysis of the model suggests that the pituitary-adrenal system can support self-sustained ACTH and glucocorticoid oscillations with a physiological ultradian frequency , even under conditions of constant CRH drive to the anterior pituitary ( Figure 1C–1D ) . In this model , the ACTH and glucocorticoid oscillations have the same frequency , but they are not synchronous—there is a small phase difference , with ACTH oscillations preceding glucocorticoid oscillations ( Figure 1D ) . The model also predicts that the capacity for this oscillatory response depends on the degree of hypothalamic drive , with higher levels of CRH resulting in damped oscillations to steady-state ( i . e . , constant ) levels of hormone ( Figure 1E–1F ) . Here we test these modelling predictions in vivo . Our data show that a constant level of CRH can activate the pituitary-adrenal system to produce ultradian hormone oscillations with a physiological frequency , and that this oscillatory activity is critically dependent on the level of hypothalamic drive , with higher levels of CRH resulting in a loss of oscillation . These results demonstrate that pulsatile secretion of hypothalamic CRH is not required for ultradian oscillatory activity in the pituitary-adrenal system , and support our theoretical hypothesis that rapid glucocorticoid inhibition at the level of the anterior pituitary is the primary factor regulating the ultradian dynamics of the system .
To test our hypothesis that the sub-hypothalamic pituitary-adrenal system functions as an ultradian hormone oscillator , we investigated the dynamics of glucocorticoid responses to different levels of constant CRH stimulation in conscious freely behaving male rats . Experiments were performed during the nadir of the circadian cycle ( 0700–1300 h ) , when there is minimal endogenous CRH [34] , [35] , and no pulsatile secretion of corticosterone , the main glucocorticoid in rodents ( Figure 1B ) . Animals were infused intravenously at a constant ( i . e . , non-pulsatile ) rate for this 6-h period with concentrations of CRH in the range 0–2 . 5 µg/h . Circulating levels of corticosterone were measured in blood samples collected prior to and throughout the infusion using an automated blood-sampling system ( see Materials and Methods ) . To check that the animals were in a physiological basal state throughout the procedure , we measured corticosterone levels in a group of control animals infused with saline ( Figure 2A–2B ) . Corticosterone levels remained low throughout the duration of the saline infusion and were not significantly different from corticosterone levels measured during the same time period ( 0700–1300 h ) in a group of untreated animals ( data not shown ) , as assessed by analysis of the area under the curve ( AUC , p>0 . 38 ) . In response to CRH infusion , corticosterone levels rose rapidly ( Figure 2C–2H ) and the overall effect was dose dependent ( AUC , p<0 . 0001 ) ( Figure 2I ) . There was a significant difference between the groups treated with saline and 0 . 5 µg/h CRH ( p<0 . 001 ) , and between the groups treated with 0 . 5 and 1 . 0 µg/h CRH ( p<0 . 001 ) , but there was no significant difference between the groups treated with 1 . 0 and 2 . 5 µg/h CRH ( p>0 . 41 ) . This suggests that both of these higher levels of CRH result in maximal pituitary-adrenal activation , which implies a systems-level “ceiling effect” in the pituitary-adrenal network . Computational modelling suggests that the dynamic activity of the pituitary-adrenal system is fundamentally dependent on the level of hypothalamic stimulation [33] . In agreement with this prediction , constant infusion of CRH at different doses gave rise to different temporal patterns of corticosterone secretion . In line with the modelling hypothesis , constant infusion of CRH ( 0 . 5 µg/h ) induced ultradian corticosterone oscillations that persisted throughout the infusion period ( Figure 2C–2D ) . In contrast , and also consistent with the qualitative predictions of the model , higher doses of CRH ( 1 . 0 and 2 . 5 µg/h ) caused a rapid activation of the adrenals , but the oscillatory component of the response was damped to constant , elevated levels of steroid ( Figure 2E–2F , CRH 1 . 0 µg/h; Figure 2G–2H , CRH 2 . 5 µg/h ) . Although CRH is the predominant ACTH secretagogue in humans and the rat [36] , its ability to stimulate ACTH secretion can be potentiated by other hypothalamic neuropeptides , most notably vasopressin [8] . However , the consistency between animals in the timing of the initial corticosterone response to CRH , and the subsequent synchrony in oscillation throughout the infusion ( Figure 2J ) indicates that the corticosterone response is not dependent on the release of any other endogenous hypothalamic factors , including vasopressin . This is in keeping with previous observations that blocking vasopressin actions on the anterior pituitary ( using a vasopressin V1b receptor antagonist ) has no effect on endogenous corticosterone oscillations during the circadian peak [37] , suggesting that vasopressin is not a significant factor in regulating the ultradian dynamics of the system . If the mechanism regulating endogenous corticosterone oscillations during the circadian peak is the same mechanism that is activated by the constant infusion of exogenous CRH , there should be good agreement between the characteristic frequencies of endogenous and CRH-induced oscillations . To test this , we computed the dominant frequency component in the CRH-induced oscillations , and compared this with the dominant frequency component in endogenous corticosterone oscillations during the circadian peak ( see Materials and Methods ) . In the CRH-induced oscillations , there was a peak frequency of ∼1 pulse/h for all animals ( Figure 3A–3B; mean peak frequency = 0 . 89 pulses/h; peak frequency range = 0 . 79–0 . 93 pulses/h ) . We then measured corticosterone levels in untreated control rats during the circadian peak , when endogenous CRH is maximal [34] , [35] and corticosterone is pulsatile ( Figure 1B ) . Corticosterone oscillations were observed in all animals ( Figure 3C ) , and frequency analysis of the data ( Figure 3D ) revealed no significant difference between these endogenous oscillations and the oscillations induced by constant CRH infusion ( p>0 . 57 ) ( Figure 3E ) . The consistency in oscillation frequency between different animals infused with constant CRH suggests that the oscillations are regulated by a generic mechanism at a sub-hypothalamic level . Furthermore , the maintenance of this dominant frequency component throughout the period of CRH infusion ( Figure 3F ) suggests that the underlying oscillator is deterministic—as opposed to stochastic—in agreement with our modelling hypothesis [33] . Numerical simulations of the model indicate that glucocorticoid oscillations induced by constant CRH stimulation are driven by oscillations in ACTH ( Figure 1C–1D ) . Ultradian ACTH oscillations have been observed in the rat [18] , and have been shown to be a critical factor in regulating pulsatile glucocorticoid secretion from the adrenal cortex [38] . Moreover , coordinated ACTH and glucocorticoid oscillations have been observed in man [39] . To confirm that the constant CRH infusion generates oscillations in both hormones , we infused CRH ( 0 . 5 µg/h ) from 0700–0940 h and measured circulating levels of both ACTH and corticosterone in samples collected at 20-min intervals throughout the infusion ( see Materials and Methods ) . In agreement with the modelling predictions , CRH induced ACTH and corticosterone oscillations that persisted throughout the infusion period ( Figure 4A–4B ) . A key feature of the oscillation predicted by numerical simulations is a small phase shift between the two hormones—ACTH oscillations preceding glucocorticoid oscillations ( Figure 1C–1D ) . This small phase shift could not be detected in the CRH-induced ACTH and corticosterone oscillations ( Figure 4A–4B ) because of the sampling frequency ( 20-min inter-sample interval ) . Therefore , we measured both ACTH and corticosterone at a higher sampling frequency ( 5-min inter-sample interval ) over the first 25 min of the CRH infusion ( 0 . 5 µg/h ) , covering the initial activation phase of the first pulse . Both hormones rose rapidly in response to CRH ( ACTH , p<0 . 0001; corticosterone , p<0 . 005 ) , with the ACTH increase preceding a delayed rise in corticosterone ( Figure 4C–4D ) . Specifically , ACTH was significantly different from basal ( time zero ) by 10 min ( p<0 . 005 ) , whereas corticosterone was not significantly different from basal ( time zero ) until 20 min ( p<0 . 05 ) . This phase shift between ACTH and corticosterone presumably reflects the time taken for de novo biosynthesis and release of corticosterone from the adrenal cortex [40] . We then checked that this phase relationship is maintained over the full pituitary-adrenal oscillation . Since collection of large volumes of plasma ( required for sensitive ACTH assay ) can activate a stress response [41] , this precludes high-frequency blood sampling ( for ACTH measurement ) for prolonged periods in the rat . We therefore used an alternative experimental approach in which animals were killed by decapitation at 10-min intervals throughout the first 70 min of the CRH infusion ( 0 . 5 µg/h ) ; ACTH and corticosterone levels were measured in plasma obtained from trunk blood ( see Materials and Methods ) . As observed in the case of high-frequency sampling ( Figure 4C–4D ) , the CRH-induced increase in corticosterone was delayed relative to the increase in ACTH , and this phase shift was maintained over the duration of the pulse ( Figure 4E ) . These results challenge the long-standing view that glucocorticoid oscillations are a consequence of pulsatile CRH secretion from a neuronal “pulse generator” within the hypothalamus [26] , [27] . Our approach was based on the premise that feedback is a key regulatory feature of biological oscillators [42] , and on our numerical results [33] , suggesting that a systems-level dynamic balance between positive feedforward and negative feedback—CRH stimulation against rapid glucocorticoid inhibition of ACTH secretion—provides a mechanism for generating ultradian oscillations in ACTH and glucocorticoid hormone secretion . Testing this modelling prediction in vivo , our results show that constant CRH stimulation of the anterior pituitary is sufficient to generate ACTH and glucocorticoid oscillations at a physiological ultradian frequency , providing good evidence for an oscillating mechanism outside the central nervous system . The hormone oscillations generated by this system are not simply a dynamic epiphenomenon of the pituitary-adrenal interaction , but have significant biological impact . Glucocorticoid oscillations are essential for optimal transcriptional regulation [20] , [21] , and are also likely to be important for more rapid non-transcriptional mechanisms of steroid action in the brain [43] that can alter behavioural function within minutes [44] , [45] . Exposure of tissues to abnormal glucocorticoid levels due to prolonged stress [46] and raised CRH [47] , [48] , or due to the loss of ultradian pulsatility that has been detected in ageing animals [49] , will modify glucocorticoid signalling , and could be an important factor for both stress- and age-related disease .
All experiments were conducted on adult male Sprague-Dawley rats ( Harlan Laboratories , Inc . ) weighing ∼250 g at the time of surgery . Animals were housed in groups of four per cage and were given at least 1 wk to acclimatize to the housing facility prior to surgery . Rats were maintained under standard environmental conditions ( 21±1°C ) under a 14-h light , 10-h dark schedule ( lights on at 0500 h ) . Food and water were freely available throughout the experiments . All animal procedures were conducted in accordance with Home Office guidelines and the UK Animals ( Scientific Procedures ) Act , 1986 . Animals were anaesthetized with a combination of Hypnorm ( 0 . 32 mg/kg fentanyl citrate and 10 mg/kg fluanisone , IM; Janssen Pharmaceuticals ) and diazepam ( 2 . 6 mg/kg , IP; Phoenix Pharmaceuticals ) . Two silastic-tipped ( Merck ) polythene cannulae ( Portex ) were pre-filled with pyrogen-free heparinized ( 10 IU/ml ) isotonic saline . The right jugular vein was exposed and both cannulae inserted into the vessel until they lay close to the entrance to the right atrium . This permits simultaneous intravenous blood sampling ( via the sampling cannula ) and substance infusion ( via the infusion cannula ) . The free ends of the cannulae were exteriorized through a scalp incision and tunnelled through a protective spring that was anchored to the parietal bones using two stainless steel screws and self-curing dental acrylic . Following recovery , animals were individually housed in a room containing an automated blood-sampling system . The end of the protective spring was attached to a two-channel swivel ( Instech Laboratories , Inc . ) , which rotates through 360° in the horizontal plane and 180° in the vertical plane , providing the animals with maximal freedom of movement . Animals were given a 5-d recovery period prior to experiments . Throughout this time , both cannulae were flushed daily with heparinized saline to maintain patency . Rats received a constant intravenous infusion of either 0 . 9% saline solution ( vehicle control animals ) or synthetic human/rat CRH ( University of Bristol Peptide Synthesis Service ) dissolved in 0 . 9% saline solution . In all experiments , drugs were infused through the infusion cannula at a volume infusion rate of 0 . 5 ml/h using an automated infusion pump ( PHD ULTRA syringe pump; Harvard Apparatus , Ltd . ) . In the experiments measuring basal corticosterone levels over 24 h , corticosterone levels in response to saline or CRH infusion , or basal corticosterone oscillations during the circadian peak , blood samples were collected from the sampling cannula using an automated blood-sampling system [50] . In the experiment measuring basal corticosterone levels over 24 h ( Figure 1B ) , blood samples were collected every 10 min from 1400–1350 h . In the experiments measuring corticosterone levels in response to saline or CRH infusion ( Figures 2A–2H , 2J , and 3A ) , rats were constantly infused with saline or CRH ( 0 . 5–2 . 5 µg/h ) from 0700–1300 h; blood samples were collected every 5 min from 0600–1300 h . In the experiments measuring basal corticosterone oscillations during the circadian peak ( Figure 3C ) , blood samples were collected every 5 min from 1530–2100 h . At the end of each experiment , the plasma fraction was separated by centrifugation ( 15 min , 3 , 120 g , 4°C ) and stored at −20°C until processed for corticosterone measurement . In the experiments measuring ACTH and corticosterone oscillations in response to CRH infusion ( Figure 4A–4B ) , rats received a constant CRH infusion ( 0 . 5 µg/h ) from 0700–0940 h and blood samples were collected manually from the sampling cannula every 20 min throughout the infusion . In the experiments measuring ACTH and corticosterone levels in response to CRH during the initial activation phase ( Figure 4C–4D ) , rats received a constant CRH infusion ( 0 . 5 µg/h ) from 0700–0725 h and blood samples were collected manually from the sampling cannula every 5 min throughout the infusion . Blood samples from both experiments were immediately mixed with EDTA ( 10 µl , 0 . 5 M , pH 7 . 4 ) and placed on crushed ice . The plasma fraction was separated by centrifugation ( 15 min , 3 , 120 g , 4°C ) within 20 min of sample collection and stored at −20°C until processed for ACTH and corticosterone measurement . In the experiment measuring ACTH and corticosterone levels in response to CRH infusion over the first pulse ( Figure 4E ) , rats received a constant CRH infusion ( 0 . 5 µg/h ) from 0700–0810 h . At 10-min intervals throughout the CRH infusion , rats were killed by decapitation following an overdose of 0 . 3 ml sodium pentobarbital ( Euthatal , 200 mg/ml; Merial ) . Trunk blood was collected and immediately mixed with EDTA ( 50 µl , 0 . 5 M , pH 7 . 4 ) and placed on crushed ice . The plasma fraction was separated by centrifugation ( 15 min , 3 , 120 g , 4°C ) within 20 min of sample collection and stored at −20°C until processed for ACTH and corticosterone measurement . Corticosterone was measured in plasma by radioimmunoassay ( RIA ) [51] . Samples were diluted in a citrate buffer ( pH 3 . 0 ) to denature the binding globulin . Antisera was supplied by G Makara ( Institute of Experimental Medicine , Budapest , Hungary ) , and [125I] corticosterone was purchased from Izotop ( Institute of Isotopes Co . Ltd . , Budapest , Hungary ) . The intra- and inter-assay coefficients of variation of the corticosterone RIA were 14 . 1% and 15 . 3% , respectively . ACTH was measured in plasma by immunoradiometric assay ( IRMA; DiaSorin Ltd . ) , in accordance with the manufacturer's protocol . The intra- and inter-assay coefficients of variation of the ACTH IRMA were 2 . 8% and 6 . 4% , respectively . Overall corticosterone responses to saline or CRH infusion were assessed by AUC . Characterization of oscillatory corticosterone responses induced by CRH infusion , and of endogenous corticosterone oscillations recorded during the circadian peak , was performed using spectral methods . Missing data points were linearly interpolated and data were detrended using the Smoothness Priors Approach ( SPA ) [52] with the smoothing parameter set at λ = 30 . This parameter value was chosen so as to remove long term changes in the mean ( i . e . , low-frequency fluctuations ) , while keeping the higher-frequency ultradian fluctuations that were the focus in this study . To define the frequency of the oscillatory corticosterone data , we computed the power spectrum of the detrended data using the Discrete Fourier Transform ( DFT ) applied to a time window corresponding to the period of CRH infusion ( 0700–1300 h ) , or to the period of sampling ( 1530–2100 h ) for the basal corticosterone oscillations recorded during the circadian peak . The peak frequency was then taken as the frequency value corresponding to the maximum spectral power of the DFT , which was calculated using a quadratic interpolation . Spectrograms were computed using the Short-Time Fourier Transform ( STFT ) . Statistical significance level was set at p<0 . 05 . Saline-infused corticosterone responses and untreated corticosterone profiles were compared using the non-parametric Mann-Whitney U test . The overall effect of CRH treatment on AUC was assessed using the non-parametric Kruskal-Wallis ANOVA on ranks test , and post hoc multiple comparisons were performed using the Mann-Whitney U Test with the statistical significance level adjusted using the Bonferroni correction . Peak frequencies obtained using the DFT , for the groups with CRH-induced corticosterone oscillations and endogenous corticosterone oscillations during the circadian peak , were compared using the Mann-Whitney U test . The ACTH and corticosterone response to constant CRH infusion was analyzed using one-way ANOVA , followed by Fisher Least Significant Difference ( LSD ) post hoc test .
|
Glucocorticoid steroid hormones , such as cortisol and corticosterone , provide a rapid response to both physical and psychological stress , and act on areas of the brain that influence learning , memory , and behaviour . Glucocorticoids are released from the adrenal glands in near-hourly pulses , which results in oscillating glucocorticoid levels in the blood and in target organs . These hormone oscillations can become disrupted during ageing and in stress-related disease ( e . g . , major depression ) , so it is important to identify the underlying mechanisms that govern their dynamics . Although the origin of the oscillations is not known , it is assumed that they are generated by a neuronal “pulse generator” within the brain . In this study , we present data that challenge this hypothesis . We characterize a peripheral hormonal system and show that constant levels of corticotrophin-releasing hormone can induce and regulate hormone oscillations independent of the brain . We also describe mechanisms that can disrupt these oscillations . These findings have important implications for our understanding of glucocorticoid signalling in both health and disease , and will be important for the design of novel treatment strategies that take into account timing of hormone administration to patients undergoing steroid therapy for inflammatory or malignant disease .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"pituitary",
"applied",
"mathematics",
"anatomy",
"and",
"physiology",
"homeostatic",
"mechanisms",
"neuroscience",
"hormones",
"endocrine",
"physiology",
"mathematics",
"endocrine",
"glands",
"endocrine",
"cells",
"theoretical",
"biology",
"endocrinology",
"diabetes",
"and",
"endocrinology",
"biology",
"nonlinear",
"dynamics",
"neuroendocrinology",
"systems",
"biology",
"neural",
"homeostasis",
"physiology",
"computational",
"biology",
"neurophysiology",
"genetics",
"and",
"genomics"
] |
2012
|
The Origin of Glucocorticoid Hormone Oscillations
|
Most of the ATP in living cells is produced by an F-type ATP synthase . This enzyme uses the energy of a transmembrane electrochemical proton gradient to synthesize ATP from ADP and inorganic phosphate . Proton movements across the membrane domain ( FO ) of the ATP synthase drive the rotation of a ring of 8–15 c-subunits , which induces conformational changes in the catalytic part ( F1 ) of the enzyme that ultimately promote ATP synthesis . Two paralogous nuclear genes , called Atp9-5 and Atp9-7 , encode structurally different c-subunits in the filamentous fungus Podospora anserina . We have in this study identified differences in the expression pattern for the two genes that correlate with the mitotic activity of cells in vegetative mycelia: Atp9-7 is transcriptionally active in non-proliferating ( stationary ) cells while Atp9-5 is expressed in the cells at the extremity ( apex ) of filaments that divide and are responsible for mycelium growth . When active , the Atp9-5 gene sustains a much higher rate of c-subunit synthesis than Atp9-7 . We further show that the ATP9-7 and ATP9-5 proteins have antagonist effects on the longevity of P . anserina . Finally , we provide evidence that the ATP9-5 protein sustains a higher rate of mitochondrial ATP synthesis and yield in ATP molecules per electron transferred to oxygen than the c-subunit encoded by Atp9-7 . These findings reveal that the c-subunit genes play a key role in the modulation of ATP synthase production and activity along the life cycle of P . anserina . Such a degree of sophistication for regulating aerobic energy metabolism has not been described before .
Most energy-transducing membranes in bacteria , mitochondria and chloroplasts contain an F-ATP synthase . Historically this enzyme has been described in terms of an integral membrane domain ( FO ) and a peripheral domain ( F1 ) . High-definition structural data now available reveal that FO and F1 are further divided into distinct oligomeric sub-structures [1–4] . The most hydrophobic FO subunits associate in a unit , composed of one a-subunit adjacent to a ring of 8–15 c-subunits , that transfers protons from one side of the membrane to the other . The remaining FO subunits assemble a stalk contacting at one end the a-subunit in the membrane and attached at the other end to the external periphery of the soluble F1 domain . F1 is made of a globular structure ( [αβ]3 hexamer ) that contains three catalytic sites , one in each β-subunit , and 2–3 other proteins that constitute a central stalk connecting these sites at one end with the c-ring at the other end . Energy released by proton translocation through the FO activates the rotation of the c-ring together with the central stalk relative to the non-moving parts of ATP synthase , which forces conformational changes in the catalytic sites that facilitate product ( ATP ) release from the enzyme [5] . The proton channel in F-ATP synthases extends along a proteinacious interface that is formed by transmembrane a-helices of subunits a and c . The individual monomers of the c-ring adopt a hairpin structure , composed of two membrane-spanning α-helices ( numbered 1 and 2 ) joined by a short hydrophilic linker . In cross section , the c-ring shows a double annular structure with inner and outer circles comprised of helices 1 and 2 , respectively . According to this arrangement , only helix 2 contributes to the proton channel . A recently proposed model of the a/c-ring complex defines the proton channel by α-helices 4 and 5 in the a-subunit and helix 2 of the c-subunit [6] . The most notable feature of this domain is a universally conserved acidic amino acid ( Glu or Asp ) in c-subunit that participates directly in proton transfer . The acidic residue is located in the middle of helix 2 where it would project into the hydrophobic phase of the bilayer around the perimeter of the c-ring . Hence , its carboxyl group is assumed to be protonated in all of the c-subunits except for the monomer that is located at the a/c-ring interface . Here , the hydrophilic environment inside the ion channel would favor the ionized carboxylate . In the proposed series of events that occur during proton translocation the carboxylate group inside the channel is neutralized with a proton that originates from the p-side of the membrane , followed by rotation of the c-ring , which moves the newly protonated c-subunit to the lipid phase and the adjacent monomer into the channel . Deprotonation of the incoming carboxyl group transfers a proton to the channel , and ultimately to the n-side of the membrane to complete the process of uni-directional proton translocation . Previous work revealed the existence in the filamentous fungus Podospora anserina of two separate nuclear genetic loci ( Atp9-5 and Atp9-7 ) that encode two different c-subunit isoforms that share 66% sequence identity ( in their mature part ) [7] . Work by others has revealed the presence of multiple isogenes encoding the c-subunit also in the nuclear genomes of mammals [8 , 9] and the parasite Trypanosoma brucei [10] . All three mammalian isoforms are identical beyond the mitochondrial leader cleavage site , whereas those from T . brucei show between them ( after the predicted processing site ) sequence variation in the first 20% of the primary sequence . All three isoforms are produced and assembled in the T . brucei enzyme during the different life stages of the parasite , though it remains to be determined if individual enzyme molecules are mosaic with respect to the c-subunit or if each isoform preferentially assembles with the same type such that there are different populations of the F-ATP synthase in the cell . The c-subunits in P . anserina are distinguished on the basis that the sequence variation is distributed throughout the entire length of the mature proteins . The vegetative phase of P . anserina is initiated by ascospore germination , which gives rise to a network ( thallus ) of branched filaments or hyphae that spread out to form a mycelium . Unlike plant and animal cells , those of filamentous fungi form a continuous multi-nucleated cytoplasm , the growth of which is restricted to the tip ( apex ) of hyphae ( polarized growth ) . For simplification , “proliferating or apical cells” will be used henceforth to designate the growing apex , while “non proliferating cells” will designate the non-growing part of hyphae backwards the apex . The current paper shows that in vegetative cultures of P . anserina , Atp9-5 and Atp9-7 are expressed in different locations; the former at the apex of hyphae and the latter in non-proliferating cells that comprise the bulk of the fungal mycelium . We also show that the proteins encoded by Atp9-5 and Atp9-7 ( referred to as ATP9-5 and ATP9-7 respectively ) have antagonist effects on the longevity of P . anserina and confer to the mitochondria different bioenergetics properties . These findings reveal that the two c-subunit genes play a crucial role in the modulation of mitochondrial energy transduction along the life cycle of this filamentous fungus .
The relative abundance of mRNA transcripts from Atp9-5 and Atp9-7 , versus a constitutively expressed gene ( Gpd ) , was determined by real-time quantitative reverse transcription PCR as described under Materials and Methods . RNA extracts were prepared from whole mycelium cultured on solid media for 1 day ( w-1d ) , 2 days ( w-2d ) , or 5 days ( w-5d ) . Atp9-5 transcripts were 20-fold more abundant than Atp9-7 transcripts in w-1d RNA samples ( Fig 1 , w-1d ) . With time , the level of Atp9-5 transcripts declined to near zero while Atp9-7 transcripts rose rapidly , reaching a steady state by day 2 ( Fig 1 , w-2d , w-5d ) . These results suggested that Atp9-5 and Atp9-7 are expressed differentially in proliferating and non-proliferating cells , respectively . The apex of hyphae establishes a zone of proliferation around the perimeter of the discoidal mycelium in plate cultures of P . anserina . These mitotically active cells comprise a significant percentage of total cells in 1-day-old mycelium , but they are far surpassed in number by non-dividing cells that accumulate during radial expansion of the mycelium . As such , any transcript unique to apical cells ( e . g . Atp9-5 ) would be a minor species in the total RNA extracted from whole preparations of 5-day-old mycelium , and difficult to detect ( Fig 1 , w-5d ) . Our hypothesis was confirmed in experiments that targeted cells isolated from different regions of mycelium . Non-proliferating cells collected from the center of 5-day-old mycelium were abundant in Atp9-7 transcripts and almost completely deficient for Atp9-5 transcripts ( Fig 1 , c-5d ) . Instead , peripheral samples collected from the same mycelium ( Fig 1 , p-5d ) , enriched for apical cells , contained a significant amount of Atp9-5 transcripts , but it was not clear if the Atp9-7 transcripts that were co-detected originated from mitotically active apical cells or from stationary cells present in the peripheral sample . Therefore , protoplasts derived from apical cells of 2-day-old and 5-day-old mycelia were isolated and used to prepare RNA samples that better reflected the transcriptional activity unique to proliferating cells . In the end we found that Atp9-5 transcripts clearly dominated Atp9-7 transcripts in these cells ( Fig 1 a-2d , a-5d ) . Cumulatively , these results lead us to propose that the origin of c-subunit mRNA in P . anserina is dictated by the mitotic status of the cells . It is also noteworthy that the levels of Atp9-7 transcripts in w-2d and w-5d were much lower compared to those of Atp9-5 in w-1d . Hence it would appear that the transcriptional switch from Atp9-5 in proliferative cells to Atp9-7 in non-proliferative cells is accompanied by a considerable decrease in the rate of c-subunit synthesis . Unfortunately , we failed to raise antibodies that specifically recognized the ATP9-5 and ATP9-7 proteins , which could have been especially useful to determine directly how these proteins are expressed along the life cycle of P . anserina . Furthermore , while these proteins can functionally substitute for yeast subunit 9 [11] adding tags to them severely compromises ATP synthase function ( S1 Fig ) . The use of fluorescence protein markers , like GFP , under control of the regulatory sequences of Atp9-7 and Atp9-5 did not seem to us a good alternative too . Indeed , though the transcripts data indicate that ATP9-5 is preferentially , if not exclusively , synthesized in proliferating cells , this does not mean that it will not be present in non-proliferating cell . As a result , and considering the well-known stability of GFP , this protein will likely distribute throughout the entire mycelium even if it is exclusively synthesized at the apex . As an alternative approach , we took advantage of the sensitivities of P . anserina to some chemical inhibitors to evaluate cis-regulatory sequences in Atp9-5 and Atp9-7 for the initiation of growth and the continued propagation of fungal mycelium on selective media . Since growth occurs only at the apex of filaments ( see above ) , any gene that is required to confer resistance to the inhibitor must be actively transcribed in apical cells . Instead , if the required gene is silent or transcribed too poorly at the apex to provide a sufficient amount of the necessary protein , the mycelium will retain inhibitor sensitivity and show a growth defect on the inhibitor-containing plates . We previously constructed a strain of P . anserina in which Atp9-5 expression is controlled by Atp9-7 regulatory sequences and vice versa for Atp9-7 ( strain [75][57] ) , and showed that it completes the full life cycle [7] . We interpreted this finding to indicate that the two c-subunit isoforms are equivalent functionally . However , having now examined the vegetative growth stage in greater detail , a functional difference between isoforms encoded by Atp9-5 and Atp9-7 was made apparent . Indeed , swapping the regulatory sequences of Atp9-5 and Atp9-7 ( strain [75][57] ) leads to a significant increase in longevity with respect to the wild type ( 123% , p<2x10-5 , Fig 3A and S3 Table ) , as evaluated by the linear length the mycelium reaches before dying . To pursue further the idea that selective expression of Atp9-7 at the apex causes an increase of longevity , the life span was measured for additional strains in which the expression of Atp9-5 and/or Atp9-7 is regulated differently than in wild type . Compared to wild type , the life span was reduced significantly for strain 55 ( 79% , p<5x10-5 , Fig 3 and S3 Table ) , which is completely deleted for Atp9-7 and this phenotype persisted in strain [75]55 , which harbors a second copy of the Atp9-5 coding sequence that is flanked by Atp9-7 regulatory sequences ( S1 Table ) . Instead , for the case in which the only source of the c-subunit is Atp9-7 , fungal longevity was increased ( 129% , p<3x10-6 , Fig 3 and S3 Table ) as long the strain carries a copy of the gene that is regulated by Atp9-5 sequences ( strains [57] and 77[57] ) . Cumulatively , these results show that the exclusive production of c-subunit encoded by Atp9-7 correlates with an increased fungal life span while longevity is compromised in strains for which production of the c-subunit is limited to the Atp9-5 coding sequence . Estimation of the life span by the number of days by which 50% of the cultures were still alive ( half-live ) further supported this conclusion ( Fig 3B , S3 Table ) . The results of the longevity experiments were interesting in light of previous work that showed the life span of P . anserina is sensitive to factors that impact mitochondrial oxidative phosphorylation: fungal longevity was increased in mutants carrying a genetic defect linked to one of the respiratory complexes [19–24] ( see Discussion ) . We hypothesized that some aspect of the F-ATP synthase related to fungal bioenergetics and life span might be modified by the incorporation of different c-subunit isoforms in its structure . P . anserina strains [57] and 55 ( described above , see S1 Table ) were ideally suited for experiments to investigate this idea , first because each strain is genetically pure for only one of the c-subunit isoforms . Second , the regulatory sequences from Atp9-5 control the gene transcription for both types of c-subunits and this guarantees uniformity between strains with respect to the amount of c-subunit mRNA that is produced in the apical cells of the mycelial cultures ( S2 Fig ) . Mitochondria were isolated from the apical cells of strains [57] and 55 . For simplicity , the samples are hereafter referred to as MitoATP9-7 and MitoATP9-5 , respectively to denote which c-subunit isomer was produced in the cells of origin . BN-PAGE analyses revealed that the F-ATP synthase was in similar , if not identical , amounts in MitoATP9-5 and MitoATP9-7 samples that contained the same amount of porin ( Fig 4A and 4B; and S4 Fig where quantification of ATP synthase from two independent experiments is provided ) . Oxygen consumption and ATP synthesis rates were measured using NADH as a respiratory substrate . It is to be noted that external NADH cannot be oxidized by complex I because the catalytic site of this complex is located in the mitochondrial matrix [25] . NADH can nevertheless deliver electrons to ubiquinone ( Q ) via two monomeric NADH dehydrogenases ( NDE1 and NDE2 ) located on the outer side of the inner membrane and these electron transfers are not coupled to the translocation of protons across the mitochondrial inner membrane [19] . Thus , using external NADH , the proton gradient for ATP synthesis is generated by electron transfer from reduced ubiquinone ( QH2 ) to oxygen , through respiratory complexes III and IV , which is coupled to the pumping of protons out of the mitochondrial matrix . A critical factor for interpreting these experiments was that the respiratory data could be used quantitatively provided there was not another pathway for oxygen consumption in the mitochondria . On this point , P . anserina contains a gene ( Aox ) encoding an alternative oxidase ( AOX ) that bypasses complexes III and IV and transfers electrons directly from ubiquinol to oxygen without generating a proton gradient . In other words , AOX consumes oxygen in a reaction that is not coupled to ATP synthesis and , if present in MitoATP9-5 or MitoATP9-7 mitochondria , would complicate calculations of oxidative phosphorylation parameters . Fortunately this gene is expressed in the fungus only under peculiar conditions , for instance when the external medium is poisoned with chemicals ( like antimycin , myxothiazol , cyanide ) that inhibit complexes III or IV and in strains with loss-of-function mutations in complex III or complex IV [19–24] . In accord with our expectations , there was hardly any evidence of Aox transcripts detected by qRT-PCR with mycelial RNA samples from four different fungal strains , including those ( [57] and 55 ) used to isolate mitochondria for the studies described in this section ( S2 Fig , bottom panel ) . In fact , the Aox mRNA level was even too low to be quantified relative to the same control transcript ( Gpd ) used to quantify Atp9-5 and Atp9-7 expression ( S2 Fig , top and middle panels ) . For this reason the Aox transcript levels are reported relative to those from a poorly expressed control gene ( Pdf2 ) and shown on a different scale in the bottom panel of the figure . Consistent with these transcript analyses , we failed to detect by Western blot the AOX protein in the MitoATP9-5 and MitoATP9-7 samples ( Fig 4B ) . Finally , respiration was in both mitochondria only minutely ( 5% ) inhibited by SHAM , a specific inhibitor of AOX ( Fig 4C ) . As such , we are confident that oxidative phosphorylation was the only significant pathway available for oxygen consumption in the MitoATP9-5 and MitoATP9-7 samples . Oxygen consumption ( Fig 4D ) and ATP synthesis rates ( Fig 4E ) in the presence ADP ( state-3 ) were significantly higher in MitoATP9-5 versus MitoATP9-7 ( see also S4 Table ) . However , the similar rates in oxygen consumption that were measured for both samples in the presence of an uncoupling agent ( CCCP ) indicated that the maximal respiratory capacity and potential to generate a proton motive force were actually quite comparable ( Fig 4D , S4 Table ) . In contrast , the capacity of the F-ATP synthase to use the proton motive force for ATP synthesis differed significantly; the yield in ATP per pair of electrons transferred to oxygen was higher in MitoATP9-5 with respect to MitoATP9-7 ( 1 . 51 versus 1 . 25 , Fig 4F ) . Most importantly , the differences were reproducibly observed in three independent experiments ( S4 Table ) . With equivalent amounts of F-ATP synthase , the c-subunit was the only factor relevant to energy coupling that was different between the MitoATP9-5 and MitoATP9-7 samples . The most straightforward interpretation of these experiments is that in P . anserina different isoforms of the c-subunit are incorporated in the F-ATP synthase to regulate the ATP/O ratio and control how energy is utilized by cells .
A previous paper reported that the c-subunit of the mitochondrial F-ATP synthase is encoded by two paralogous genes , Atp9-5 and Atp9-7 , in the nuclear genome of the filamentous fungus P . anserina [7] . The results from early work to characterize isogene expression showed that both genetic loci were active throughout the fungal life-cycle [7] . Also , Atp9-7 proved to be transcribed preferentially during spore maturation while Atp9-5 expression was clearly dominant in germinating spores . We have since completed a detailed transcriptional analysis in vegetative mycelium with respect to both the levels of Atp9-5 and Atp9-7 mRNAs ( Fig 1 ) and properties of the 5’ and 3’ sequences that flank the Atp9 coding sequences ( Fig 2 ) . The reason for the profound difference in the level of Atp9-5 transcripts between whole mycelium samples on the first and last days ( w-1d , w-5d ) was not immediately clear . Though if Atp9-5 expression was limited to only the proliferating cells , the data trend was in accord with the diminishing contribution of these to the total fungal mass as the mycelium expanded radially . Since gross dissection of peripheral region from whole mycelium ( Fig 1 , p-5d ) does not provide a pure sample of only the actively proliferating cells , an accurate description of the expression pattern necessitated the preparation of protoplasts originating from apical cells . Such samples faithfully reported the properties exclusive to apical cells . Combining the results of these samples ( a-2d , a-5d ) with those from RNAs that were isolated from the stationary phase mycelium removed from the center of the thallus ( c-5d ) reveals clearly that the expression of c-subunit isogenes in P . anserina tracks with the mitotic status of the cells; the non-proliferating cells , which comprise most of the vegetative apparatus in developed mycelium , accumulated a steady state level of Atp9-7 transcripts and little , if any , from the other gene , while Atp9-5 was preferentially and very highly expressed in the growth zone where the apical cells of filaments are located . Chimeric genes in which the coding sequence for a reporter protein is flanked with the upstream and downstream regions from either the Atp9-5 or Atp9-7 provided insight to how the isogenes are regulated in vivo . In short , only the sequences from Atp9-5 permitted expression of the chimeric genes at a high enough level in order to initiate and support growth on nourseothricin ( Fig 2A ) or oligomycin ( Fig 2B ) . These findings indicated that the near absence of Atp9-7 transcripts from apical cells is solely a function of the flanking sequences , which do not respond to a transcriptional activation signal ( or signals ) in apical cells , and has nothing to do with the coding region of the gene . We wish to point out here that the impetus to mutagenize P . anserina , both the wild type ( 7755 ) and the strain that was optimized to enable high production of the c-subunit encoded by Atp9-7 in the proliferating cells at the apex ( 77[57] ) , was to isolate oligomycin-resistant alleles of both c-subunit isomers for construction of the chimeric gene reporters . In fact , the mutant alleles that encode the c-subunit with either the F124S substitution ( in ATP9-5 ) or F135Y substitution ( in ATP9-7 ) are interesting in their own right and relevant to the broad field of ATP synthase bioenergetics . The efficacy of oligomycin resistance , which we observed for single substitution at either F124 or F135 , was not surprising because both phenylalanine residues are identically conserved in the c-subunit from Saccharomyces cerevisiae ( numbered F55 and F64 in the yeast protein ) and map to the oligomycin binding site in the ATP synthase of this organism ( [15] , see S5 Fig ) . Moreover , published work had already identified an oligomycin-resistance conferring mutation affecting F64 in the yeast c-subunit [16] . Hence , we are not the first to show that a single mutation of the amino acid that occupies the position of F135 in the c-subunit of P . anserina is sufficient to prevent oligomycin from interfering with proton translocation in the ATP synthase . On a conceptual basis , the expression pattern of Atp9-5 and Atp9-7 in mycelium was similar to that observed during early development [7]; the two isogenes were differentially expressed in individual cell populations , as a function of differences in the regions that flank the protein coding sequences , and the amount of mRNA for the c-subunit was significantly higher in active versus resting cells . This regulatory strategy has important implications for aerobic energy metabolism because the cellular content of F-ATP synthase can be controlled by how much c-subunit is available . Together these results indicate that the transcriptional regulation of Atp9-5 and Atp9-7 is used to modulate ATP synthase production during the life cycle of P . anserina: when large amounts of ATP synthase need to be produced , i . e . during spore germination and vegetative growth , the c-subunit is mainly , if not exclusively , expressed by Atp9-5 , whereas Atp9-7 becomes the main source of c-subunit in sedentary cells and maturing spores that require a lesser amount of newly-synthesized F-ATP synthase . While it makes intuitive sense that ATP synthase production needs to be modulated along the life cycle of P . anserina , it was not at all obvious why as part of this strategy two different c-subunit gene isoforms are used . Indeed , one single gene controlled by a nutrient sensing system could a priori be sufficient . For example in humans and other mammals , there is a single species of the c-subunit that is produced from three different genetic loci , in which unique characteristics of the flanking sequences allow the modulation of ATP synthase levels in a cell- and tissue-specific manner [26] . The results from our experiments with mitochondria isolated from strains that produced equivalent amounts of one or the other c-subunit isoform have provided initial evidence of a two-pronged regulatory mechanism that is based not only on the availability of c-subunit but also on characteristics of the protein itself . Despite the similarities in ATP synthase content and rates of oxygen consumption in the presence of CCCP , the ATP/O ratio was measured to be 1 . 51 in MitoATP9-5 and 1 . 25 in MitoATP9-7 ( Fig 4 , S4 Table ) . The implication of this finding is that the individual c-subunit isomers might confer different functional properties to the ATP synthase . Indeed , the two mitochondrial samples show almost the same passive permeability to protons as evidenced by their similar state 4 respiration rates . One way the ATP9-7 and ATP9-5 proteins could modify the functional properties of the ATP synthase is through modification of the seal between the central stalk of F1 and the c-ring caused by structural differences in the two proteins . One could also speculate that the lower RCR value in MitoATP9-5 compared to MitoATP9-5 is due to lower maximal turnover of ATP synthase under load ( in contrast to the uncoupled state ) , which could be an alternative functional difference between the two forms of the enzyme . Alternatively , c-rings assembled from the c-subunit encoded by Atp9-5 might contain less monomer than those made from the c-subunit encoded by Atp9-7 . As a result , fewer protons would be needed to make one ATP with the former than with the latter . Recent studies have shown that the c-ring stoichiometry is quite variable among species ( from 8 to 15 ) due to structural specificities in the N-terminal α-helix of c-subunit [27–32] . Although it is impossible to predict the c-ring stoichiometry from the primary sequences of the proteins encoded by Atp9-5 and Atp9-7 , it is interesting to note that amino-acid differences exist between them in positions presumed to be critical for the packing of c-subunits ( S5 Fig ) . It will be interesting in the future to determine the efficiency of translation of ATP9-5 and ATP9-7 and the turnover of the two forms of ATP-synthase , which could reveal an additional level of regulation that could explain why two structurally different c-subunit isomers are used in P . anserina . While much remains to be learned , one thing is clear: by invoking different c-subunit isoforms P . anserina demonstrates a degree of sophistication for regulating aerobic energy metabolism that has not been described before . There is a large body of observations in the literature indicating there is a link between longevity and mitochondrial function in many organisms including P . anserina ( for review , see [33] ) . For instance , mutants of this fungus severely defective in complex I , III or IV , display a huge increase in longevity [19–21 , 34] . It has been argued that the induction of AOX that occurs in these mutants was possibly involved by some mechanism in the observed changes in longevity . However , while AOX becomes essential for cellular viability when electrons can no longer be transferred to oxygen by complexes III and IV , the deletion of the AOX gene is tolerated well by complex I mutants and these remained long-lived [19] . Interestingly , overexpressing NDI1 , an alternative mitochondrial NADH dehydrogenase , and AOX partially reversed the longevity phenotypes of complexes I , III and IV mutants [19 , 35 , 36] , indicating that the rate of electron flow is a key factor that influences longevity . According to the ‘mitochondrial free radical theory of aging’ [37] , increasing this flow possibly enhances the production of reactive oxygen species ( ROS ) that can damage any type of biomolecules , owing to a higher diversion of electrons from their normal pathway to oxygen . While supported by many data , this theory was however challenged by controversial results [38] . For instance , no increase in carbonylation of mitochondrial proteins was observed during aging of P . anserina [39] , and eliminating the gene of the mitochondrial localized ROS detoxifying protein PaSod3 does not affect life span [40] . Furthermore , the long-lived complex I mutants of P . anserina do not show any change in production and/or scavenging of free radical species [19] . These observations point to the rate of production of ATP , the end product of respiration , as a possible life span modulator . The results of this study support this hypothesis . Indeed , reducing the rate of ATP production in apical cells by expressing there Atp9-7 instead of Atp9-5 leads to a longer life span while the mitochondrial electron transfer capacity remains unchanged .
All the strains are derived from the S wild type strain of Podospora anserina that harbors two different genes encoding the c-subunit of ATP synthase: Atp9-5 ( Pa_5_9140 ) and Atp9-7 ( Pa_7_20 ) . The short-hand nomenclature of the strains describes what allele is actively transcribed ( 5 and 7 in regular point size ) and the origin of the 5’ and 3’ flanking sequences that control expression of the alleles ( 5 or 7 in superscript just before the allele ) : 55 77 is for the wild type strain . The wild type ( 55 and 77 ) and the inactivated ( Δ5 and Δ7 that do not appear in the name of the strain ) Atp9 alleles have been combined with various Atp9 ectopic transgenes expressed under the control of switched cis-regulatory sequences ( 57 ) and ( 75 ) by successive genetic crosses [7] . The previous nomenclature of the strains was modified and adapted for coherence with this manuscript: 77[57] ( instead of Δ5 75 ) , [75][57] ( Δ5 Δ7 57 75 ) , [75]55 ( Δ7 57 ) , [57] ( Δ7 Δ5 75 ) and 55 ( Δ7 ) . The selection of strains with oligomycin resistant Atp9 alleles ( 55OR and [57OR] ) , the construction of strains with new transgenes ( 55 , 5nat , 7nat ) and their combination by genetic cross to obtain the various strains used in this work are detailed in Supporting Information ( S1 Methods ) . Complete genotypes and nomenclature of all strains are detailed in S1 Table . Vegetative cultures were initiated from small pieces of mycelium freshly obtained from germinated spores . They were grown at 27°C in Petri dishes containing solid M2 minimal media supplemented , when necessary , with antibiotics ( http://podospora . igmors . u-psud . fr/ and Text SI ) . When mycelium had to be collected ( e . g . for protoplast preparation or RNA extraction ) it was grown on plates overlaid with cellophane . For longevity measurements at least 32 independent cultures for each tested genotype ( see S1 Methods , S3 Table ) were grown on solid M2 medium in 30cm-race tubes . Longevities were determined in centimeters of linear growth from the culture initiation until the apical front stops growing or by the estimation of the half-life ( in days ) , which is the number of days by which 50% of the cultures were still alive . For each strain , 20 liquid cultures were initiated with fragmented mycelium as described ( see S1 Methods for details ) . Briefly , the grown mycelium was drained and cell walls were digested for 3 hours with 40 mg/ml Glucanex ( Novozyms ) , after which the protoplasts were collected by filtration and centrifugations . Mitochondria were released from 109 protoplasts by osmotic shock in 0 . 33 M saccharose , 1 mM EGTA , 0 . 2% BSA pH 6 . 8 at 4°C . Mitochondria were isolated by differential centrifugation 4°C in the same buffer , washed , and suspended at 10 mg/ml in the same buffer . The procedure yielded ~1 . 5 mg mitochondrial protein per ~200 g starting material . Protein concentrations were determined using the BIORAD protein assay . Oxygen consumption was measured with a Clark-type O2 electrode ( Hansatech ) at 28°C in a 1 ml-oxytherm thermostated chamber that contained 0 . 65 M mannitol , 0 . 36 mM EGTA , 10 mM Tris/maleate , 5 mM Tris/Phosphate pH 6 . 8 and 0 . 3% BSA . Freshly prepared mitochondria and NADH were added to 0 . 15 mg/ml and 4 mM , respectively . State 3 respiration was initiated with ADP ( 150 μM ) and the reaction was followed until the system converted back to the resting rate of respiration ( state 4 ) , which indicated that all of the ADP had been phosphorylated . The maximal ( uncoupled ) rate of respiration was measured with carbonyl cyanide m-chlorophenyl hydrazone ( CCCP 5μM ) included in the reaction mix . The sensitivity of respiration to SHAM was measured with protoplasts prepared from apical cells ( see S1 Methods ) before or adding KCN . KCN and SHAM were each used at a concentration of 1mM . For measurement of ATP synthesis rates , the chamber buffer contained excess ADP ( 1mM ) . Samples ( 50 ul ) were removed at 15 second intervals and added to a solution of 2 . 3% perchloric acid and 8 . 3 mM EDTA to stop the reaction . Acid-quenched samples were brought to pH 6 . 5 with 0 . 3 M MOPS , 2 N KOH and the amount of ATP was determined using a luciferin/luciferase assay ( ATPlite 1step , PerkinElmer; Turner Scientific ( Reporter ) bioluminometer ) . Published methods [41 , 42] were used for electrophoretic separation of mitochondrial proteins on 12% SDS-polyacrylamide or 5–10% Blue Native ( BN ) gels . For BN-PAGE analysis , mitochondria were solubilized with 2% digitonin , and after the electro-transfer of the proteins onto nitrocellulose membrane ATP synthase was detected using polyclonal antibodies against yeast α-F1 protein ( a gift from J . Velours ) at a 1:10 , 000 dilution . The immunological signals obtained on a X-ray film were scanned as 16-bit tiff images and quantification using ImageJ software [43] . Antibodies against porin and AOX ( provided by H . D . Osiewacz and T . Elthon ) were used after 1:5000 and 1:100 dilutions , respectively . Mycelium cultures were initiated from single pieces of mycelium on solid M2 medium overlaid with a cellophane sheet and grown at 27°C for 1 , 2 or 5 days . Depending on the experiment , the entire mycelium , or only part of it , was collected with a scalpel blade , added to a screw-capped tube , and disrupted with glass beads . Alternatively , 5x107-108 protoplasts acquired from 2 or 5 day-liquid cultures were used . Mycelium and protoplasts were frozen in liquid nitrogen then ground in the Qiagen RLT buffer in a FasPrep apparatus ( speed 6 for 45sec; MP Biomedical ) . Total RNA was extracted using the RNAeasy plant kit ( Qiagen ) . The number of thalli collected was adapted according to the time of culture ( 24 hours-5 days ) in order to obtain 25–50 μg RNA . Quantification of transcripts was conducted in a two-step procedure . Reverse transcription of 2 μg total RNA was performed using the SuperScript II ( Invitrogen ) kit and primed with oligo ( dT ) 15 . cDNA levels were then quantified by real time PCR in the Light Cycler 480 system using couples of primers described in S5 Table and reaction kits containing SYBR Green I ( Roche ) . Data were analyzed using the ‘second derivative maximum’ method for quantification . Amplification efficiency was determined for each couple of primers , based on standard curves established using serial dilutions of one of the cDNA samples . The abundance of Atp9-5 and Atp9-7 transcripts in at least three independent RNA extractions was estimated relative to that of the constitutively expressed Gpd gene ( Pa_3_5410 ) . Results were expressed as mean values relative to the abundance of Atp9-7 in a one day-old culture of the wild type ( w-1d , see Fig 1 ) . It is to be noted that during protoplast preparation living cells are subject to a long , possibly stressful , period ( 3 hours in the presence of glucanex; see S1 Methods for details ) that might affect gene expression . We therefore tested the influence of the procedure by repeating the transcript analyses on w-5d mycelium exposed to these conditions before extracting RNA . The relative abundance of Atp9-5 and Atp9-7 transcripts was essentially unchanged compared to that measured without exposure of the mycelium to these conditions ( S3 Fig ) , which precludes any procedural bias .
|
In mitochondria , the ATP synthase ( also referred to as complex V ) catalyzes the late steps of oxidative phosphorylation ( OXPHOS ) , which is a process that provides aerobic eukaryotes with most of their energy requirements by generating adenosine triphosphate ( ATP ) molecules . While the structure and mechanism of ATP synthase are mostly well established , much remains to be learned about how cells and tissues modulate the production and activity of this enzyme . Herein we report the existence in the filamentous fungus Podospora anserina of a two-pronged energy regulatory mechanism that involves two nuclear genes ( Atp9-5 and Atp9-7 ) that encode structurally different c-subunits of ATP synthase . This system enables a proper production of ATP synthase and optimizes the rate of ATP synthesis in mitochondria along the rather complex life cycle of this fungus .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"protons",
"medicine",
"and",
"health",
"sciences",
"fungal",
"genetics",
"gene",
"regulation",
"rna",
"extraction",
"fungal",
"structure",
"physiological",
"processes",
"mycelium",
"mitochondria",
"bioenergetics",
"cellular",
"structures",
"and",
"organelles",
"extraction",
"techniques",
"oxygen",
"consumption",
"research",
"and",
"analysis",
"methods",
"respiration",
"mycology",
"gene",
"expression",
"nucleons",
"physics",
"biochemistry",
"cell",
"biology",
"nuclear",
"physics",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"energy-producing",
"organelles"
] |
2016
|
Regulation of Aerobic Energy Metabolism in Podospora anserina by Two Paralogous Genes Encoding Structurally Different c-Subunits of ATP Synthase
|
Despite eliciting a potent CD8+ T cell response , Brucella abortus is able to persist and establish a chronic infection inside its host . We have previously reported that the infection of human monocytes/macrophages with B . abortus inhibits the IFN-γ-induced MHC-I cell surface expression down-modulating cytotoxic CD8+ T cell responses . MHC-I down-modulation depends on bacterial viability and results from the capacity of B . abortus to retain the MHC-I molecules within the Golgi apparatus . Furthermore , we recently demonstrated that epidermal growth factor receptor ( EGFR ) pathway is involved in this phenomenon and that this is an early event during infection . However , the components and mechanisms whereby B . abortus is able to down-modulate MHC-I remained to be elucidated . In this study we demonstrated that the down-modulation of MHC-I expression is not mediated by well-known Brucella virulence factors but instead by B . abortus RNA , a PAMP associated to viability ( vita-PAMP ) . Surprisingly , completely degraded RNA was also able to inhibit MHC-I expression to the same extent as intact RNA . Accordingly , B . abortus RNA and its degradation products were able to mimic the MHC-I intracellular retention within the Golgi apparatus observed upon infection . We further demonstrated that TLR8 , a single-stranded RNA and RNA degradation products sensor , was involved in MHC-I inhibition . On the other hand , neutralization of the EGFR reversed the MHC-I inhibition , suggesting a connection between the TLR8 and EGFR pathways . Finally , B . abortus RNA-treated macrophages display diminished capacity of antigen presentation to CD8+ T cells . Overall , our results indicate that the vita-PAMP RNA as well as its degradation products constitute novel virulence factors whereby B . abortus , by a TLR8-dependent mechanism and through the EGFR pathway , inhibits the IFN-γ-induced MHC-I surface expression on human monocytes/macrophages . Thus , bacteria can hide within infected cells and avoid the immunological surveillance of cytotoxic CD8+ T cells .
Host control of brucellosis requires a set of cells and components of the immune system which together promote a complex response against Brucella spp . [1] . Yet , from the many defensive resources that adaptive immunity brings into play , cytotoxic CD8+ T cells are determinant to restrain Brucella infection . The importance of these cells resides in their capacity to eliminate Brucella-infected target cells [2 , 3] . Previous studies in humans , mice and bovines have shown that specific CD8+ T cells are developed during Brucella infection [1 , 4] , confirming the ability of Brucella-infected macrophages to present bacterial antigens on MHC-I molecules and activate cytotoxic CD8+ T cell responses . Despite this immune response , Brucella is able to persist inside these cells establishing a chronic infection . Therefore , as a successful chronic and persistent pathogen , Brucella must own an effective strategy to subvert the challenge of highly outfitted CD8+ T cells . We have previously reported that infection of human monocytes/macrophages with B . abortus inhibits the IFN-γ-induced MHC-I cell surface expression . As a consequence , B . abortus-infected macrophages display diminished capacity of antigen presentation to CD8+ T cells [5] . MHC-I down-modulation results from the capacity of B . abortus to induce the retention of these molecules within the Golgi apparatus [5] . However , the components of B . abortus involved in this phenomenon remained unknown . Interestingly , B . abortus-mediated MHC-I down-modulation is dependent on bacterial viability as was demonstrated by the incapacity of heat-killed bacteria to inhibit the expression of such molecules [5] . Furthermore , we have recently reported that two B . abortus mutant strains devoid of key virulence factors , B . abortus RB51 ( a rough LPS mutant ) and B . abortus virB10- ( a VirB type IV secretion system mutant ) , are capable of inhibiting the IFN-γ-induced MHC-I surface expression to the same extent as wild-type B . abortus [6] . These B . abortus mutant strains are unable to persist inside human monocytes for a long period despite their preserved capacity of infecting cells [7–9] . Consistent with this , we observed that the phenomenon of MHC-I inhibition is triggered at early time points and can be observed at 8 h post-infection . At 24 h and 48 h post-infection it became even stronger [6] . Overall these results led us to think that the components involved in the inhibition of IFN-γ-induced MHC-I surface expression should be associated with bacterial viability . In turn , our results with the mutant strains gave us the idea that these bacterial components should be expressed early during infection . It has been recently demonstrated that the immune system is capable of sensing the most essential characteristic of microbial infectivity , microbial viability itself [10] . The immune system can thus detect pathogen-associated molecular patterns ( PAMPs ) which are present in live bacteria but rapidly eliminated when bacteria lose their viability [10] . These PAMPs are lost since they are intimately linked to the metabolic activity and replicative capacity of the microorganisms . In order to differentiate them from traditional PAMPs , structural components that are preserved after loss of bacterial viability ( such as LPS , lipoproteins and DNA , among others ) , this special class of PAMPs were named viability-associated PAMPs ( vita-PAMPs ) , among which prokaryotic RNA is included [10 , 11] . Recognition of nucleic acids in general and RNA in particular by receptors of the innate immune system is a complicated and interesting field of investigation . The immune system must discriminate between ‘self’ ( host ) and ‘foreign’ ( invading pathogen ) nucleic acids [12] . This principle is based on three criteria: the availability of nucleic acid ligands , the localization of such nucleic acids and their structure ( characterized by sequence motifs , conformation and chemical modification ) . In most cases , a combination of these aspects contributes to the proper recognition of foreign nucleic acids and the induction of adequate immune responses [12] . Most of the receptors involved in the immune sensing of nucleic acids have been identified . Among them , the TLRs located in endosomes/phagolysosomes are the most studied: TLR9 senses CpG DNA motifs; TLR3 and TLR7 are capable of recognizing double-stranded and single-stranded RNA respectively and TLR8 is not only able to recognize single-stranded RNA but it has been recently described as a RNA degradation products sensor as well [13 , 14] . Taking our previous results into account , we hypothesized that the components of B . abortus involved in the inhibition of MHC-I could be vita-PAMPs such as B . abortus RNA , since they are found exclusively in live bacteria and are actively expressed during early stages of infection . Thus , the aim of this study was to characterize the components , signaling pathways and mechanisms implicated in MHC-I down-modulation . Overall , our results indicate that the vita-PAMP RNA as well as its degradation products constitute novel virulence factors whereby B . abortus , by a TLR8-dependent mechanism and through the EGFR pathway , inhibits the IFN-γ-induced MHC-I surface expression on human monocytes/macrophages .
Our previous results had demonstrated that B . abortus-mediated MHC-I inhibition is dependent on bacterial viability [5] . On the other hand , we have recently reported that B . abortus rough LPS mutant RB51 and a mutant in the B . abortus type IV secretion system VirB , two mutant strains in key virulence factors , are capable of inhibiting the IFN-γ-induced MHC-I surface expression to the same extent as wild-type B . abortus [6] . These results led us to think that human monocytes/macrophages could be able to detect a component associated with bacterial viability independently of its virulence factors . In order to address this hypothesis , we used different B . abortus mutant strains on key virulence factors and evaluated whether their live and heat-killed ( HK ) forms could inhibit the IFN-γ-induced MHC-I surface expression on THP-1 cells . The mutant strains used were: RB51 ( rough LPS mutant ) , virB10- ( mutant in VirB type IV secretion system ) , btpA , btpB single mutants and a btpAbtpB double mutant ( mutants of TIR domain-containing proteins which interfere with TLRs signaling pathways ) , and Bpe159 ( mutant in B . abortus putative effector protein BPE159 , which is secreted into the host cytosol independently of the VirB secretion system [15] ) . Confirming and extending our previous results , B . abortus S2308 ( wild type , WT ) and all mutant strains studied were able to diminish the IFN-γ-induced MHC-I surface expression in a dose-dependent manner after 48 h . However , this phenomenon occurred exclusively when bacteria were alive ( Fig 1A–1E and S1 Fig , Panels i and iii ) . The heat-killed forms of these bacteria lost the capacity of inhibiting MHC-I , even at the highest concentration used ( 1 x 109 bacteria/ml ) ( Fig 1A–1E and S1 Fig , Panels ii and iv ) . These results confirm that the inhibition of MHC-I surface expression is dependent on B . abortus viability but independent of the studied virulence factors . In addition , these results suggest that MHC-I inhibition is not caused by B . abortus structural components , which are conserved in heat-killed bacteria . To corroborate our results , we next studied the effect of different structural components of B . abortus on MHC-I surface expression , such as: B . abortus lipopolysaccharide ( Ba LPS ) ; its outer membrane protein 19 ( Omp19 ) , a prototypical lipoprotein of B . abortus , on its lipidated ( L-Omp19 ) and unlipidated ( U-Omp19 ) versions and B . abortus DNA ( Ba DNA ) . None of the evaluated structural components was able to inhibit the IFN-γ-induced MHC-I surface expression ( Fig 2A and 2B ) . Overall , these results confirm that B . abortus-mediated inhibition of MHC-I surface expression requires bacterial viability regardless of the presence of more specialized factors that regulate microbial virulence . In addition , they show that the bacterial component involved in this phenomenon seems to be associated with bacterial viability . Prokaryotic RNA has been recently characterized as a special class of viability-associated PAMP ( vita-PAMP ) , as it is present only in viable bacteria [10] . To investigate whether B . abortus RNA was the component involved in the inhibition of MHC-I surface expression on human monocytes , we used B . abortus WT RNA to stimulate THP-1 cells in the presence of IFN-γ for 48 h at different concentrations . The expression of MHC-I was then evaluated by flow cytometry . B . abortus WT RNA significantly down-regulated the IFN-γ-induced surface expression of MHC-I molecules in a dose-dependent manner ( Fig 3A ) , mimicking what was previously observed with viable B . abortus . Given that certain phenol traces could still be present in the purified RNA , we performed the RNA extraction in the absence of bacteria and used it as control ( TRIzol bar ) . This treatment was not able to down-modulate MHC-I ( Fig 3A ) . Moreover , RNA purified with a method other than TRIzol was equally able to inhibit MHC-I expression on THP-1 cells ( S2 Fig ) . In turn , RNA purified from the mutant strains RB51 and virB10 was also able to inhibit the IFN-γ-induced MHC-I surface expression on THP-1 cells to the same extent as B . abortus WT RNA ( Fig 3B and 3C ) . This inhibition was not due to a loss of cell viability in B . abortus RNA stimulated cultures , since MHC-I inhibition was observed gating only on viable cells ( 7-AAD negative cells ) . Furthermore , B . abortus RNA treatment did not induce early and late apoptosis or necrosis as determined by the Annexin V assay , even at the highest evaluated concentration ( 10 μg/ml ) ( Fig 3D ) . On the contrary , high levels of early and late apoptosis or necrosis were obtained on cells treated with the positive control paraformaldehyde ( PFA ) . In agreement with the inability of heat-killed B . abortus ( HKBA ) to inhibit MHC-I surface expression , we determined by gel electrophoresis that HKBA lacked RNA and that the products recovered from the HKBA RNA extraction process were unable to inhibit MHC-I surface expression ( S3 Fig ) . Moreover , B . abortus RNA was able to complement the absence of this molecule in HKBA , making it capable of down-modulating MHC-I expression on human monocytes ( S4 Fig ) . In another set of experiments , THP-1 cells were stimulated with B . abortus RNA alone or introduced into the cell by transfection with lipofectamine reagent . Stimulation with B . abortus RNA , independently of the procedure , was able to inhibit MHC-I expression suggesting that RNA without transfection could gain access to endosomal sensors ( 3E and F ) . Confirming these results , the endocytosis inhibitor Nystatin was able to reverse MHC-I inhibition mediated by stimulation with B . abortus RNA ( S5 Fig ) . On the other hand , other prokaryotic RNAs ( from Bacillus cereus , Salmonella typhimurium , Escherichia coli and Klebsiella pneumonia ) were able to inhibit MHC-I surface expression . On the contrary , eukaryotic RNA ( from peripheral blood mononuclear cells , PBMCs ) was unable to inhibit MHC-I surface expression , even at the highest concentration utilized ( S6 Fig ) . Overall , these results indicate that RNA is a component associated with bacterial viability which is employed by B . abortus to inhibit the IFN-γ-induced surface expression of MHC-I molecules on human monocytes . More importantly , this is not an exclusive phenomenon of B . abortus RNA as it could be extended to other prokaryotic although not to eukaryotic RNAs . As traces of DNA and proteins could contaminate the RNA fractions , we decided to further purify our preparations of B . abortus RNA by eliminating either residual DNA or proteins . For this , B . abortus RNA fractions were digested with a DNase or a proteinase ( Proteinase K; PK ) . After that , we verified that the treatments with the enzymes had not affected the integrity of the RNA ( Fig 4A , lane 3 and 4 ) . The products of such digestions were then employed to stimulate THP-1 cells in the presence of IFN-γ for 48 h . Then , the expression of MHC-I molecules was evaluated by flow cytometry . The preparations of DNase- and PK-digested RNA were still able to inhibit MHC-I expression in the same manner as intact RNA , indicating that contaminating DNA and proteins do not mediate the phenomenon of MHC-I inhibition ( Fig 4B and 4C ) . B . abortus RNA was next digested with a prokaryotic RNA-specific RNase and this product was employed to stimulate THP-1 cells in the presence of IFN-γ for 48 h . The RNase used was RNase I from Escherichia coli which degrades single-stranded RNA in a mixture of mono- , di- , and tri-nucleotides . RNA preparations digested with RNase I completely lost the integrity of the RNA ( Fig 4A , lane 5 ) . Surprisingly , products from RNase I-digested RNA were still able to inhibit the IFN-γ-induced MHC-I surface expression to the same extent as intact RNA ( Fig 4D and 4E ) . MHC-I down-modulation was not due to the presence of the RNase in the culture , since the negative control with merely RNase I was unable to reproduce the phenomenon . Overall , these results indicate that B . abortus RNA and its degradation products are the components involved in the inhibition of IFN-γ-induced MHC-I surface expression . While THP-1 cells are a good model of human monocytes , we next evaluated whether our results could be extended to primary cultures of monocytes/macrophages . For this purpose , peripheral blood-isolated human monocytes or murine bone marrow-derived macrophages ( BMM ) were stimulated with different concentrations of B . abortus RNA and then the expression of MHC-I molecules was evaluated by flow cytometry . B . abortus RNA was significantly able to inhibit MHC-I expression in both human primary monocytes and murine BMM in a dose-dependent manner ( Fig 5A and 5B ) . Thus , B . abortus RNA does not only inhibit MHC-I expression on THP-1 cells but also on human primary monocytes and murine macrophages . The most known receptors capable of detecting RNA are TLRs located in endosomes/phagolysosomes . Among them , TLR3 is capable of recognizing double-stranded RNA , TLR7 is capable of recognizing single-stranded RNA and TLR8 is also able to recognize single-stranded RNA and recently it was described as capable of recognizing RNA degradation products [13 , 14] . Since the capacity of Brucella RNA to form secondary structures is still unknown and considering that TLR3 has been implicated in many functions mediated by viral double-stranded RNAs , we first wanted to evaluate whether TLR3 could be involved in B . abortus RNA-mediated inhibition of MHC-I molecules . TLR3 is the unique TLR that transduces its signal through the adapter protein TRIF . We therefore evaluated the effect of B . abortus RNA in BMM from TRIF KO mice . B . abortus RNA was able to inhibit the IFN-γ-induced MHC-I surface expression in BMM from TRIF KO mice to the same extent as in BMM from WT mice ( Fig 6A ) . To confirm the fact that TLR3 was not involved in the inhibition mediated by B . abortus RNA we used a TLR3 inhibitor ( TLR3/dsRNA Complex Inhibitor ) . Yet in the presence of a TLR3 inhibitor , B . abortus RNA down-regulated MHC-I expression confirming that TLR3 is not involved in this phenomenon ( Fig 6B ) . Having discarded the participation of TLR3 , we focused our attention on TLR7 and TLR8 . Specific agonists have been described for either TLR7 or TLR8 , or both . We used the human TLR7 ( hTLR7 ) agonist Gardiquimod , the human TLR7/8 ( hTLR7/8 ) agonist Resiquimod ( R848 ) and the human TLR8 ( hTLR8 ) agonists ssRNA40/LyoVec and ORN06/LyoVec . Gardiquimod was unable to inhibit the IFN-γ-induced MHC-I surface expression on THP-1 cells ( Fig 6C ) . However , R848 was able to mimic the inhibition of MHC-I expression mediated by B . abortus RNA ( Fig 6D ) . These results allowed us to discard hTLR7 and postulate hTLR8 as a possible receptor . To corroborate this , THP-1 cells were stimulated with the hTLR8 agonists ORN06 and ssRNA40 . As shown in Fig 6E and 6F , both ORN06/LyoVec and ssRNA40/LyoVec were able to mimic the effect of B . abortus RNA on MHC-I surface expression . Although TLR8 is not functional in mice [16] , it has been demonstrated that TLR7 performs its function [17 , 18] . Since ssRNA40 is not only an agonist of hTLR8 but also of murine TLR7 ( mTLR7 ) [17] , we evaluated its effect on BMM . Corroborating our results , ssRNA40 was able to inhibit the expression of MHC-I in BMM ( Fig 6G ) . To confirm these results BMM from WT or TLR7 KO mice were infected with B . abortus or stimulated with B . abortus RNA . Our results showed that the inhibition of MHC-I surface expression mediated by B . abortus and B . abortus RNA was abolished in BMM from TLR7 KO mice ( Fig 6H and 6I ) . Altogether these results demonstrate that the MHC-I inhibition by B . abortus and its RNA is mediated by hTLR8/mTLR7 . Recently , we have demonstrated that the EGFR pathway is involved in the inhibition of MHC-I surface expression mediated by B . abortus infection [6] . In order to extend this finding and taking into account that TLR8 is the receptor involved in the phenomenon of B . abortus RNA-mediated MHC-I inhibition on human monocytes , we decided to evaluate the connection between TLR8 and EGFR signaling pathways . For this , THP-1 cells were stimulated with the hTLR8 agonist ORN06/LyoVec in the presence of an EGFR ligand-blocking antibody ( Cetuximab ) . Neutralization of the EGFR significantly reversed the inhibition of MHC-I surface expression mediated by ORN06 ( Fig 6J ) . This reversion was not due to a dysfunction of TLR8 caused by Cetuximab , as the secretion of pro-inflammatory cytokines downstream of NF-κB was unchanged between Isotype control and Cetuximab-treated cells ( S7 Fig ) . Overall , our results indicate that B . abortus RNA inhibits the IFN-γ-induced MHC-I surface expression on human monocytes/macrophages by a TLR8-dependent mechanism and through the EGFR pathway . We have previously demonstrated that B . abortus infection induces the intracellular retention of MHC-I molecules within the Golgi apparatus [5] . Thus , we evaluated whether B . abortus RNA was able to mimic this phenomenon . For this , the localization of MHC-I molecules was determined by confocal microscopy in cells infected with B . abortus or stimulated with B . abortus RNA in the presence of IFN-γ for 48 h . MHC-I expression was determined with an anti- HLA-ABC monoclonal antibody followed by Alexa 546-labelled secondary antibody . At 48 h of culture , cells treated only with IFN-γ showed MHC-I expression confined predominantly to the cellular membrane ( Fig 7A ) . On the contrary , both B . abortus-infected monocytes as well as monocytes treated with B . abortus RNA , showed MHC-I expression restricted to the cellular interior concomitantly with a marked decrease of MHC-I surface expression ( Fig 7A and 7B ) . Next , we examined the subcellular localization of retained MHC-I molecules . For this , THP-1 cells were treated with B . abortus RNA in the presence of IFN-γ for 48 h . MHC-I was detected as described previously and the subcellular compartments were detected with specific primary mAbs followed by Alexa 488-labelled secondary Ab . No colocalization was detected with either the early endosome marker EEA1 , the lysosome marker LAMP-2 or the reticulum endoplasmic marker calnexin ( Fig 7C ) . On the contrary , in 50% of the B . abortus RNA-treated monocytes that retained MHC-I , these molecules colocalized with the Golgi apparatus marker GM130 ( Fig 7C and 7D ) . Taking into account that RNA degradation products were also able to inhibit the MHC-I surface expression , we next evaluated their capacity to induce intracellular retention of these molecules . Completely degraded RNA was able to induce the intracellular retention of MHC-I molecules within the Golgi apparatus to the same extent as intact RNA ( Fig 8 ) . Altogether these results demonstrate that B . abortus RNA and its degradation products mimic the intracellular retention of MHC-I within the Golgi apparatus observed with B . abortus infection . Finally , we evaluated whether the diminished MHC-I surface expression induced by B . abortus RNA was associated with changes in Ag presentation to MHC-I-restricted CD8+ cytotoxic T cells . For this , murine BMM from WT and TLR7 KO mice were treated with different doses of B . abortus RNA in the presence of murine IFN-γ ( mIFN-γ ) for 48 h , followed by incubation with OVA peptide ( SIINFEKL ) and a B3Z T-cell hybridoma specific for OVA-Kb , which carries a β-galactosidase construct driven by NF-AT elements from the IL-2 promoter . BMM from WT and TLR7 KO mice treated solely with mIFN-γ presented Kb-restricted OVA peptide efficiently after 6 h and onwards ( Fig 9A and 9B ) , as evidenced by the ability of these cells to induce LacZ activity in B3Z cells . Treatment of BMM from WT mice with B . abortus RNA ( 1–10 μg/ml ) in the presence of mIFN-γ significantly inhibited presentation of OVA peptide since it diminished the response of B3Z cells , compared to mIFN-γ-only treated cells ( Fig 9A ) . However , treatment of BMM from TLR7 KO mice with B . abortus RNA did not affect antigen presentation to CD8+ T lymphocytes compared to mIFN-γ-only treated cells ( Fig 9B ) . Taken together , these results indicate that inhibition of MHC-I expression by B . abortus RNA correlates with diminished Ag presentation to MHC-I-restricted CD8+ cytotoxic T cells . In addition , our results demonstrate that inhibition of Ag presentation to CD8+ T cells by B . abortus RNA is mediated by mTLR7 .
Human monocytes were isolated exclusively from healthy adult blood donors in accordance with the guidelines of the Ethical Committee of the IMEX Institute . All adult blood donors provided their informed written consent prior to the study . Mouse bone marrow-derived macrophages ( BMM ) were generated by differentiation of bone marrow progenitors from female C57BL/6 mice ( aged 2–3 months ) . Mice were kept under specific pathogen-free conditions in positive-pressure cabinets and provided with sterile food and water ad libitum . All animal procedures were performed according to the rules and standards for the use of laboratory animals of the National Institutes of Health . Animal experiments were approved by the Animal Care and Use Committee of the IMEX Institute . The protocol license number assigned by this Committee is: 020/2016 . B . abortus S2308 , Salmonella typhimurium SL14028 , Bacillus cereus B10502 , Escherichia coli 11105 and Klebsiella pneumoniae 700603 , and B . abortus RB51 , B . abortus virB10 , B . abortus btpA , B . abortus btpB , B . abortus btpAbtpB or B . abortus Bpe159 mutant strains were cultured in tryptose-soy agar supplemented with yeast extract ( Merck ) . The number of bacteria on stationary-phase cultures was determined by comparing the OD at 600 nm with a standard curve . To obtain heat-killed B . abortus strains , bacteria were washed in PBS , heat killed at 70°C for 20 min and stored at -70°C until used . Total absence of B . abortus viability subsequent to heat killing was verified by the absence of bacterial growth in tryptose-soy agar . All live Brucella manipulations were performed in biosafety level 3 facilities , located at the Instituto de Investigaciones Biomédicas en Retrovirus y SIDA ( Buenos Aires , Argentina ) . Lipoproteins were expressed and purified as previously described [19] . To eliminate LPS contamination , recombinant Omps were adsorbed with Sepharose-polymyxin B ( Sigma-Aldrich ) . Both proteins contained less than 0 . 25 endotoxin U/μg of protein as assessed by Limulus Amebocyte Lysate assay ( Lonza ) . The protein concentration was determined by the BCA protein assay ( Pierce ) using bovine serum albumin as standard . The purified proteins were aliquoted and stored at -70°C until used . B . abortus 2308 LPS was provided by I . Moriyón ( University of Navarra , Pamplona , Spain ) . The purity and characteristics of these preparations have been described elsewhere [20] . LPS was solubilized in water by sonication at the appropriate concentration and autoclaved before use . B . abortus DNA was purified by extraction with phenol:chloroform:isoamyl alcohol and ethanol precipitation [21] . To eliminate LPS contamination , DNA was adsorbed with Sepharose-polymyxin B ( Sigma-Aldrich ) . DNA contained less than 0 . 25 endotoxin U/μg of protein as assessed by Limulus Amebocyte Lysate assay ( Lonza ) . All experiments were performed at 37°C in 5% CO2 atmosphere and standard medium composed of RPMI-1640 supplemented with 25 mM Hepes , 2 mM L-glutamine , 10% heat-inactivated fetal bovine serum ( Gibco ) , 100 U of penicillin/ml and 100 μg of streptomycin/ml . THP-1 cells were obtained from the American Type Culture Collection ( Manassas , VA ) and cultured as previously described [19] . To induce maturation , cells were cultured in 0 . 05 μM 1 , 25-dihydroxyvitamin D3 ( EMD Millipore ) for 72 h . Peripheral blood mononuclear cells ( PBMCs ) were obtained by Ficoll-Hypaque ( GE Healthcare ) gradient centrifugation from human blood collected from healthy adult individuals . Monocytes were then purified from PBMCs by Percoll ( GE Healthcare ) gradient and resuspended in standard medium . Purity of the isolated CD14+ monocytes was more than 80% as determined by flow cytometry . Viability of cells was more than 95% in all the experiments as measured by trypan blue exclusion test . Mouse bone marrow-derived macrophages ( BMM ) were generated by differentiation of bone marrow progenitors from C57BL/6 wild type mice , TRIF KO or TLR7 KO mice ( provided by Federal University of Minas Gerais , Belo Horizonte , Brazil ) with rM-CSF ( PeproTech ) , as previously described [22] . THP-1 cells at a concentration of 0 . 5 x 106/ml were infected in round-bottom polypropylene tubes ( Falcon ) with different multiplicities of infection ( MOI ) of B . abortus S2308 , B . abortus RB51 , B . abortus virB10 , B . abortus btpA , B . abortus btpB , B . abortus btpAbtpB or B . abortus Bpe159 mutants . All infections were done in the presence of 150 U/ml IFN-γ ( Endogen ) for 2 h in standard medium containing no antibiotics . In another set of experiments , BMM from WT or TLR7 KO mice at a concentration of 0 . 5 x 106/ml were infected in a 24-well plate with different MOI of B . abortus S2308 . Infections were done in the presence of 10 ng/ml mIFN-γ ( Peprotech ) for 2 h in standard medium containing no antibiotics . In all cases , cells were extensively washed to remove uninternalized bacteria and infected cells were maintained in culture in the presence of IFN-γ or mIFN-γ , 100 μg/ml gentamicin and 50 μg/ml streptomycin for an additional 48 h . For viability assay , THP-1 cells at a concentration of 0 . 5 x 106/ml were treated with different doses of B . abortus RNA in the presence of IFN-γ for 48 h . THP-1 cells treated with 2% paraformaldehyde ( PFA ) were also included as positive control . After 48 h , cells were washed and incubated with Annexin V-FITC and Propidium Iodide ( BD Biosciences ) for 10 min on ice in darkness . Then , cells were evaluated in the quadrants of Annexin V+/PI- ( early apoptosis ) , Annexin V+/PI+ ( late apoptosis ) and Annexin V-/PI+ ( necrosis ) . After labelling , cells were analyzed on a FACSCalibur flow cytometer ( BD Biosciences ) and data were processed using CellQuest software ( BD Biosciences ) . 5–10 x 106 PBMCs or 5 x 108 CFU were resuspended in 1ml of TRIzol Reagent ( Invitrogen ) and total RNA was extracted according to the manufacturer’s instructions . OD at 260 was measured for RNA quantification . In another set of experiments , B . abortus RNA was purified with Quick-RNA MiniPrep ( Zymo Research ) according to the manufacturer’s instructions . The purity of B . abortus RNA was assessed using a DeNovix DS-11 Spectrophotometer ( DeNovix Inc . ) with a ratio of absorbance 260/280 > 2 . 0 and a ratio of absorbance 260/230 > 1 . 8 . In one set of experiments , RNA was treated with DNase RQ1 ( Promega ) , Proteinase K ( PK ) ( Promega ) or E . coli RNase I ( Life Technologies ) prior to cell stimulation . RNA preparations ( B . abortus RNA , DNase-treated B . abortus RNA , PK-treated B . abortus RNA and RNase I-treated B . abortus RNA ) were further visualized by 1% agarose gel electrophoresis . The RNA was detected using UV light and the size of the RNA was determined using standard 100 bp Plus DNA ladder ( TransGen Biotech Co . , Ltd . ) . Cells at 0 . 5 x 106/ml were stimulated with B . abortus RNA , other prokaryotic or eukaryotic RNAs , DNase-treated B . abortus RNA , PK-treated B . abortus RNA , RNase I-treated B . abortus RNA , HK B . abortus strains , different structural components of B . abortus ( LPS , DNA and lipoproteins ) or TLR ligands in the presence of 150 U/ml IFN-γ for 48 h in standard medium containing antibiotics . In another set of experiments , THP-1 cells were treated with B . abortus RNA complexed with Lipofectamine 2000 ( Invitrogen ) . Briefly , Lipofectamine was mixed with bacterial RNA ( 1:3 ratio ) in 100 μL/well serum-free RPMI and incubated for 20 min at room temperature . Then , complexes were added to the cells in the presence of 1 . 25% FBS and cell cultures were incubated for 48 h at 37°C in a 5% CO2 atmosphere . In all cases , MHC-I expression was evaluated by flow cytometry . After B . abortus infection or stimulation; THP-1 cells or human primary monocytes were stained with FITC-labelled anti-human HLA-ABC ( clone G46-2 . 6; BD Pharmingen ) or isotype-matched control mAbs . In the experiments with murine macrophages , BMM were infected with B . abortus , or treated with B . abortus RNA or TLR ligands in the presence of 10 ng/ml recombinant murine IFN-γ ( PeproTech ) for 48 h . To determine MHC-I surface expression , cells were stained with PE- or FITC-labelled anti-mouse H-2Kd/H-2Dd ( clone 34-1-2S; BioLegend ) . In all cases , cells were washed and incubated with 7-Amino-Actimycin D ( 7-AAD; BD Biosciences ) for 10 min on ice in darkness . MHC-I expression was evaluated gating on viable cells ( 7-AAD negative cells ) . After labelling , cells were analyzed on a FACSCalibur flow cytometer ( BD Biosciences ) and data were processed using CellQuest software ( BD Bioscience ) or FlowJo 7 . 6 software . THP-1 cells were incubated in chambers-slides ( 2 x 105 cells/well ) with 10 ng/ml PMA ( Sigma-Aldrich ) for 24 h to promote adherence . Then , cells were infected with B . abortus or stimulated with B . abortus RNA or RNase I-treated B . abortus RNA in the presence of IFN-γ for 48 h , fixed with 2% paraformaldehyde , permeabilized with 0 . 1% saponin and incubated with anti-HLA-ABC class I mAb W6/32 , ( purified from murine hybridoma culture supernatants ) and Alexa 546-labelled secondary Ab ( Invitrogen ) . Subcellular compartments were detected using mAbs specific for EEA1 ( early endosomes ) , LAMP-2 ( late endosomes/lysosomes ) , GM130 ( Golgi ) and calnexin ( ER ) ( BD Biosciences ) following Alexa 488-labelled secondary Ab ( Invitrogen ) . Slides were mounted with PolyMount ( Polysciences ) and analyzed using FV-1000 confocal microscope with an oil-immersion Plan Apochromatic 60X NA1 . 42 objective ( Olympus ) . Presentation of OVA epitope 257–264 on Kb ( SIINFEKL ) was detected using the T cell hybridoma B3Z , which carries a β-galactosidase construct driven by NF-AT elements from the IL-2 promoter [23] . For Ag presentation assays , B . abortus RNA-treated BMM from WT or TLR7 KO mice were exposed to 20 ng/ml of the SIINFEKL epitope during 20 min at 37°C . Then cells were washed , suspended in complete medium , and cultured in the presence of the T cell hybridoma B3Z . After 0 , 4 , 6 and 18 h of culture , cells were washed with PBS , and the cross-presentation was evaluated by a colorimetric assay using o-nitrophenyl-p-D-galactoside ( ONPG ) ( Sigma-Aldrich ) as substrate to detect the LacZ activity in B3Z lysates . Antibody targeting EGFR ( Cetuximab ) was purchased from Merck Serono . Gardiquimod , R848 ( Resiquimod ) , ssRNA40/LyoVec and ORN06/LyoVec were purchased from InvivoGen . The 257–264 OVA peptide ( SIINFEKL ) was provided by Dr . S . Amigorena ( Institut Curie , Paris , France ) . TLR3/dsRNA Complex Inhibitor was purchased from Calbiochem . Human TNF-α and IL-1β were measured in culture supernatants by sandwich ELISA , using paired cytokine-specific mAbs according to the manufacturer’s instructions ( BD Pharmingen ) . Results were analyzed with one-way ANOVA followed by post hoc Tukey test using the GraphPad Prism software .
B . abortus is an intracellular pathogen capable of surviving inside macrophages [24] . The success of B . abortus as a chronic pathogen relies on its ability to orchestrate different strategies to evade the adaptive CD8+ T cells responses that it elicits . Previously , we have demonstrated that B . abortus infection inhibits the IFN-γ-induced MHC-I surface expression on human monocytes down-modulating cytotoxic CD8+ T cell responses [5] . Moreover , we have recently deepened into various aspects of this event , such as its kinetics and the participation of the EGFR pathway [6] . Two striking features of the phenomenon of MHC-I inhibition allowed us to shed light on the B . abortus components involved . First of all , heat-killed B . abortus is incapable of inhibiting MHC-I expression [5] . Secondly , the phenomenon is triggered early during infection [6] . Together , these results indicated that only metabolically active viable bacteria can inhibit MHC-I expression and that it must occur during the time span before the bacteria are removed and/or mediated by a product generated early in response to infection . In this study , we could corroborate that what Brucella employs to inhibit MHC-I on monocytes/macrophages is a component associated with bacterial viability itself regardless of the most relevant bacterial virulence factors . Particularly , we elucidated that this component is B . abortus RNA . Moreover , our experiments demonstrated that not only wild-type B . abortus RNA but also the RNA of two mutants strains , RB51 and virB10 , were equally able to inhibit MHC-I . These results together with those shown in Fig 1 and S1 Fig corroborate that what the cells sense is a general determinant of bacterial viability different from its virulence factors . Supporting these results , we demonstrated that the inability of heat-killed B . abortus to inhibit MHC-I surface expression is due to the absence of RNA in these bacterial preparations . In line with this evidence , it is widely known that live vaccines trigger more vigorous immune responses than their killed counterparts , even when live microorganisms are attenuated by elimination of their virulence factors [25] . Since structural bacterial components are present in both live and dead microorganisms , this suggested that there should be non-characterized determinants in live bacteria important for the induction of an effective protective immune response . In this sense , it has been demonstrated that macrophages can directly sense microbial viability through detection of prokaryotic messenger RNA ( mRNA ) , a vita-PAMP present only in viable bacteria , triggering a unique innate and a robust adaptive antibody responses[10] . Notably , the innate response evoked by viability and prokaryotic mRNA was thus far considered to be reserved for pathogenic bacteria , but Sander et al in their study show that even non-pathogenic bacteria in sterile tissues can trigger similar responses , provided they are alive [10] . Furthermore , our results also demonstrated that inhibition of MHC-I is not restricted to B . abortus as it could be extended to other prokaryotic RNAs , suggesting the broad implications of this immune regulation in the context of other infectious processes . One issue that merits discussion is how , during B . abortus infection , the human monocyte/macrophages are able to sense bacterial RNA . This question can be answered if we review the biology of B . abortus and the localization of receptors capable of detecting RNA . It has been described that viral and bacterial RNA are sensed by pattern recognition receptors ( PRRs ) , among which the TLRs family has gained more attention [26 , 27] . TLR3 , TLR7 and TLR8 are the ones preferentially expressed in intracellular vesicles of the endoplasmic reticulum ( ER ) , endosomes , and lysosomes [28] . With respect to the intracellular cycle of the bacterium , B . abortus is able to enter , survive and replicate within vacuolar phagocytic compartments of macrophages [29] . Once inside the macrophages , Brucella dwells in an acidified compartment that fuses with components of the early and late endosomal/lysosomes pathway [7 , 30] . There , the vast majority of the ingested bacteria are rapidly killed . However , the establishment of a persistent infection depends on the ability of the bacterium to form a Brucella-containing vacuole ( BCV ) , which traffics from the endocytic compartment to the endoplasmic reticulum ( ER ) [7 , 29 , 31] . Once inside the replicative BCV , bacteria are resistant to further attack and begin to multiply dramatically [7 , 31] . Starr et al demonstrated that Brucella replication in the ER is followed by BCV conversion into multi-membrane LAMP-1-positive vacuoles with autophagic features ( aBCV ) . Furthermore , aBCVs were required to complete the intracellular Brucella lifecycle and for cell-to-cell spreading [32] . In this context , it is possible that while B . abortus traffics through early and late endosomes/lysosomes the bacterial RNA released during phagocytosis activate endolysosomal TLRs . On the other hand , B . abortus mutant strains on virulence factors are also capable of infecting human monocytes/macrophages and transiting by early and late endosomes/lysosomes , but , unlike wild-type B . abortus , they are unable to replicate in BCVs and thus persist into the cell host . However , our results demonstrated that these strains are equally able to down-regulate MHC-I than wild-type B . abortus . As a consequence , the RNA of these bacteria could also gain access to TLR3 , TLR7 or TLR8 in their transit through endosomes and lysosomes , although they do not persist in macrophages . In accordance with this , it was reported that human TLR8 is activated upon recognition of Borrelia burgdorferi RNA in the phagosome of human monocytes [33] . Furthermore , in line with endosomal TLRs sensing in an infectious context , we have previously demonstrated that MHC-I inhibition is an early event during infection , already observed at 8 h post-infection [6] . This goes along with the time that elapses in the passage of the bacteria by the early and late endocytic/lysosomal pathway [30] . In our in vitro experiments of stimulation with purified RNA , either in the presence or the absence of transfection , the entry of RNA by endocytosis gaining access to the endosomal TLRs can perfectly mimic what happens in an infectious context . Regarding the receptor involved , our results indicated that an hTLR7/8 agonist such as R848 was able to mimic the MHC-I inhibition mediated by B . abortus RNA . However , an hTLR7 agonist per se was unable to reproduce MHC-I inhibition . This led us to propose hTLR8 as a possible receptor , which was corroborated by the specific human TLR8 agonists ssRNA40 and ORN06 . The greater efficiency of the synthetic oligonucleotide ORN06 in MHC-I inhibition may be due to the fact that it contains 6 repeats of the UUGU sequence motif , identified as the minimal motif responsible for ssRNA40 immunoactivity [34] . In addition , the involvement of hTLR8 in the B . abortus RNA-mediated MHC-I inhibition was corroborated in murine BMM using the agonist ssRNA40 which is also specific for murine TLR7 , since TLR7 acts as the human TLR8 in mice [17 , 18] . Moreover , our results with BMM from TLR7 KO mice confirm that the inhibition of MHC-I surface expression by B . abortus and its RNA is mediated by hTLR8/mTLR7 . In agreement with these results , the inhibition of antigen presentation to CD8+ T cells by B . abortus RNA was abolished in BMM from TLR7 KO mice . Further experiments in vivo are needed to determine the involvement of TLR8/TLR7 in the cytotoxic CD8+ T cell responses and chronicity of B . abortus-infected mice . Single-stranded RNA has been identified as the natural ligand of TLR7 and TLR8 [17 , 35] . Of note , a recent report identifying the molecular structure of TLR8 showed that this receptor recognizes degradation products of RNA , specifically an uridine mononucleoside at one binding site and oligonucleotides like UG or UUG at a distinct second binding site [14] . The concept of recognition of RNA degradation products by TLR8 raises the hypothesis that phosphatases and/or nucleases of bacterial or host origin might play a role upstream of TLR8 activation [36] , in analogy to the requirement for lysosomal endonuclease DNase II for the activation of TLR9 [37 , 38] . In agreement with this evidence , we could observe that RNA digested by a specific prokaryotic RNase was able to inhibit MHC-I in the same manner as intact RNA . These results indicate that these degradation products could be sensed by TLR8 . Although our results indicated that human TLR3 and TLR7 are not involved in MHC-I inhibition , Campos et al have recently demonstrated that both receptors play an important role in sensing B . abortus RNA to induce the production of pro-inflammatory cytokines and type I IFN expression in murine DCs . However , these receptors were not required to control Brucella infection in vivo [39] . To explain the latter , they hypothesized that TLR13 , a PRR involved in sensing a specific sequence from bacterial 23S rRNA [40 , 41] , could play a role in B . abortus RNA sensing . However , we did not focus our attention on TLR13 since it is a receptor present in mice but not in humans [40 , 41] . Beyond TLRs , RIG-I and MDA5 have been characterized as cytosolic receptors capable of sensing RNA . More specifically , MDA5 was identified to initiate antiviral signaling in response to long stretches of viral double-stranded RNA , whereas RIG-I is a sensor of short double-stranded or single-stranded RNA with 5’-triphosphate termini [42 , 43] . Moreover , RIG-I was also involved in the recognition of bacterial RNA . It has been demonstrated that RIG-I detects infection with live Listeria monocytogenes by sensing the RNA secreted into the cytosol of infected cells [44 , 45] . Although RIG-I and MDA5 could also be involved in the MHC-I inhibition mediated by B . abortus RNA , taking into account the cytosolic location of such receptors , the RNA should be able to be transferred from the phagosomes into the cell cytosol . However , this does not occur for all bacteria as it was demonstrated for B . burgdorferi RNA [33] . Regarding Brucella , it was demonstrated that B . abortus DNA can activate cytosolic molecules such as AIM2 and STING [46 , 47]; however , whether RNA might gain access to cytosolic receptors during phagocytosis has not been investigated yet . We have previously demonstrated that the EGFR pathway is involved in the inhibition of MHC-I mediated by B . abortus infection [6] . Moreover , that EGF and TGF-α are EGF-like ligands involved in the phenomenon of MHC-I inhibition [6] . In light of the results that we obtained in this study and those recently published [6] , we next investigated if there was a connection between RNA sensed by TLR8 and the EGFR pathway . EGFR neutralization with Cetuximab led to a partial reversion of the TLR8 agonist mediated-MHC-I inhibition suggesting a clear connection between the TLR8 and EGFR pathways . To our knowledge , this is the first report describing a link between these pathways . Overall , the results obtained in this study support a model in which infection with B . abortus induces the release of its RNA and RNA degradation products into the Brucella-containing endosomes . These molecules via TLR8 induce the secretion of EGF-like ligands such as EGF and TGF-α which bind ErbB receptors on the cell surface causing their activation . These effects finally lead to the retention of MHC-I molecules within the Golgi apparatus . MHC-I molecules are therefore unable to reach the cell surface and present bacterial Ags to CD8+ T cells ( Fig 10 ) . Here we elucidate that the vita-PAMP RNA is a component employed by B . abortus to inhibit MHC-I expression , an event whereby the bacteria could avoid the cytotoxic CD8+ T cell immunological surveillance establishing a chronic infection .
|
Brucella abortus is one of the intracellular bacterial species that cause brucellosis , a zoonotic worldwide disease . An intricate enigma of Brucella immunity is its long-term persistence inside host despite a vigorous and specific immune response . Our study describes a novel immune evasion strategy exploited by B . abortus: the down-modulation of the IFN-γ-induced expression of the molecules responsible for antigen presentation to CD8+ T cells on the surface of monocytes , Major Histocompatibility Complex Class I ( MHC-I ) molecules . We found that the bacterial component responsible for this phenomenon is its RNA , a component associated with bacterial viability itself . Specifically , we demonstrated that not only does intact RNA down-modulate MHC-I but also the RNA degradation products . Bacterial RNA is sensed by pattern recognition receptors ( PRRs ) , among which TLRs family has gained more attention . In this study , we demonstrated that the receptor involved in this phenomenon is TLR8 , a single-stranded RNA and RNA degradation products sensor . By means of this mechanism , Brucella impairs antigen presentation to CD8+ T cells , hiding within infected cells and avoiding the immunological surveillance of cytotoxic cells . On balance , these results provide new evidence to understand how B . abortus can survive inside the host and persist chronically .
|
[
"Abstract",
"Introduction",
"Results",
"Methods",
"Discussion"
] |
[
"blood",
"cells",
"flow",
"cytometry",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"rna",
"extraction",
"microbiology",
"organisms",
"cytotoxic",
"t",
"cells",
"extraction",
"techniques",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"white",
"blood",
"cells",
"animal",
"cells",
"medical",
"microbiology",
"t",
"cells",
"brucella",
"microbial",
"pathogens",
"spectrophotometry",
"cytophotometry",
"biochemistry",
"rna",
"double",
"stranded",
"rna",
"cell",
"biology",
"monocytes",
"nucleic",
"acids",
"virulence",
"factors",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"spectrum",
"analysis",
"techniques"
] |
2017
|
B. abortus RNA is the component involved in the down-modulation of MHC-I expression on human monocytes via TLR8 and the EGFR pathway
|
Osteosarcomas ( OS ) are complex bone tumors with various genomic alterations . These alterations affect the expression and function of several genes due to drastic changes in the underlying gene regulatory network . However , we know little about critical gene regulators and their functional consequences on the pathogenesis of OS . Therefore , we aimed to determine microRNA and transcription factor ( TF ) co-regulatory networks in OS cell proliferation . Cell proliferation is an essential part in the pathogenesis of OS and deeper understanding of its regulation might help to identify potential therapeutic targets . Based on expression data of OS cell lines divided according to their proliferative activity , we obtained 12 proliferation-related microRNAs and corresponding target genes . Therewith , microRNA and TF co-regulatory networks were generated and analyzed regarding their structure and functional influence . We identified key co-regulators comprising the microRNAs miR-9-5p , miR-138 , and miR-214 and the TFs SP1 and MYC in the derived networks . These regulators are implicated in NFKB- and RB1-signaling and focal adhesion processes based on their common or interacting target genes ( e . g . , CDK6 , CTNNB1 , E2F4 , HES1 , ITGA6 , NFKB1 , NOTCH1 , and SIN3A ) . Thus , we proposed a model of OS cell proliferation which is primarily co-regulated through the interactions of the mentioned microRNA and TF combinations . This study illustrates the benefit of systems biological approaches in the analysis of complex diseases . We integrated experimental data with publicly available information to unravel the coordinated ( post ) -transcriptional control of microRNAs and TFs to identify potential therapeutic targets in OS . The resulting microRNA and TF co-regulatory networks are publicly available for further exploration to generate or evaluate own hypotheses of the pathogenesis of OS ( http://www . complex-systems . uni-muenster . de/co_networks . html ) .
Osteosarcoma ( OS ) is a rare type of cancer frequently occurring in children and young adolescents [1] . It is a complex tumor typically accompanied by severe genomic instability and extensive mutations hampering the identification of a genetic root [2]–[4] . These genomic alterations affect several genes to a varying extent depending on patient and OS subtype . For instance , frequent mutations and deletions of the tumor suppressor genes TP53 , RB1 , and CDKN2A and mutations and amplification of the MYC locus [5] , [6] . However , their interactions in the molecular pathogenesis and the underlying cellular network of OS are poorly characterized . Recently , attention has been focused on the impact of microRNAs in OS . Besides transcription factors ( TFs ) that transcriptionally regulate gene expression , microRNAs are a class of small , conserved , non-coding RNA molecules generally acting on the post-transcriptional level . They are mono- or polycistronically transcribed , processed to mature molecules and subsequently incorporated into the RNA Induced Silencing Complex ( RISC ) . Once integrated in RISC , microRNAs are able to select their target genes via binding to partially complementary sequences in the 3′-UTRs of mRNAs that lead to mRNA degradation or translational inhibition . Computational prediction methods revealed that individual microRNAs regulate hundreds of target genes and one target gene might be regulated by several microRNAs [7] . According to Friedman et al . [8] around 60% of human genes are predicted to be regulated by multiple microRNAs in a cooperative manner . This huge number of target genes suggests a widespread control of biological processes including differentiation , proliferation , migration , and apoptosis [9] . In cancer , microRNAs might serve as onco- and/or tumor suppressor-microRNAs . Amplification or over-expression of oncogenic microRNAs can down-regulate tumor suppressor proteins . Likewise , deletion or under-expression of tumor suppressor microRNAs might lead to the up-regulation of oncogenes [10] . In addition , more than 50% of microRNA genes are located within fragile sites in the genome and are frequently subjected to chromosomal alterations [11] . In this manner , they can affect cancer development and progression . MicroRNAs share several regulatory concepts with TFs , e . g . they simultaneously regulate many target genes and cooperate with other regulators . However , TFs activate or repress their target gene expression whereas microRNAs regulate their targets primarily through repression to fine-tune cell-specific gene regulatory programs [12] . Because the expression of microRNAs often depends on TF regulation and vice versa , it is not surprising that both families of regulators are tightly related to each other in gene regulatory networks . The coordinated transcriptional regulation of microRNAs and their target genes by TFs is a recurrent network motif . The two types of gene regulators frequently form 3-node feedforward loops ( FFLs ) with common target genes [13] . Recently , Sun et al . [14] extended this regulatory motif to 4-node FFLs by integrating additional TF target genes . The extension of 3-node to 4-node motifs illustrated a more detailed model of the oncogenesis of glioblastoma by recruiting additional disease genes not directly targeted by microRNAs . Several studies have shown an involvement of microRNAs in the pathogenesis of OS . They demonstrated down-regulation of miR-143 in OS progression [15] , up-regulation of the oncogenic miR-17∼92 cluster in OS cells [16] , and regulatory functions for miR-199a-3p [17] , miR-21 [18] , and miR-125b [19] in OS cell proliferation and migration . Additional genome-wide microRNA analyses suggested sets of microRNAs to discriminate OS from osteoblasts and bone tissue [20]–[23] . All studies proposed the use of microRNAs as biomarkers in OS that might correlate with clinico-pathological parameters . However , those studies lack a comprehensive analysis of the functional consequences of aberrant microRNA expression in OS . Analyzing microRNAs in the context of their microRNA and TF co-regulatory networks might provide a deeper understanding of the pathogenesis of OS . In this study , we joined different data sources to analyze the contribution of microRNA and TF co-regulatory 3-node and 4-node motifs to the proliferative activity of OS cells . First , we divided seven OS cell lines into high and low proliferation groups by performing proliferation assays . Expression analysis based on these groups yielded differentially expressed ( DE ) microRNAs and mRNAs . Second , high efficacy microRNA target genes were obtained by integrating computational predicted targets with DE mRNAs . Only microRNAs with significantly enriched target genes were considered in the analysis . Third , microRNA target genes were clustered according to their functional similarity to explore their distinct biological processes . Fourth , transcription factor binding site ( TFBS ) information was added to assemble 3-node and 4-node motifs of non-random microRNA and TF co-regulator pairs . Fifth , the resulting 3-node and 4-node motifs were merged to form microRNA and TF co-regulatory networks to examine the coordinated regulation of microRNAs and TFs ( Figure 1 ) . Here , we present the first study analyzing microRNA and TF co-regulation in OS and uncover critical microRNA players of the functional processes implicated in OS cell proliferation .
In order to investigate the deregulated microRNA and TF co-regulatory networks of proliferative OS cells , we used seven authenticated OS cell lines . The cell lines were divided according to their proliferative activity by performing a proliferation assay . Four OS cell lines exhibited a high proliferative activity with a doubling time <10 hours while three showed less proliferation ( Table 1 ) . MNNG/HOS , U2-OS , and SJSA-1 showed additional extensive migratory capabilities . The expression analysis of the microRNAs was based on these two proliferation groups . The analysis yielded nine down-regulated and eight up-regulated microRNAs that passed the differential expression criteria ( False discovery rate ( FDR ) <0 . 05 & log2 Fold change ( FC ) ≥|1| , Table 2 ) . The derived DE microRNAs have been reported in association with neoplastic disease either due to oncogenic or tumor suppressor properties . Hierarchical clustering of them clearly separated the OS cell samples according to their proliferative activity ( Figure 2 ) . Hence , we selected the DE microRNAs as candidates that might affect OS cell proliferation for further analysis . To explore the functional consequences of DE microRNAs on OS cell proliferation , we determined their target genes by integrating gene expression profiles with computational predicted target genes . First , the expression analysis of mRNAs resulted in a total of 666 up-regulated and 610 down-regulated mRNAs . We applied loose filter criteria for DE mRNAs without correcting for multiple tests ( p-value<0 . 05 & log2 FC≥|0 . 7| ) because microRNA regulation might lead to subtle changes in gene expression . Next , we superimposed the DE genes with predicted microRNA targets to obtain target genes affecting OS cell proliferation . We assumed that microRNAs exhibit an inverse regulatory relationship to their functional target genes , i . e . microRNA expression is inversely correlated to its target gene expression . Hence , down-regulated targets were assigned to up-regulated microRNAs in high proliferative OS cells and vice versa . To exclude DE microRNAs with random association to OS cell proliferation , we tested for microRNA target gene enrichment within the list of DE genes . Among the 17 DE microRNAs , 12 are significantly enriched due to their targets ( FDR<0 . 05 , Table S1 ) . To account for different numbers of targets that might influence the enrichment analysis , we also computed the target gene enrichment of 1 , 000 permuted samples . The permutation procedure confirmed previous results ( Figure S1 ) . Consequently , we excluded 5 microRNAs from further analyses ( miR-92b , let-7f , miR-9-3p , miR-151-5p , and miR-100 ) . The remaining 12 OS proliferation-related microRNAs are implicated in the regulation of 474 target genes . Hierarchical clustering of the target genes resulted in a distinct separation of the high and low proliferative OS cell samples ( Figure 3A ) . We further investigated the underlying biological processes of OS proliferation-related microRNAs . We classified microRNA target genes according to their functional similarities of their gene ontology ( GO ) biological process terms using fuzzy c-means clustering ( FCM ) . After determining FCM parameters ( Figure S2 ) , we obtained two clusters . Principal Component Analysis ( PCA ) supported the results . The first two components separate the determined clusters ( Figure 3B ) . Cluster C1 consists of 172 members and cluster C2 contains 212 members . The remaining 90 microRNA targets could not be annotated with a GO biological process term and were excluded from further analysis . The clustering suggested that the microRNA targets can be classified into two broad functional classes . GO enrichment analyses revealed that members of C1 are mainly involved in metabolic processes like protein modification , nucleic acid metabolism , and carbohydrate metabolism , whereas members of C2 are implicated in signal transduction pathways leading to proliferation , differentiation , apoptosis , and migration . Both clusters demonstrate that cancer cells adapt metabolic processes for cell proliferation and survival [24] . A comparison between the five most informative GO terms ( FDR<0 . 05 ) illustrating the specific biological aspects of each cluster is shown in Figure 3C . Transcriptional regulation of TFs is tightly coupled with the post-transcriptional regulation of microRNAs . We utilized their 3-node and 4-node co-regulatory motifs to study DE microRNA and TF co-regulation in OS cell proliferation for each functional cluster . Every possible 3-node and 4-node FFL motif was determined to assess significant microRNA and TF combinations ( FDR<0 . 2 ) by using the hypergeometric test ( Table S2 and S3 ) . For the 3-node FFL , we obtained non-random microRNA and TF pairs with common target genes ( Figure 4A ) . For 4-node FFL motifs , we assessed non-random microRNA and TF pairs that regulate gene neighbors in the protein interaction network ( Figure 4B ) . The individual 3-node and 4-node FFL motifs are listed in Table S4 and S5 , respectively . Subsequently , we analyzed the co-regulated target genes of significant microRNA and TF combinations . The results are summarized in Table 3 . Noticeably , the microRNA and TF duo with the highest number of co-regulated target genes in both functional clusters is miR-9-5p and SP1 ( Figure 5 ) indicating a prominent role in OS cell proliferation . Further , we examined the co-expression of genes co-regulated by the same microRNA and TF pairs . We computed the Pearson correlation coefficients between co-regulated gene pairs as a measure of their co-expression . The distribution of the resulting correlation values was compared to the correlation distribution of random genes by the Kolmogorov-Smirnov ( KS ) test . The co-expression of co-regulated gene pairs tends to be significantly higher than for random genes ( p-value<2 . 2×10−16 , Figure 4C ) . This result supports the hypothesis of non-random microRNA and TF co-regulation within the list of their common or interacting target genes and suggests a similar functional context for their targets . Subsequently , we constructed the microRNA and TF co-regulatory networks that highlighted the combinatorial regulation patterns and regulated biological processes of microRNAs and TFs . The networks of C1 and C2 were generated by joining all significant co-regulatory relationships of microRNAs and TFs ( Table 3 ) . The resultant microRNA and TF co-regulatory networks are provided for full exploration on our website ( http://www . complex-systems . uni-muenster . de/co_networks . html ) . To assess the contribution of individual nodes in the co-regulatory networks on the networks' stability and robustness , we calculated the node degree and betweenness centrality parameters . The node degree distributions are highly right skewed . A large fraction of nodes shows a low degree and only few nodes have high degrees ( Figure S3 ) . Almost all microRNAs and TFs are located at higher node degrees as indicated by their average node degrees ( C1: microRNAs 19 and TFs 19 , C2: microRNAs 25 and TFs 49 ) . We expected that finding as microRNA and TF co-regulation is the main subject of the present study . Each network contains three types of nodes , namely microRNAs , TFs , and target genes . We ranked the nodes according to their node degrees and node type . The top 25% of microRNAs and TFs and the top 5% of target genes were considered as hubs in the C1 and C2 networks ( Table S6 ) . We detected the hub microRNA miR-214 and the hub TFs CREB1 , SP1 , and ZIC2 in both networks suggesting a central function in OS cell proliferation . Strikingly , around 50% of microRNA and TF target gene hubs in the two networks are TFs themselves . The microRNA and TF co-regulatory network derived from C1 contains ATF6 , GTF2A1 , HIVEP2 , KLF5 , LMO3 , NFKB1 , and TBPL1 and the C2 network comprises BCL6 , BCL6B , E2F4 , HES1 , JUN , LMO4 , RARA , REST , SIN3A , TCF7L2 , and ZBTB16 . Some of these TFs ( ATF6 , E2F4 , JUN , RARA , and REST ) are implicated in building 3-node and 4-node FFL motifs in one or two networks . The remaining TFs were either not existent in the UCSC conserved TFBS track or do not produce any significant FFL . Additionally , we found epigenetic modulators and genes involved in protein modification processes , like protein ubiquitination and phosphorylation . As already mentioned , the microRNA targets in C2 are associated with signal transduction for maintaining OS cell proliferation . Among the hub genes in the network derived from C2 , 30% of target gene hubs ( AMOT , ARF6 , CACNA1A , CTNNB1 , GRB2 , NOTCH1 , PDGFRB , PIK3R1 , SMAD7 , and TGFBR2 ) are related to signaling pathways that participate in cell proliferation , survival , and migration . Moreover , we assessed over-represented functional pathways derived from the KEGG database [25] . The enriched categories ( FDR<0 . 05 ) are shown in Table S7 . We detected an enrichment of genes involved in the cell cycle and cancer related pathways in both networks . We expected to observe these functional categories as we analyzed the proliferative potential of OS tumor cell lines . Further , we observed a similar functional trend between the C1 and C2 networks as for their corresponding functional clusters . Within the co-regulatory network of C2 , signaling pathways are significantly over-represented such as the MAPK- , TGFB- , and WNT-signaling pathways . In contrast , the network of C1 comprises a significant number of genes required for the basal transcription machinery . After examining the global co-regulation patterns of microRNAs and TFs in both networks , we were interested in sets of microRNAs and TFs that co-regulate densely connected network modules . To investigate the local structure of the OS proliferation-related co-regulatory networks , the walktrap algorithm was applied [26] . The algorithm obtained six modules within the metabolic network of C1 ( Figure S4 ) and six modules in the signaling network of C2 ( Figure S5 ) . The size and node types within each module are indicated in Table 4 . Strikingly , miR-9-5p is located in the largest module and is regulated through the TFs ATF2 , BACH1 , CREB1 , and SP1 in both networks . As mentioned before , miR-9-5p and SP1 co-regulate the largest number of target genes and thus indicate a prominent function in OS cell proliferation . Further , we run the Functional Annotation clustering Tool of the DAVID database [27] to classify the distinct network modules according to their GO biological process and molecular function terms . We annotated each module with the biological aspect of its maximum enrichment score ( ES ) . Among the 12 modules , five are mainly involved in transcriptional regulation processes , which is in accordance with previous studies that illustrated that microRNAs function via TFs to regulate various biological processes like cell proliferation [28] , [29] . Despite the top scored biological associations , one module ( C2 . 1 ) is related to negative regulation of differentiation ( ES>3 . 9 ) , particularly to osteoblast differentiation due to the genes CDK6 , MEN1 , SKI , SMAD3 , and SOX2 , which might provide a link to the pathogenesis of OS ( Figure S5A ) . Within this module , the TF MYC ( node degree 83 ) co-regulates several targets with miR-138 ( node degree 25 ) . The top ranked target gene in this module is SIN3A ( node degree 57 ) .
The microRNA and TF co-regulatory networks modeling OS cell proliferation are based on 12 proliferation-related microRNAs . Among these microRNAs , 11 were previously mentioned in OS [20] , [22] , [23] , [18] , whereas miR-138 was exclusively obtained in this study . Previous studies focused on global microRNA alterations in OS with respect to osteoblast cells and bone tissue . However , microRNA expression was partially inconsistent between different studies . Namløs et al . [23] hypothesized that contradictory microRNA regulation in different genome-wide studies might be explained due to distinct differentiation stages of OS progenitor cells . In this study , the DE microRNAs were assessed between high and low proliferative OS cells . The varying proliferation activity between these cell lines might reflect distinct differentiation stages of their progenitor cells . This might explain the detection of proliferation-related microRNAs in global OS studies . In different biological contexts the down-regulated microRNAs have been associated with decreased proliferation [30]–[33] . In turn , up-regulated microRNAs have been related to an increased proliferation potential [34]–[37] . These studies emphasize the proposed influence of the derived microRNAs on OS cell proliferation . The assembly of non-random microRNA and TF co-regulators revealed several interesting combinations . Based on the huge number of co-regulated target genes , the most notably co-regulation duo is the miR-9-5p and SP1 pair . Depending on the derived cluster , these co-regulators seem to affect distinct functional pathways due to their target genes . The C1 derived miR-9-5p and SP1 co-regulated target genes seem to participate in NFKB-signaling . Commonly regulated target genes include the TFs NFKB1 , NFKB2 , RELA , RELB , and BCL3 and the inhibitors NKRF , NFKBIA , and TNIP2 that cooperatively activate or block target gene expression of NFKB , respectively [38] . This pathway is implicated in OS cell proliferation [39] , and NFKB1 is an experimentally validated target gene of miR-9-5p [40] . Furthermore , a regulatory circuit including SP1/NFKB1/HDAC and miR-29b is known to induce leukemic growth [41] . Thus , miR-9-5p might function in a similar context in OS than miR-29b in leukemia . On the other hand , the C2 derived miR-9-5p and SP1 co-regulated genes might be involved in focal adhesion . We observed cadherins ( CDH1 , CDH2 , DSC2 , and DSG2 ) , further cell adhesion molecules ( e . g . FN1 , ITGA6 , ITGB1 , JUP , PKP3 , and VCL ) , and calcium signaling receptors ( e . g . CACNA1A and CALR ) . According to StringDB [42] , almost all commonly regulated genes interact with each other indicating a functional association ( Figure S6 ) . A possible pathway of miR-9-5p and SP1 co-regulation could implicate CDH1 and CTNNB1 that modulate cell proliferation [43] . CTNNB1 is not a commonly regulated target gene of miR-9-5p and SP1 . However , we observed some of its binding partners ( e . g . CDH1 , CDH3 , CTNNBPI1 , RUNX2 , and SMAD7 ) . Additionally , CTNNB1 is a hub gene in the C2 co-regulatory network . Moreover , increased expression of miR-9-5p resulted in down-regulation of the NFKB-SNAIL pathway and simultaneously to up-regulation of CDH1 in melanoma cells [44] . All these observations suggest a central role of miR-9-5p and SP1 co-regulation in OS cell proliferation . Merging significant microRNA and TF co-regulators resulted in co-regulatory networks of C1 and C2 . The networks provided a global view on microRNA and TF co-regulation in OS cell proliferation . To analyze the local co-regulation patterns within the networks , we specifically extracted densely connected modules . One of these modules , namely the C2 . 1 module , is implicated in negative regulation of differentiation , particularly osteoblast differentiation . It contains the hubs miR-138 , MYC , and SIN3A . We hypothesize a function for members of C2 . 1 that might be specific for the pathogenesis of OS . According to StringDB [42] , the genes in the module form a densely connected network illustrating a tight functional relationship ( Figure S7 ) . The module comprises members of the cell cycle ( CCND1 , CCND3 , CDKN1A , CDKN2C , and CDK6 ) , all involved in the RB1-pathway [45] . The complex of the module members SIN3A , NCOR1 , SKI , and HDAC can bind to RB1 and repress E2F target genes [46] . Therefrom , we assumed a connection between module members and RB1 , which has been reported to be frequently deregulated in OS [5] . Further , SIN3A is an experimentally validated target gene of miR-138 [47] . In the global co-regulatory network of C2 as well as in C2 . 1 module , it depicts a hub gene . Its role in cancer is contradictory , on the one hand it shows tumor suppressor functions [48] . On the other hand , it acts in tumor growth [49] . Taken together with the fact that SIN3A can interact with RB1 , we suggest a possible role for SIN3A in the pathogenesis of OS . After examining and discussing the structural and functional aspects of the co-regulatory networks , we integrated the main results of the present study into a potential model of OS cell proliferation ( Figure 6 ) . The focus of the model is on microRNA and TF co-regulation of the microRNAs miR-9-5p , miR-138 , and miR-214 and the TFs SP1 and MYC . In proliferative active OS cell lines , miR-9-5p , miR-138 , and miR-214 are significantly down-regulated leading to the up-regulation of their direct target genes CDK6 , E2F4 , HES1 , ITGA6 , NFKB1 , NOTCH1 , and SIN3A . CDK6 phosphorylates RB1 and therewith the RB1/SIN3A/SKI/NCOR1/HDAC complex cannot repress E2F4 target gene expression [45] , [46] . Activated NOTCH1 induces HES1 and sustains NFKB-signaling through NFKB1 , NFKB2 , RELA , RELB [50] . CTNNB1 stabilizes cell-cell adhesions in complex with CDH1 . Unbound CTNNB1 can translocate to the nucleus [43] . All these signals end up in the nucleus where they induce expression of proliferation promoting genes like microRNAs up-regulated in high proliferative OS cells , CCND1 , and FN1 . In turn , they repress genes implicated in growth-arrest . The resulting microRNA and TF co-regulatory networks display a detailed picture of the regulation of OS cell proliferation . In the present study , we concentrated on distinct functional aspects unraveled from the networks . The outcome suggests that down-regulation of miR-9-5p , miR-138 , and miR-214 results in a strong proliferative phenotype of OS cells due to their impact on NFKB- and RB1-signaling and on focal adhesion . Our study provides potential therapeutic targets in OS and proposes concepts for further research . In addition , we demonstrated how systems biological approaches support the analyses of complex diseases .
We used microRNA and mRNA expression data of seven authenticated OS cell lines , six from our previously published study [16] and one additional ( ZK-58 ) . Prior to microarray analyses , RNA was isolated and further processed as described in [16] . MicroRNA and mRNA expression profiles were determined on Exiqon's miRCURY LNA and Affymetrix's Human Gene 1ST arrays , respectively . Conserved TFBSs ( hg19 ) were downloaded from the UCSC Table Browser [51] . The track contained predicted TFBSs conserved in the human/mouse/rat alignment that were determined by using the Transfac Matrix Database 7 . 0 [52] . Protein interaction data were obtained from BioGRID release 3 . 1 . 92 [53] . The OS cell lines were evaluated regarding their proliferative , migrative , and invasive potential by using in vitro-assays ( BD Biosciences ) . Cells utilized in the assays showed 60 to 80% confluence growth . Prior to the assays cell lines were synchronized to ensure a uniform cell growth . To analyze OS cell proliferation , duplicates of each cell line ( 1×105 cells ) were seeded in 25 cm2 cell culture flasks . Cells were harvested at 24 , 48 , 72 , 96 , and 168 hours of growth . The cell number was determined by an automated cell counter ( Beckman Coulter ) . For each cell line and time point , the mean cell number was calculated to estimate the growth rate and subsequently the doubling time . Further , the Biocoat Matrigel Invasion Assay and a migration assay ( BD Biosciences ) were performed for each cell line in duplicate with matrigel-coated and uncoated inserts , respectively . Experiments were performed according to the manufacturer's instructions . Evaluation of invaded and migrated cells was done after 24 and 48 hours by light-microscopic analysis . Ten visual fields ( magnification 10× ) were analyzed by counting stained cells on the membranes . The expression data sets were analyzed using the Bioconductor package limma [54] . DE genes between high and low proliferative OS cells were determined using eBayes [54] . MicroRNA expression data were annotated with miRBase release 18 [55] , background corrected by normexp+offset 10 [56] , and normalized with printtiploess followed by RQuant [57] . In the differential expression analysis we considered the top 75% of microRNA probes that showed largest variation over all samples . Multiple test correction was performed using Benjamini and Hochberg's FDR approach [58] . The mRNA expression data were preprocessed by the Bioconductor package affy [59] . The Affymetrix Human Gene 1 ST array contains probes mapping among the whole transcript . Therefore , we filtered probes that mapped to exons present in at least 80% of a gene's transcripts to get one stable expression value per gene . Transcripts were derived from Ensembl release 63 [60] . The raw probe intensities were background corrected , normalized , and summarized to the gene-level by applying the robust multi-array average algorithm ( rma ) [59] . Predicted microRNA targets were obtained by running the local perl scripts targetscan_60 . pl and targetscan_61_context_scores . pl that were online available at the TargetScan website ( http://www . targetscan . org/ ) [61] . Mature microRNA sequences were downloaded from miRBase release 18 [55] . To derive high efficacy targets , we filtered target predictions with a context score≤−0 . 1 [61] . To determine TF target genes , we downloaded the transcriptional start sites ( TSSs ) of all genes included in the mRNA expression data set from the hg19 assembly of the UCSC Table Browser [51] and the TSSs from our proliferation related microRNA genes from miRStart [62] . Further , we defined the promoter region to −/+ 2000 nucleotides around the TSSs . Genes having a TFBS that completely overlapped their promoter regions were considered as TF targets . We just considered human TFs that were expressed in at least one proliferation group of our OS cell samples ( log2 intensity≥8 ) . To assess the enrichment of microRNA target genes in the list of DE genes , a hypergeometric test was performed . Multiple test correction was done by determining the FDR according to Benjamini and Hochberg [58] . To account for the different number of target genes of the DE microRNAs , a permutation procedure was applied . We randomly selected the number of DE genes out of the genes in our mRNA expression data set and counted the number of random microRNA target genes . For each permutation an enrichment score ( ES , -log10 p-value ) was calculated . The permutation procedure was repeated 1 , 000 times . The resulting permutation p-values were obtained by counting the number of permuted ESs exceeding the observed one . This was done for all DE microRNAs separately . To classify DE microRNA targets according to their functional similarity , their GO semantic similarity scores based on biological process terms were computed using Resnik's information content approach of the GOSim package [63] . The resulting functional similarity scores for any target gene pair were listed in a similarity matrix , which was further utilized as distance matrix for clustering . We applied FCM [64] to classify microRNA target genes according to their functional similarity using the function fanny from the R cluster package [65] . The fuzziness parameter was estimated by Dunn's coefficient [66] among a range of 1 . 1 to 1 . 5 and the cluster number was estimated over a range of 2 to 15 using Dunn's index [67] . The Dunn coefficient ( range 0–1 ) indicates the fuzziness of a cluster [66] . The Dunn index compared the between cluster variation to the within cluster variation and measures the cluster separation . A Dunn index >1 indicates a satisfying clustering [67] . The derived functional clusters were evaluated by running a GO enrichment analysis with the Bioconductor package GOStats [68] . Multiple test correction was performed by using Benjamini and Hochberg's FDR approach [58] . We tested for non-random microRNA and TF 3-node and 4-node motifs by using the hypergeometric test adapted from Sun et al . [14] . In contrast to them , we applied a different null model to derive p-values specific for the underlying microRNA and TF co-regulation pairs . For the 3-node motifs we tested if a microRNA and TF pair had significantly more commonly DE target genes than computationally predicted target genes . In turn , co-regulation of microRNA and TF pairs in 4-node motifs was tested based on commonly regulated secondary target genes and compared to the genes with corresponding TFBS in the whole 1st-neighbor protein interaction network . The 1st-neighbor protein interaction network was determined by extracting all 1st-neighbors of microRNA target genes from the protein interaction data . Benjamini and Hochberg's FDR was used to adjust for multiple testing [58] . Furthermore , evaluation of significant pairs of microRNAs and TFs was performed by assessing the coexpression of genes targeted by the same microRNA and TF pair . The Pearson correlation was used as a measure for coexpression . Statistical significance was determined by a permutation procedure . We randomly chose the same number of genes targeted by 3-node and 4-node FFLs out of all genes annotated in the mRNA expression data and computed their correlation coefficients . The permutation procedure was repeated 1 , 000 times . Finally , we tested if the coexpression of the genes in the FFLs was significantly greater than in randomly selected gene pairs using the KS test . The microRNA and TF co-regulatory networks were constructed by merging all 3-node and 4-node FFL motifs . The networks were modeled as graphs with nodes and edges . Nodes corresponded to microRNAs , TFs , or target genes and edges corresponded to microRNA-target regulation , TF-target regulation , or protein interactions . To identify crucial network players , we computed network centralities , namely node degree and betweenness , using the R package igraph [26] . The node degree is defined as the number of direct neighbors of a node in a network . Nodes having a high number of direct neighbors are thought to be important regulatory hubs inside the network . In contrast , a node's betweenness is a measure of the number of shortest paths between any pair of nodes that run through it [69] . To detect tightly connected groups of nodes in the network , we run the walktrap algorithm [70] . This algorithm finds modules in connected graphs . It is based on random walks and assumes that the random walker is trapped in dense parts of a network [26] . For further network evaluation we used the Functional Annotation Tool of the DAVID database [27] . The networks were visualized with Cytoscape 2 . 8 . 3 [71] and Cytoscape Web 1 . 0 . 2 [72] .
|
Osteosarcomas ( OS ) are bone tumors most frequently affecting children and young adolescents . We do not know much about its molecular pathogenesis hampering personalized therapies to more effectively cure patients . To this day , almost all patients receive adjuvant chemotherapies independent of its necessity . Hence , we need to gain a comprehensive understanding of the molecular pathogenesis of OS to uncover molecular components that can be used as therapeutic targets . MicroRNAs and transcription factors ( TFs ) are master regulators of the cellular system . They control the amount of genes expressed in a cell at a specific time point that ultimately results in a distinct cellular phenotype . Here , we investigated microRNA and TF co-regulatory networks in OS cell growth as one hallmark in cancer . We uncovered key microRNA and TF regulators that cooperatively control growth-related pathways in OS and proposed potential therapeutic targets . This study illustrates the benefit of analyzing complex diseases from a network perspective because no molecular component functions isolated from the underlying cellular system .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
How MicroRNA and Transcription Factor Co-regulatory Networks Affect Osteosarcoma Cell Proliferation
|
Tachyzoite to bradyzoite development in Toxoplasma is marked by major changes in gene expression resulting in a parasite that expresses a new repertoire of surface antigens hidden inside a modified parasitophorous vacuole called the tissue cyst . The factors that control this important life cycle transition are not well understood . Here we describe an important transcriptional repressor mechanism controlling bradyzoite differentiation that operates in the tachyzoite stage . The ApiAP2 factor , AP2IV-4 , is a nuclear factor dynamically expressed in late S phase through mitosis/cytokinesis of the tachyzoite cell cycle . Remarkably , deletion of the AP2IV-4 locus resulted in the expression of a subset of bradyzoite-specific proteins in replicating tachyzoites that included tissue cyst wall components BPK1 , MCP4 , CST1 and the surface antigen SRS9 . In the murine animal model , the mis-timing of bradyzoite antigens in tachyzoites lacking AP2IV-4 caused a potent inflammatory monocyte immune response that effectively eliminated this parasite and prevented tissue cyst formation in mouse brain tissue . Altogether , these results indicate that suppression of bradyzoite antigens by AP2IV-4 during acute infection is required for Toxoplasma to successfully establish a chronic infection in the immune-competent host .
Toxoplasma gondii is an obligate intracellular parasite that exhibits a multi-host and multi-stage developmental life cycle . Sexual stages are restricted to the gut mucosa of the feline definitive host and asexual stages of the intermediate life cycle occur within any warm-blooded host , including humans . Acute disease is generally asymptomatic in immune-competent hosts , however , primary or recrudescent infection from latent bradyzoites in humans with AIDS , those undergoing chemotherapy or in the unborn cause significant disease and death [1] . While the tachyzoite lytic cycle is responsible for disease pathology in human hosts , the interconversion of the tachyzoite stage into the bradyzoite stage underlies chronic infection and ensures host to host transmission [2] . Evidence indicates that the tachyzoite cell cycle is intricately tied to bradyzoite differentiation with the choice to continue tachyzoite replication or develop into the latent bradyzoite containing tissue cyst made during S phase and/or mitosis [3 , 4] . Transcriptome data demonstrates that unique changes in mRNA expression occur in the tachyzoite cell cycle and during development [5–7] . An estimated ~5% of all transcripts are exclusive to a single developmental stage with nearly 40% of the mRNAs in the tachyzoite division cycle periodically expressed . How these changes are controlled is largely unknown . Early genome mining for known gene specific transcription factors revealed two important observations . While the general transcriptional machinery is present in the genomes of Apicomplexa species , initial studies failed to identify classic gene specific transcriptional regulators common to higher eukaryotes . Second , an overall lack of genes encoding DNA binding proteins suggested a limited arsenal from which to regulate these dynamic changes in parasite developmental gene expression . In 2005 , a family of DNA binding proteins ( ApiAP2 factors ) distantly related to the APETALA family plant transcription factors was discovered encoded in the genomes of Apicomplexa species [8] . In Plasmodium spp . , ApiAP2 factors bind DNA with distinct sequence specificity [9 , 10] via a novel domain swapping mechanism [11] and have non-transcriptional roles in sub-telomeric chromosome biology [12] . Examples of ApiAP2 gene-specific functions in Plasmodium falciparum are ookinete ( AP2-O ) and sporozoite ( AP2-Sp ) ApiAP2s that serve as stage specific transcriptional activators regulating motile stages within the mosquito definitive host [13–15] . Genetic disruption of AP2-O results in non-invasive ookinetes [15] whereas disruption of the AP2-Sp locus yields a parasite unable to form viable sporozoites [14] . The Toxoplasma genome encodes 67 ApiAP2 domain-containing proteins ( ToxoDB and ref . [16] ) , with 24 of these genes expressed cyclically during the tachyzoite division cycle [5] . In Toxoplasma , ApiAP2s have been implicated in virulence and invasion mechanisms [17] , as part of chromatin remodeling complexes [18 , 19] and RNA processing machinery [20] and there is evidence for ApiAP2 factors regulating bradyzoite development . AP2XI-4 is up-regulated during bradyzoite development and the loss of AP2XI-4 blocks the stress-induction of some bradyzoite mRNAs , including the canonical marker , BAG1 [21] . The novel stress-inducible transcriptional repressor AP2IX-9 acts to prevent premature commitment to bradyzoite development through direct interaction with bradyzoite specific promoters [22] . Here we describe the discovery of a new level of developmental control in the intermediate life cycle that is required to establish the chronic tissue cyst stage in animals . AP2IV-4 is exclusively expressed in the tachyzoite division cycle with peak expression of the encoded mRNA and protein during early mitosis . Surprisingly , genetic knockouts of AP2IV-4 demonstrate it is non-essential to the replicating tachyzoite but is instead critical for the suppression of bradyzoite surface antigens and cyst wall proteins in the tachyzoite stage . Results from animal studies determined that AP2IV-4 silencing of bradyzoite gene expression is critical to enable tachyzoites to escape an effective immune response and produce the tissue cysts required for transmission .
Previous studies identified Toxoplasma ApiAP2 genes that are cyclically transcribed once per tachyzoite cell cycle with the peak timing of mRNA levels distributed throughout the division cycle [5] . The functions of periodically expressed ApiAP2 factors is largely unknown , although it is proposed they control the remarkable "just-in-time" cell cycle transcriptome of asexual stage Apicomplexa parasites [5 , 23] . A Group-of-12 of these periodic ApiAP2 mRNAs share overlapping cyclical profiles that reach maximum expression during the S through mitotic phases ( S/M ) of the Toxoplasma tachyzoite cell cycle ( Fig 1A ) . We made multiple attempts to knockout each of the Group-of-12 ApiAP2 genes ( Fig 1B ) in a Type I RH strain ( RHΔku80Δhxgprt = RHQ strain ) engineered for enhanced homologous recombination [24 , 25] . The results from this series of genetic experiments were mixed; half the Group-of-12 ApiAP2 genes were successfully disrupted in the RHQ strain at relatively high prevalence except AP2III-2 ( Fig 1B ) , while knockouts of the other half failed repeated attempts . A recent whole genome CRISPR screen performed in human fibroblast cells ( HFF ) [26] supports the preliminary RHQ experimental sorting of Group-of-12 ApiAP2 genes into dispensable versus required ( Fig 1B , CRISPR column ) . Alternate developmental expression ( and possible function ) , may help explain why half of the Group-of-12 ApiAP2s are not required for tachyzoite growth . AP2IX-4 [27] and AP2XI-4 [21] are also expressed in bradyzoites and recent studies indicate important roles for these factors in tissue cyst development; similarly AP2III-2 , AP2VI-1 , and AP2XI-1 are expressed in tachyzoites and bradyzoites [28] . Notably , AP2VI-1 mRNA is expressed at high levels across the Toxoplasma intermediate and definitive life cycles ( the only ApiAP2 with this profile ) and AP2III-2 is highly expressed in unsporulated oocysts [28] . Three of the Group-of 12 ApiAP2 factors that were expressed in tachyzoites and not induced by stress conditions ( AP2IV-4 , AP2XII-2 , AP2XII-9 ) [28] also failed knockout attempts in RHQ parasites and had significant negative phenotype scores in the HFF/CRISPR study ( Fig 1B ) . At the mRNA level AP2IV-4 stood out as one of the most dynamic of the Group-of-12 ApiAP2s with a >10-fold change in mRNA abundance over a ~2 h period in the second half of the tachyzoite cell cycle ( Fig 1A , red curve ) . The failure to disrupt AP2IV-4 in RHQ parasites indicated an important function in tachyzoite replication . To verify AP2IV-4 protein is cell cycle regulated , we introduced three copies of the HA epitope tag in frame with the C-terminus of the AP2IV-4 coding region by genetic knock-in , which preserved the native promoter and genomic flanking contexts . The gene model for AP2IV-4 ( http://toxodb . org/toxo/app/record/gene/TGME49_318470 ) predicts a single exon structure that encodes a large protein with a single AP2 DNA binding domain ( S1A Fig , diagram ) , which was verified by Western analysis of AP2IV-4HA expressing parasites ( S1A Fig , gel right ) . As with previously tagged Toxoplasma ApiAP2s ( e . g . ref . [5 , 22] ) , AP2IV-4HA in the RHQ strain localized exclusively to the nucleus ( Fig 2A ) and was cell cycle regulated ( Fig 2A and S1B Fig ) with a timing similar to its mRNA expression profile ( Fig 1A , red curve ) . In a randomly growing tachyzoite population , AP2IV-4HA was detectable in 30% of vacuoles ( S2A and S2B Fig ) due to cell cycle periodicity . To pinpoint the exact cell cycle expression of AP2IV-4HA , co-staining with two daughter cytoskeleton markers was utilized to define the timing of initiation , accumulation and degradation of the AP2IV-4HA fusion protein in comparison to the earlier expressing cyclical factor AP2VI-1HA ( Fig 1A , green curve ) also produced by genetic knock-in . Antibodies for the inner membrane complex ( S1B Fig , α-TgIMC1 ) [29] and the apical cap ( Fig 2 , α-TgISP1 ) [30] permit the late S phase through mitotic periods of the tachyzoite cell cycle to be resolved in time . AP2IV-4HA first appeared in tachyzoites lacking internal daughter IMC structures as did AP2VI-1HA , although by first AP2IV-4HA detection AP2VI-1HA had already reached maximum expression in these parasites ( S1B Fig , a vs . b images ) . The detection of AP2IV-4HA prior to internal daughter structures indicates initiation of expression in late S phase just prior to the start of mitosis and before nuclear division . The rapid accumulation of AP2IV-4HA paralleled the formation of ISP1 rings of the daughter parasites ( Fig 2A , a-c images ) and the growth of the IMC1 daughter scaffold ( S1B Fig , c , e images ) , while during these same cell cycle transitions AP2VI-1HA declined rapidly to undetectable levels ( Fig 2B , d , e images; S1B Fig , d , f images ) . AP2IV-4HA was highly expressed throughout budding ( distinct mitotic U-shaped nuclear morphology , DAPI staining , Fig 2A ) and began to disappear following nuclear division in late cytokinesis ( Fig 2A , images d , e ) and was no longer detectable after resolution of the mother IMC , which is consistent with the lack of AP2IV-4 mRNA expression in the G1 phase ( Fig 1A ) . Expression of AP2IV-4HA in tachyzoites induced to differentiate into bradyzoites by alkaline stress followed a pattern that was consistent with known changes in replication associated with this developmental pathway ( S2 Fig ) [3] . Alkaline-stress conditions caused AP2IV-4HA parasites to loose intravacular synchronous growth and this led to a heterologous pattern of AP2IV-4HA expression reflective of asynchronous cell cycle distributions in a single vacuole ( S2B Fig ) . However , co-staining alkaline-stressed populations for AP2IV-4HA and centrin , showed that the cell cycle restriction of AP2IV-4 to the second half of the cell cycle was preserved in the differentiating parasites ( representative image , S2C Fig ) . It is challenging to resolve centrosome counts in differentiating populations due to the lack of rosetting and parasite stacking , however , it is clear in the representative example shown that parasites lacking AP2IV-4HA possess a single centrosome . Further , AP2IV-4HA positive parasites are wider than their more slender "vacuolar mates" , which is a recognized morphological difference between replicating tachyzoites and dormant mature bradyzoites . Thus , during alkaline-stress induction AP2IV-4HA expression is marking those parasites that are replicating and are likely at an earlier stage in development as growth arrest in the G1 period is the ultimate outcome of this developmental pathway [3 , 31 , 32] . The expression of AP2IV-4 during early differentiation shares similarities with another of the Group-of-12 cell cycle ApiAP2 factors that we have studied , AP2IX-4 [27] . The failure to disrupt the AP2IV-4 gene in the RHQ strain ( Fig 1B ) suggested this factor was essential for tachyzoite growth , although other explanations such as low frequency double crossover or growth defects preventing the recovery of AP2IV-4 knockout parasites could explain the knockout failure . To investigate whether low frequency recombination was responsible , we applied Cre-Lox methods [33] to disrupt AP2IV-4 using the rapamycin-inducible diCRE model recently introduced into the Type I RHΔhxgprtΔku80 strain ( RHCre ) [33] . To “flox” the AP2IV-4 gene in the RHCre strain , we performed serial epitope tagging by genetic knock-in of the AP2IV-4 ( 3xHA tag ) and TGGT1_318480 ( 3xmyc tag ) genes ( see Fig 3A diagram ) , which are in a sequential head to tail configuration on chromosome IV . The AP2IV-4HA fusion protein that resulted from the production of the RHCre-AP2IV-4floxed strain ( RHCre-parent in these studies ) preserved the native promoter and reproduced the identical 3xHA fusion protein as was generated in the RHQ-AP2IV-4HA strain above ( see Fig 2A ) . Gene TGGT1_318480 is expressed at very low levels in tachyzoites or bradyzoites ( <30th mRNA percentile , ToxoDB . org ) and was undetectable by both IFA and Western blot following tagging with 3xmyc . Cre-mediated excision of the AP2IV-4 locus was induced by a 6 h incubation with rapamycin ( 50nM ) of RHCre-parent parasites ( Fig 3A ) . In contrast to the failure to knockout AP2IV-4 in the RHQ strain by conventional methods ( Fig 1B ) , ~20% ( 10/51 ) of isolated clones following rapamycin treatment of RHCre-parent parasites lacked the AP2IV-4 gene ( also no longer HA+ ) , and for two clones we verified the absence of AP2IV-4 mRNA ( S3A Fig ) . The isolation of viable RHCre-Δap2IV-4 transgenic parasites indicated AP2IV-4 is not essential for tachyzoite growth . Absent a requirement for AP2IV-4 in cell division , it was not clear what role AP2IV-4 serves in the tachyzoite . In order to build clues to understand AP2IV-4 function , duplicate total RNA samples from the RHCre-parent and RHCre-Δap2IV-4 tachyzoites were isolated ( standard tachyzoite conditions , pH 7 . 0 ) , converted to cRNA and hybridized to a custom Toxoplasma Affymetrix GeneChip [34] . In total , 40 mRNAs were altered >2-fold in RHCre-Δap2IV-4 tachyzoites , including 26 mRNAs that were up-regulated ( Fig 3A , results of selected genes; complete results in S1 Dataset ) . Remarkably , the loss of AP2IV-4 caused increased expression of mRNAs encoding known bradyzoite surface antigens ( e . g . SRS9 , SAG4 . 2 ) [35 , 36] and cyst wall components ( e . g . BPK1 , MCP4 ) [6 , 37] in the tachyzoite . These results indicated the major function for AP2IV-4 is to repress the transcription of key bradyzoite genes in replicating tachyzoites . This new level of developmental control of bradyzoite gene expression in the tachyzoite is distinct from the stress-induced AP2IX-9 mechanism we described previously [22] . Importantly , AP2IV-4 and AP2IX-9 combined appear to transcriptionally silence 66% ( 14/21 , S2 Dataset ) of the bradyzoite genes thought to be activated by AP2XI-4 [21] , which is also one of Group-of-12 ApiAP2s . The RHCre-Δap2IV-4 transgenic strain was complemented with the cosmid PSBM794 [38] that carries a Toxoplasma genomic DNA fragment spanning the AP2IV-4 gene and RNA samples from the resulting RHCre-Δap2IV-4::AP2IV-4 transgenic strain were analyzed on the Toxoplasma GeneChip . The results from this experiment determined that reintroduction of the AP2IV-4 gene restored mRNA repression to RHCre-parent levels for >80% of the mRNAs with the remaining mRNAs substantially reduced from the de-repressed levels of RHCre-Δap2IV-4 tachyzoites ( S1 Dataset ) . To confirm the function of AP2IV-4 in a second genetic lineage , we employed the same double-tagging strategy to "flox" the AP2IV-4 gene in the Type II Prugniaud strain ( PruQ-parent in this study: Pru-Δku80Δhxgprt/AP2IV-4floxed ) [39] followed by knockout of the AP2IV-4 gene by transient transfection of pMIN-CRE-eGFP plasmid ( S3B Fig , diagram ) [40] . Two confirmed PruQ-Δap2IV-4 clones lacking the AP2IV-4 gene were recovered from 178 independent clones screened ( S3C Fig ) . The successful disruption of the AP2IV-4 locus in PruQ confirms the dispensability of this factor for tachyzoite growth in a second genetic lineage . Microarray analysis of PruQ-Δap2IV-4 tachyzoites identified very similar gene expression changes to RHCre-Δap2IV-4 parasites; >90% of genes altered up or down by the loss of AP2IV-4 in these knockout strains were shared ( Fig 3A; complete lists S1 Dataset ) . In the PruQ-Δap2IV-4 parasites , bradyzoite mRNA fold changes were often less than in the RHCre-Δap2IV-4 tachyzoites due to higher starting baseline levels of mRNA expression in the PruQ-parent strain ( Fig 3A and S1 Dataset ) . Type II Pru strains have significant capacity to spontaneously form bradyzoites [28] and this raised the population baseline expression of bradyzoite mRNAs ( Fig 3A , PruQ-parent ) . Nonetheless , microarray studies of PruQ-Δap2IV-4 parasite mRNA expression clearly validated the conclusion that a major function of AP2IV-4 is to silence bradyzoite surface and cyst wall gene transcription in replicating tachyzoites . To verify bradyzoite-specific proteins are expressed in replicating tachyzoites lacking AP2IV-4 , we completed immunofluorescence assays ( IFA ) of PruQ-Δap2IV-4 ( Fig 3B and 3C ) as well as RHCre-Δap2IV-4 tachyzoites ( S4 Fig ) using antibodies to four bradyzoite specific proteins ( α-BPK1 , α-MCP4 , α-CST1 , α-SRS9 ) and also evaluated the formation of cyst walls using biotin-labeled Dolichos biflorus agglutinin ( DBA ) . Microarray probes for the recently discovered bradyzoite cyst wall protein CST1 [41] were not included in Toxoplasma GeneChip , although we suspected the CST1 gene could be a target of AP2IV-4 suppression in tachyzoites . This was confirmed at the mRNA level by RT-qPCR using CST1-specific primers ( see S3 Dataset for CST1 primer designs ) . Staining of PruQ-Δap2IV-4 ( Fig 3B and 3C ) and RHCre-Δap2IV-4 tachyzoites ( S4A Fig ) with CST1 antibodies confirmed the mis-timing of expression of this large cyst wall protein ( >250 kDa ) , which was reversed by genetic rescue of these knockout strains with a cosmid genomic clone carrying the AP2IV-4 gene ( Fig 3B and S4A Fig ) . Similar to CST1 , the cyst wall pseudokinase BPK1 and structural protein MCP4 as well as bradyzoite surface protein SRS9 were all increased in PruQ- and RHCre-Δap2IV-4 tachyzoites as was the number of DBA+-vacuoles , which was again reversed by genetic complementation ( Fig 3B and 3C and S4B and S4C Fig ) . In comparison to alkaline-induced PruQ-parental bradyzoites , PruQ-Δap2IV-4 tachyzoites expressed the bradyzoite mRNAs and proteins shown here within a normal range expected of bradyzoites and the proteins were properly localized to either cyst walls or the parasite surface ( Fig 3C , see also S4B Fig ) . The , expression of bradyzoite-specific genes in the Δap2IV-4 tachyzoites of either strain was the result of developmental mis-timing . Type I RH parasites are known to be resistant to developmental induction [42] , and therefore , it was remarkable that deletion of a single ApiAP2 factor could accomplish what strong alkaline stress fails to do in this strain . The expression of bradyzoite-specific antigens in Δap2IV-4 populations was not 100% for either strain , which likely reflects the restricted S/M cell cycle window that AP2IV-4 operates ( Figs 1 and 2 ) . Intriguingly , we reported previously that increased baseline bradyzoite mRNA expression occurs during the S/M periods of synchronized tachyzoites [5] . This hypothesis was examined further by co-staining PruQ-Δap2IV-4 and PruQ-parent tachyzoites with α-SRS9 and α-centrin antibodies ( S5A Fig ) . This IFA analysis determined that the majority of SRS9+/PruQ-Δap2IV-4 parasites possessed duplicated centrosomes ( S/M phases ) confirming that SRS9 misexpression was occurring primarily in the mitotic half of the tachyzoite cell cycle . Similarly , most of the 14 . 3% of PruQ-parental parasites that spontaneously expressed SRS9 also possessed duplicated centrosomes ( only 3% of parental parasites were in the G1 phase and also SRS9+ ) . Altogether , these results are consistent with the unique relationship between bradyzoite differentiation and tachyzoite mitosis that we discovered more than a decade ago [3] . ApiAP2 factors have been shown to regulate gene expression through the binding of target promoters in a sequence specific manner [9 , 10 , 12 , 14 , 15 , 17 , 22] . To assess DNA binding specificity for AP2IV-4 , a GST-AP2IV-4 fusion protein ( AP2 domain only ) was expressed , purified and incubated on a microchip containing all possible 10-mer DNA fragments ( Fig 3D , protein binding microarray results ) . A resulting 8-nucleotide “consensus” sequence motif bound specifically by the GST-AP2IV-4 fusion protein contains homopolymeric poly ( dC ) :poly ( dG ) ( Fig 3D , 5’-ACCCCCCT-3’/3’-TGGGGGGA-5’; enrichment score 0 . 497 ) . Electrophoretic mobility shift assays ( EMSA ) using DNA probes containing poly ( dC ) :poly ( dG ) repeats were used to validate the specificity of GST-AP2IV-4 binding ( Fig 3D , EMSA results ) [22] . GST-AP2IV-4 bound biotin-labeled DNA probes that contained a single instance of the “consensus” PBM motif and a second five nucleotide poly ( dC ) segment ( Fig 3D , lane 2 ) , and was successfully competed using 300x excess unlabeled poly ( dC ) competitor DNA ( Fig 3D , lane 3 ) , but not with unlabeled DNA probes that contained no poly ( dC ) stretch greater than three nucleotides ( Fig 3D , lane 4 ) . In a larger protein binding screen of 46 Toxoplasma AP2 domains , two other ApiAP2 factors , AP2VIIa-5 and AP2XII-4 , were determined to also specifically bind homopolymeric poly ( dC ) :poly ( dG ) DNA ( Kim et al , in preparation ) . In addition , a recent analysis of nucleosome-free regions for enriched DNA motifs discovered poly ( dC ) :poly ( dG ) repeats were preferentially found upstream of cell cycle and bradyzoite genes , such as SRS9 ( Wang et al , in preparation ) . The presence of poly ( dC ) :poly ( dG ) repeats in the SRS9 promoter ( Fig 3E , blue legend ) suggested AP2IV-4 might directly bind this promoter . To examine this question , we utilized the FKBP ( DD ) /Shield 1 conditional expression model [43] in order to improve the signal strength for AP2IV-4 expression , which has been very successful for studying ApiAP2 factors in P . falciparum and Toxoplasma [13 , 22 , 44] . The FKBP peptide combined with three copies of the HA epitope tag was fused to the N-terminus of the AP2IV-4 coding region ( DDAP2IV-4 ) by genetic knock-in methods . The addition of Shield 1 ( 100nM ) to RHQ-DDAP2IV-4 transgenic parasites successfully increased nuclear levels of DDAP2IV-4 , but did not disrupt the normal periodic cell cycle expression of this protein . Thus , there are likely significant post-transcriptional mechanisms regulating AP2IV-4 expression in tachyzoites as we also previously documented for AP2IX-9 expression [22] . Utilizing lysates prepared from RHQ-DDAP2IV-4 tachyzoites incubated with Shield 1 , we performed chromatin immunoprecipitation followed by quantitative PCR of eight regions covering ~1 , 200bp of the SRS9 promoter and 5'-UTR ( Fig 3E ) . The results from this experiment showed that binding of DDAP2IV-4 to the SRS9 promoter in parasite chromatin was enriched in regions 5–7 that includes a poly ( dG ) motif ( region 7 , yellow bar ) ~230bp upstream of the SRS9 ATG ( Fig 3E ) . To control for non-specific binding , we analyzed DDAP2IV-4 binding to the chromosome region ( -950 bp ) 5'-flanking of the BAG1 bradyzoite gene . In contrast to stress-induced AP2IV-3 , which we have recently reported activates BAG1 [28] , AP2IV-4 does not regulate BAG1 ( S1 Dataset ) . No enrichment of DDAP2IV-4 binding was detected to the six regions tested within the BAG1 promoter ( Fig 3E ) , whereas DDAP2IV-3 binding to this promoter is significantly enriched [28] ( Fig 3E , reference line graph ) . Our results indicated that PruQ-Δap2IV-4 parasites are tachyzoites expressing a few genes that are normally induced to high levels during bradyzoite development . Unlike bradyzoites ( S2B and S2C Fig ) , RHCre- and PruQ-Δap2IV-4 tachyzoites replicated synchronously within a shared vacuole , which is a growth behavior observed for all native tachyzoite strains . PruQ-Δap2IV-4 parasites also retained normal SAG1 surface antigen expression even as they also expressed bradyzoite-specific surface antigen , SRS9 ( S5B Fig ) . Finally , whole-cell mRNA analysis of PruQ-parent and PruQ-Δap2IV-4 tachyzoites revealed nearly identical transcriptomes ( S5C Fig ) , whereas , the transcriptomes of native Type II tachyzoites and bradyzoites have numerous differences and lower quantitative similarity ( S5C Fig ) . RHCre-Δap2IV-4 tachyzoites are a Type I strain and consistent with the limited number of genes regulated by AP2IV-4 the knockout of this gene did not disrupt the lethality of the Type I genetic background . Inoculation ( i . p ) of BALB/c mice with 500 RHCre-Δap2IV-4 ( or RHCre-parent ) parasites was fully lethal , while infections with a very high dose ( 1x107 ) of PruQ-Δap2IV-4 parasites was not ( Fig 4A ) . These data are consistent with the known difference in the virulence of Type I versus II strains , and while the PruQ-parent strain showed slightly higher virulence ( LD100 >106 ) this difference was minor compared to the acute virulence of RHCre strains lacking AP2IV-4 . By contrast , the loss of AP2IV-4 had a dramatic effect on PruQ-strain chronic infections in mice . Shifting PruQ-Δap2IV-4 parasites into alkaline media ( pH8 . 2 ) effectively slowed growth and induced high levels of DBA+ tissue cysts ( Fig 4B , in vitro = 93 . 1% ) demonstrating PruQ-Δap2IV-4 parasites were capable of forming cysts in vitro . However , PruQ-Δap2IV-4 infections of BALB/c mice failed to produce tissue cysts in brain tissue ( Fig 4B ) . The lack of tissue cyst formation was correlated with a lower parasite burden measured in BALB/c mice at 6 days post-inoculation ( Fig 4C ) , although lower parasite burden was not due to an inability of PruQ-Δap2IV-4 parasites to replicate in vivo . Mice infected intraperitoneally with PruQ-Δap2IV-4 display no reduction in the frequency of infection or growth as compared to the PruQ-parent or the complemented strain at day one post infection ( Fig 4D and S6A Fig ) , which is in agreement with the lack of any growth difference of these strains in HFF cell culture . This indicates that PruQ-Δap2IV-4 parasites efficiently invade host cells and replicate productively early in the infection , but as the infection progresses the parasites that lack AP2IV-4 are more effectively eliminated by the immune response of BALB/c mice . The development of bradyzoites and cysts in vivo is , at least partly , dependent upon immune factors . This is illustrated in immune deficient animals where bradyzoites fail to develop and mice succumb due to uncontrolled tachyzoite replication . The protective immune response during acute infection is dominated by the recruitment of inflammatory monocytes and T cell production of IFN-γ [45] . During chronic infection , IFN- γ is required to prevent parasite recrudescence . Yet , the host immune response is unable to clear the bradyzoite and cysts persist for long periods in host tissues . The pay off between protection and the development of chronic infection is poorly understood . At day 6 post Toxoplasma infection neutrophils and inflammatory monocytes , distinguished by their expression of Ly6 surface antigens and distinct morphologies ( Fig 5A and 5B ) , are present at the site of infection . Although neutrophils are important sources of IL-12 , inflammatory monocytes are the key effector cell in controlling parasite replication [46] . Analysis of the proportion of these populations following infection of parasites with an intact or disrupted AP2IV-4 gene revealed striking differences . Consistent with the response to a high inoculum of parasites [46] , infection with the PruQ-parent induced the influx of neutrophils , outweighing inflammatory monocytes nearly 3-fold ( Fig 5A , 5B and 5D ) . By contrast , PruQ-Δap2IV-4 infections at day 6 post-infection led to a significant increase in the proportion of inflammatory monocytes ( Fig 5A ) that was not observed early in the infection ( S6A Fig ) . This was confirmed by cytospin , where large numbers of polymorphonuclear neutrophils can be seen in the peritoneal exudate wash of mice infected with PruQ-parental parasites and foamy monocytes observed with PruQ-Δap2IV-4 parasite infection ( Fig 5B ) . Further , the absolute numbers of inflammatory monocytes recruited to the site of infection is greater in the absence of AP2IV-4 with a corresponding decrease in the recruitment of neutrophils ( Fig 5C and 5D ) . Although PruQ-Δap2IV-4 parasites can be found replicating in peritoneal exudate cells , cytospin counts suggest that overall , the proportion of inflammatory monocytes infected with this strain is highly reduced at day 6 post-infection ( S6B Fig ) . Genetic complementation of the PruQ-Δap2IV-4 parasites restored the innate immune response to that of mice infected with the PruQ-parent ( Fig 5 and S6B Fig ) . Together , these data indicate that in the absence of AP2IV-4 silencing of bradyzoite gene expression in the tachyzoite stage , there is an amplification of the protective innate immune response driven by inflammatory monocytes .
The life cycle of Toxoplasma is heteroxenous with a sexual definitive cycle in the felid host and a second intermediate life cycle in any endothermic animal including humans . The steps of the intermediate life cycle leading to tissues cysts in murine brain tissue illustrate this developmental process [4] . Bradyzoite/sporozoite oral infection leads to population wide development of the tachyzoite stage [47 , 48] that is followed by systemic spread of tachyzoites . In particular , spread into the vasculature resulting in the infection of endothelial cells of brain capillaries is a critical route for tachyzoites to cross the BBB into the brain [49] . Through poorly understood mechanisms , the tachyzoites slow growth [4] and alter their transcriptome to form dormant bradyzoite-tissue cysts in neurons [50] setting the stage for transmission to the next host animal . Thus , there are two competing demands of the Toxoplasma intermediate life cycle; expand tachyzoite numbers to ensure systemic spread within a host [49] and produce the dormant bradyzoite-tissue cyst required for passing the infection onto a new host [4] . How Toxoplasma mechanistically balances these competing demands is not understood . However , clues are emerging from our studies of ApiAP2 factors ( see Fig 6 model ) . Early bradyzoite development is associated with the induction of six Toxoplasma ApiAP2 genes ( AP2Ib-1 , AP2IV-3 , AP2VI-3 , AP2VIIa-1 , AP2VIII-4 , AP2IX-9 ) that are not expressed in the tachyzoite [22 , 28] . Remarkably , these factors do not operate in the same direction . AP2IX-9 , is a stress-inducible repressor of bradyzoite gene expression [22] , while AP2IV-3 is a stress-induced transcriptional activator ( and likely AP2Ib-1 ) regulating many of the same bradyzoite genes as AP2IX-9 [28] . The studies here add unexpected new complexity to bradyzoite developmental gene expression . AP2IV-4 was the first transcription factor expressed in replicating tachyzoites whose major function is the regulation of tissue cyst formation . Thus , our studies have uncovered a complex ApiAP2 transcriptional network of repressors and activators competing at the interface of tachyzoite replication and early switching to regulate tissue cyst formation ( see Fig 6 model ) . Notably , we have not yet identified an ApiAP2 that exclusively operates late in bradyzoite development . For the ApiAP2 gene family , the most distinguishing feature of mature bradyzoites is the down regulation of many ApiAP2 factors [28] . Does the lack of ApiAP2 factors specific for mature bradyzoites mean that once initiated bradyzoite development in vivo progresses to maturity ? Answering this question will be challenging given the asynchrony of bradyzoite development . However , a recent analysis of tissue cyst biology provides two important insights; tissue cyst size in the infected murine brain is related to tachyzoite vacuole size at the time of switching and average cyst numbers in tissues like murine brain become stable after an early period [28] . Add to this the observation that tissue cyst recrudesence is rare in the brain of immune-competent animals [51] , returns the discussion to the critical importance of the tachyzoite stage for fulfilling the biotic demands of the intermediate life cycle . The discovery of AP2IV-4 highlights the concept that specific life cycle decisions begin upstream in the developmental pathway and provides insight into the mechanisms that link the tachyzoite cell cycle to bradyzoite development . Together , the cell cycle AP2IV-4 and the stress-inducible AP2IX-9 , comprise two independent levels of transcriptional regulation preventing bradyzoite development in Toxoplasma . This is convincing support for the hypothesis that tachyzoite growth is the primary driver of parasite biomass and through dissemination the tachyzoite finds suitable host cell environments for which to develop ultimately end-stage bradyzoites [4 , 22] . In addition , the overlap of gene regulatory targets between AP2IV-4 and AP2IX-9 ( S2 Dataset ) indicate there is some redundancy governing the induction of the bradyzoite developmental pathway indicating the importance of preventing premature commitment to the bradyzoite stage that leads to dormancy [22] . The earliest clues to unfolding bradyzoite development in Toxoplasma revolves around the central role asexual stage replication plays in the transition to growth-arrested end-stages [3 , 31 , 52] . DNA replication in the tachyzoite is required for bradyzoite development [52] and the tachyzoite is “poised” to enter the bradyzoite developmental pathway during each round of replication [3] . The sub-transcriptome of the tachyzoite S and mitotic phases is enriched in basal bradyzoite transcripts [5] and developing populations have more 2N parasites [3] , which is a cell cycle timing that corresponds with peak AP2IV-4 expression . This places AP2IV-4 perfectly within the tachyzoite cell cycle to regulate these critical developmental processes . Repressing tissue cyst wall formation in the tachyzoite could provide the parasite with flexibility to maintain a replicative stage or quickly interpret “development” signals in the animal resulting in induction of bradyzoite differentiation when the parasite encounters the immune system and/or a tissue that favors tissue cyst longevity . In addition to controlling when and where tissue cyst formation occurs , repressors like AP2IV-4 may need to be re-expressed for pre-bradyzoites or bradyzoites to recrudesce . Consistent with this idea , our previous studies demonstrated bradyzoites from murine brain cysts re-express tachyzoite antigens prior to their first division in HFF cells and most bradyzoites that failed to re-express them did not divide [3] . These studies also demonstrate that deletion of a single ApiAP2 factor in Toxoplasma can significantly alter the course of the host immune response . Thus , host influences on ApiAP2 evolution has likely led to mechanisms that suppress bradyzoite antigens during acute infection , which we show here is required to establish a chronic infection in the murine brain . There are implications from this discovery for future vaccine development that might block tissue cyst formation in food animals , and thereby eliminate this source of human infections , which is an unmet therapeutic challenge . Our results point to inflammatory monocytes as a major component of the immune response contributing to protective immunity in the absence of AP2IV-4 ( Fig 5 ) [45 , 53] . A rapid response to the signature of a fast replicating lytic parasite is appropriate but there would be little evolutionary drive to respond equivalently to a slow replicating cyst form . Thus , changing the signatures of the parasite as we have done here with the deletion of AP2IV-4 have dramatically altered the early immune response , with bradyzoite antigens now being seen in the context of significant cell lysis . Increased inflammatory monocyte recruitment may point to a change in the ability of the parasite to be seen by the innate immune response . This could be either increased TLR recognition of bradyzoite antigens or a failure to inhibit signaling pathways by the tachyzoite . Perhaps predictably in the presence of enhanced monocyte recruitment , the T cell response is also superior , and the adaptive immune response to these parasites will be of significant interest in future studies . Alternatively , the mis-timing of bradyzoite antigen expression in the replicating and systemic tachyzoites lacking AP2IV-4 may increase the overall function of the immune response by either targeting it more rapidly to the bradyzoite or act as an adjuvant to the overall anti-parasite response . Further studies will be needed to fully understand the molecular mechanism ( s ) responsible for the shift to a more protective immune response . It is worth noting that the functions now emerging for AP2IV-4 in controlling in vivo persistence were not uncovered by cell culture models . Achieving bradyzoite switching in vitro in the mid-90's was a major breakthrough , and much has and will be learned using these models [54 , 55] . However , the complexity of parasite encounters with host cells and tissues in animals cannot be replicated by these models . Distinct tissue tropisms observed for tissue cyst formation in animals infected with Toxoplasma [47 , 56–58] suggest the parasite senses different host cell environments and relays this information to the mechanisms controlling developmental switching [4] . We know little about the molecular basis for Toxoplasma host tissue tropisms , however , it is likely that the network of ApiAP2 repressors and activators we have discovered will have critical roles in these host-parasite interactions .
Toxoplasma tachyzoites were serially passaged in vitro using confluent tissue culture flasks ( T25cm2 and T175cm2 ) containing human foreskin fibroblasts ( HFF cells; obtained from ATCC , Manassas , VA ) . Parasites were grown in confluent HFF cells and prepared for immunofluorescence as previously described [60] . Primary antibodies were used at the following concentrations: HA ( rat mAb 3F10 , 1:500 , Roche ) ; ISP1 ( mouse mAb clone 7E8 , 1:2000 , Dr . Peter Bradley , University of California , Los Angeles ) ; IMC1 ( mouse mAb , 1:1000 , Dr . Gary Ward , University of Vermont ) ; biotin-labeled Dolichos biflorus agglutinin ( DBA ) ( Vector labs , CA , 1:3000 ) ; BPK1 , MCP4 and SAG1 ( mouse polyclonal antibodies = pAbs , 1:1000 , Dr . John Boothroyd , Stanford University , ) ; CST1 ( mouse pAb , Salmon E , 1:2000 , Dr . Louis Weiss , Albert Einstein College of Medicine ) ; SRS9 ( rabbit pAb , 1:1000 , Dr . Laura Knoll , University of Wisconsin ) . Secondary antibodies by Alexa or streptavidin conjugated secondary antibodies were used at a 1:1000 dilution . All images were collected with a Zeiss Axiovert microscope equipped with 100x objective . Statistical significance was calculated using the one-tailed t-test , p values as indicated . Protein from 25x106 parasites were isolated , purified and whole parasite lysates collected as previously described [60] and subjected to electrophoresis on a SDS-PAGE gel . After transfer to nitrocellulose , the blots were probed with primary antibodies for CST1 ( mouse pAb , Salmon E , 1:2000 , Dr . Louis Weiss , Albert Einstein College of Medicine ) and the loading control TgNF3 ( mouse pAb , 1:1000 , Dr . Stan Tomavo , Pasteur Institute , Lille ) [61] . Detection of the proteins was completed using HRP conjugated antibodies ( Jackson ImmunoResearch ) followed by chemiluminescence reaction for visualization . Two independent biological replicates of total RNA were isolated from five RHCre transgenic strains ( S3 Dataset ) : RHCre-AP2IV-4floxed clone ID6 ( RHCre-parent ) , RHCre-Δap2IV-4 clones 27 and 30 , and cosmid complemented populations ( PSBM794; RHCre-Δap2IV-4::AP2IV-4 ) of each knockout . Likewise , two biological replicates were isolated from the following PruQ strain transgenics ( 3 total strains ) : PruQ-AP2IV-4floxed clone C3 ( PruQ-parent ) , PruQ-Δap2IV-4 clones 10 and 34 . RNA quality for all strains was evaluated using the Agilent Bioanalyzer 2100 ( Santa Clara , CA ) and 500ng of total RNA was prepared for hybridization on the ToxoGeneChip as described [5] . The resulting data was analyzed using GeneSpring GX software ( v11 . 5 , Agilent ) and all microarray data made available in the Gene Expression Omnibus ( GSE93531 ) . The AP2 domain ( amino acids 782–854 ) of AP2IV-4 was cloned into pGEX4T3 and expressed as a GST-fusion protein . Following affinity purification on a glutathione column , purified GST-AP2IV-4 protein was subjected to protein binding microarrays as previously described [9 , 22] . Complementary oligonucleotides were annealed to create 5’-biotinylated DNA probes of 59bp ( WT ) and 60bp ( scrambled ) . All binding reactions contained 20fmol DNA probe and 50ng of GST-AP2IV-4 protein . Non-biotinylated “cold” competitor probe was added at 300x concentration . GST-AP2IV-4-DNA complexes were resolved on a 6% polyacrylamide PAGE gel , transferred to a nylon membrane and interactions visualized using the LightShift Chemiluminescent EMSA kit as described by the manufacturer ( Thermo Fisher , Waltham , MA ) . Oligonucleotide sequences used for GST-AP2IV-4 expression and DNA probes can be found in S3 Dataset . Chromatin immunoprecipitation followed by quantitative PCR ( ChIP-qPCR ) was performed as previously published ( supplement of ref . [22] ) . In brief , RH-DDAP2IV-4 and RHΔhxgprt ( negative control ) parasites were inoculated at 3:1 MOI in T175 cm2 flasks , allowed to invade for 1 h , rinsed three times with Hanks balanced salt solution ( Gibco ) to remove free floating parasites and fresh media containing 100nM Shield 1 was added . Parasite cultures were allowed to grow 36 h prior to intracellular crosslinking with formaldehyde and isolation of nuclear fraction . Nuclear material was subjected to sonication to shear DNA into 200-1000bp fragments and soluble fraction incubated with α-HA antibody ( 5μg , ab9110 , rabbit , Abcam ) . Protein-DNA complexes were isolated using protein-G coupled magnetic beads ( Dynabeads , Invitrogen ) and DNA isolated by treatment with 1% SDS followed by phenol-chloroform extraction and ethanol precipitation . Whole genome amplification ( Sigma-Aldrich ) was performed on ChIP-DNA and purified by Qiagen Mini-Elute PCR kit . qPCR was performed using 20ng/rxn of specific ( DDAP2IV-4 ) chromatin and non-specific chromatin ( RHΔhxgprt ) using Fast SYBR green master mix on an ABI7900 according to manufacturer’s protocols . Relative enrichment was calculated with the equation: 2^- ( ΔCt target-ΔCt non-target ) where the change in Ct value of specific versus nonspecific chromatin at the SRS9 and BAG1 promoters was calculated . All ChIP-qPCR oligonucleotides used can be found in S3 Dataset . Mice were purchased from Jackson or Harlan Laboratories . 10–12 week old BALB/c mice were infected with 1x107 PruQ-parental strain or PruQ-Δap2IV-4 intraperitoneally in sterile PBS . Uninfected , age-matched mice were used as naïve uninfected controls . Mice were monitored daily and euthanized at day one and day six-post infection for study . To examine acute virulence and tissue cyst formation , 5–6 week old female BALB/c mice were injected intraperitoneally with 105 , 106 , or 107 PruQ-parent or PruQΔap2IV-4 parasites ( 4 mice per group , 107 dose only shown in Fig 4 ) . Plaque assays were performed for each sample and ensured equal viability between strains . Mice were examined daily and time to death was recorded . Serology performed on cardiac bleeds of infected mice confirmed presence of Toxoplasma . To assess cyst burden , BALB/c mice were infected with 1x106 parasites as described above and allowed to progress to chronic infection for 30 days ( 4 mice per group ) . Brains were then homogenized; homogenates were fixed , quenched , and permeabilized . Samples were blocked in 3% BSA/1xPBS/0 . 2% Triton X-100 . To visualize cyst walls , rhodamine-conjugated Dolichos biflorus lectin ( Vector Labs ) was applied at 1:250 dilution overnight at 4°C . Cyst quantification was performed as previously described [62] . Following euthanasia , peritoneal exudate cells ( PECs ) were collected from the peritoneal cavity in sterile PBS . Cells were counted using an automated cell counter , and total cell numbers were determined using the volume recovered from the peritoneum . Aliquots were used for cytospins , and stained using HEMA3 stains . For flow cytometry , cells were incubated with 1:10 FC Block ( BD , 553142 ) for 5 minutes on ice , and subsequently incubated with fluorophore-conjugated antibodies to CD45 , CD11b , CD11c , Ly6C , Ly6G . Cells were washed , resuspended in FACS buffer and samples acquired using a BD FACS Canto II flow cytometer . Analysis was conducted using Flowjo software . Peripheral tissues were placed in tissue lysis buffer for DNA isolation . DNA was isolated using a genomic DNA purification kit ( Roche , 11796828001 ) . To quantify parasite burden , quantitative PCR ( Bioline ) was conducted on isolated DNA by amplification of the Toxoplasma B1 gene ( F: 5’ TCCCCTCTGCTGGCGAAAAGT 3’ R: 5’ AGCGTTCGTGGTCAACTATCG 3’ ) . Parasite burden was quantified using a standard curve as previously described [63] .
All animal research was conducted in accordance with the animal welfare act , and all protocols were approved by the Institutional Animal Care and Use Committees at the University of California , Riverside ( approved protocol #A-20140007 ) or the Indiana University School of Medicine ( approved protocol #10852 ) .
|
The Toxoplasma biology that underlies the establishment of a chronic infection is developmental conversion of the acute tachyzoite stage into the latent bradyzoite-tissue cyst stage . Despite the important clinical consequences of this developmental pathway , the molecular basis of the switch mechanisms that control formation of the tissue cyst is still poorly understood . A fundamental feature of tissue cyst formation is the expression of bradyzoite-specific genes . Here we show the transcription factor AP2IV-4 directly silences bradyzoite mRNA and protein expression in the acute tachyzoite stage demonstrating that developmental control of tissue cyst formation is as much about when not to express bradyzoite genes as it is about when to activate them . Losing the suppression of bradyzoite gene expression in the acute tachyzoite stage caused by deleting AP2IV-4 blocked the establishment of chronic disease in healthy animals via increased protective immunity suggesting a possible strategy for preventing chronic Toxoplasma infections .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods",
"Ethics",
"statement"
] |
[
"parasite",
"groups",
"medicine",
"and",
"health",
"sciences",
"vacuoles",
"parasitic",
"cell",
"cycles",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"dna-binding",
"proteins",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"parasitology",
"developmental",
"biology",
"apicomplexa",
"tachyzoites",
"protozoans",
"toxoplasma",
"cellular",
"structures",
"and",
"organelles",
"proteins",
"gene",
"expression",
"life",
"cycles",
"biochemistry",
"eukaryota",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"organisms",
"parasitic",
"life",
"cycles"
] |
2018
|
Transcriptional repression by ApiAP2 factors is central to chronic toxoplasmosis
|
Although of fundamental importance in developmental biology , the genetic basis for the symmetry breaking events that polarize the vertebrate oocyte and egg are largely unknown . In vertebrates , the first morphological asymmetry in the oocyte is the Balbiani body , a highly conserved , transient structure found in vertebrates and invertebrates including Drosophila , Xenopus , human , and mouse . We report the identification of the zebrafish magellan ( mgn ) mutant , which exhibits a novel enlarged Balbiani body phenotype and a disruption of oocyte polarity . To determine the molecular identity of the mgn gene , we positionally cloned the gene , employing a novel DNA capture method to target region-specific genomic DNA of 600 kb for massively parallel sequencing . Using this technique , we were able to enrich for the genomic region linked to our mutation within one week and then identify the mutation in mgn using massively parallel sequencing . This is one of the first successful uses of genomic DNA enrichment combined with massively parallel sequencing to determine the molecular identity of a gene associated with a mutant phenotype . We anticipate that the combination of these technologies will have wide applicability for the efficient identification of mutant genes in all organisms . We identified the mutation in mgn as a deletion in the coding sequence of the zebrafish microtubule actin crosslinking factor 1 ( macf1 ) gene . macf1 is a member of the highly conserved spectraplakin family of cytoskeletal linker proteins , which play diverse roles in polarized cells such as neurons , muscle cells , and epithelial cells . In mgn mutants , the oocyte nucleus is mislocalized; and the Balbiani body , localized mRNAs , and organelles are absent from the periphery of the oocyte , consistent with a function for macf1 in nuclear anchoring and cortical localization . These data provide the first evidence for a role for spectraplakins in polarization of the vertebrate oocyte and egg .
The animal-vegetal axis is the first axis to form in the vertebrate embryo; however , the cellular and genetic pathways by which it is specified are the least well understood . In lower vertebrates , such as frogs and fish , this axis forms during early oogenesis , while in mouse , it becomes apparent during oocyte maturation . The earliest morphological marker of asymmetry in the vertebrate oocyte is the Balbiani body , a transient structure composed of organelles , including mitochondria , endoplasmic reticulum ( ER ) and Golgi ( reviewed in [1] and [2] ) . The Balbiani body is found in the oocytes of invertebrates such as Drosophila [3] and vertebrates , including Xenopus , mouse and human [4]–[9] . In most mammals , the molecular composition and function of the Balbiani body is unknown . In mouse , however , the Balbiani body contains Trailer hitch ( Tral ) protein , which has been implicated in mRNA localization and translation in Drosophila and in P body assembly in human cells [10] , [11] . The presence of Tral in the Balbiani body of the mouse suggests a function for this structure in RNA metabolism [6] . Studies of the Xenopus Balbiani body have shown that in addition to ER , mitochondria and Golgi , it contains germ plasm RNAs and germinal granules , and it is thought to localize these factors to the vegetal cortex of the oocyte during early oogenesis [5] , [12] , [13] ( reviewed in [2] ) . Recent studies have described a Balbiani body in zebrafish oocytes that behaves similarly to the Xenopus Balbiani body [14]–[17] . In both zebrafish and Xenopus , the Balbiani body is the first morphological marker of polarity and predicts the future animal-vegetal axis . In primary oocytes , it forms adjacent to the oocyte nucleus on the future vegetal side of the oocyte and then becomes localized to the vegetal cortex as oogenesis proceeds [5] , [14]–[18] . Upon localization to the vegetal cortex , germ plasm RNAs and germinal granules are deposited and the Balbiani body disassembles . Although a highly conserved structure , the mechanisms of Balbiani body formation , function , and disassembly have remained elusive , in large part due to a lack of genetic and molecular data . The zebrafish bucky ball ( buc ) gene is the only vertebrate gene known to be required for Balbiani body formation [15] , [17] . buc encodes a novel protein , in the absence of which a Balbiani body fails to form . In buc mutants , vegetal RNAs are not localized , reflecting a defect in animal-vegetal polarity of the oocyte [15] , [17] . We report a second zebrafish gene required for animal-vegetal polarity of the oocyte and egg . The magellan ( mgn ) mutant was identified based on a defect in which cytoplasm localizes around the yolk rather than at the animal pole of the egg [19] . We found that during oogenesis , mgn mutant oocytes display an asymmetric localization of the oocyte nucleus , a novel enlarged Balbiani body phenotype and an absence of vegetally-localized RNAs , stable microtubules , and organelles at the oocyte cortex . To identify the gene disrupted in mgn mutants , we positionally cloned the mgn gene , utilizing a novel DNA capture method to enrich for genomic DNA spanning the interval containing our candidate gene . This technique involves hybridization of region-specific oligonucleotides to long genomic DNA fragments , extension of the oligonucleotides using labeled nucleotides , capture of the labeled fragments along with the genomic DNA template and flanking regions , and finally , isolation of the captured genomic DNA , which is then processed for massively parallel sequencing . Using this technique , we identified a 31 base pair deletion in the zebrafish ortholog of microtubule actin crosslinking factor 1 ( macf1 ) , a highly conserved cytoskeletal linker protein belonging to the spectraplakin family of proteins . This is one of the first examples in which genomic DNA enrichment combined with massively parallel sequencing has been used to determine the molecular identity of a gene associated with a phenotype . macf1 function has been characterized in several types of polarized cells including epithelial cells , neurons , and muscle cells . Our analysis of the mgn mutant reveals a new role for macf1 in the oocyte , providing insight into the role of spectraplakins during vertebrate oogenesis .
In zebrafish , animal-vegetal polarity of the egg becomes morphologically apparent upon egg activation . Prior to activation , the cytoplasm of the egg is intermingled with the yolk . Activation of the egg through contact with water causes two striking changes to the egg: the cortical granules fuse with the plasma membrane and exocytose their contents into the perivitelline space causing the chorion to expand; and the egg contracts , resulting in segregation of cytoplasm from the yolk to the animal pole to form the blastodisc ( Figure 1A ) . The zebrafish mgn mutant was identified based on a defect in activated eggs . We found that mgn mutant eggs exhibit variable expansion of the chorion . In addition , we observed that following cytoplasmic segregation , cytoplasm was variably distributed around the yolk rather than being restricted to one pole ( Figure 1B ) , suggesting a defect in animal-vegetal polarity of the egg . To determine if the mgn mutation causes defects during oogenesis , we examined mutant and wild-type oocytes by histology , using hematoxylin and eosin dyes to visualize structures such as the Balbiani body , cortical granules and yolk . In wild-type mid to late stage I oocytes , the nucleus was centrally located and the Balbiani body , which forms on the future vegetal side of the nucleus and translocates to the vegetal cortex , was detected either near the nucleus ( n = 2/15 ) , between the nucleus and future vegetal cortex ( n = 6/15 ) , or at the vegetal cortex ( Figure 2A; n = 7/15 ) . In contrast , the nucleus of mutant oocytes was asymmetrically localized ( Figure 2B ) and the Balbiani body was only infrequently found at the cortex ( n = 2/20 ) . In the majority of mutant oocytes , the Balbiani body was observed near the nucleus ( Figure 2B; n = 12/20 ) or between the nucleus and oocyte cortex ( n = 6/20 ) . In addition , the Balbiani body of mutant oocytes was surrounded by pale pink staining that appeared speckled . This staining was not observed in wild-type oocytes . The asymmetric positioning of the oocyte nucleus was also seen in stage II mutant oocytes , whereas in wild-type oocytes , the nucleus remained centrally located ( compare Figure 2C and 2D ) . In addition , we found that in wild-type stage II oocytes , cortical granules were located radially around the central nucleus ( Figure 2C; n = 25/25 ) , whereas in mutant oocytes , cortical granules were asymmetrically located and were found in the middle of the oocyte , including within and around a region distinct to mutant oocytes that stains lightly with eosin ( Figure 2D; n = 31/31 ) . Stage III of oogenesis is characterized by an accumulation of yolk in the middle of the oocyte , which is thought to drive the cortical granules to the oocyte cortex ( Figure 2E; n = 17/17 ) [20] . In stage III mgn mutant oocytes , as in wild-type , yolk accumulated in the oocyte and the cortical granules typically moved to the oocyte periphery . In mutant oocytes , however , there was a variable and uneven distribution of cortical granules around the cortex and a small number of cortical granules remained within the yolk ( Figure 2F; n = 22/23 ) , likely due to their abnormal localization during stage II . These defects in cortical granule localization may result in the variable chorion expansion defect in mgn mutants ( Figure 1 ) . In zebrafish , as in Xenopus , the Balbiani body contains germ plasm components that include germ plasm RNAs and germinal granules [14] , [15] , [17] . Studies in Xenopus indicate that the Balbiani body is required for localization of mRNAs to the vegetal cortex [12] , [13] , ( reviewed in [2] ) and recent work in zebrafish shows that distinct localization elements within the 3′UTR of dazl mRNA localize it to the Balbiani body and vegetal cortex of the oocyte [14] . To investigate the function of the Balbiani body in mgn mutant oocytes , we examined the localization of dazl and a second conserved germ plasm mRNA , vasa . In wild-type oocytes , vasa mRNA is localized to the Balbiani body during early stage I of oogenesis , transiently localizes to the vegetal cortex during late stage I , and then becomes localized around the cortex of the oocyte during stage II ( Figure 3A and 3B ) . In mgn mutant oocytes , vasa mRNA localized to the Balbiani body during early stage I , but it was predominantly localized to the middle of the oocyte during late stage I and stage II ( Figure 3C and 3D; n = 15/16 ) . In wild-type oocytes , dazl mRNA is localized to the Balbiani body during early stage I and then becomes localized to the vegetal cortex during late stage I where it remains during stage II ( Figure 3E and 3F ) . In mutant oocytes , dazl mRNA initially localized to the Balbiani body during stage I , and like vasa mRNA , was predominantly found in the middle of the oocyte in late stage I and stage II , rather than localized to the vegetal cortex ( Figure 3G and 3H; n = 24/25 ) . These results show that germ plasm mRNAs localize to the Balbiani body of mutant oocytes during early stage I and that Mgn is required for the subsequent localization of the Balbiani body and associated RNAs to the vegetal cortex during late stage I . In addition , the persistence of both vasa and dazl mRNA in a central domain of stage II mutant oocytes may reflect a failure of the Balbiani body to disassemble at the end of stage I of oogenesis . Using SSLP ( simple sequence length polymorphism ) markers in bulk segregant analysis , followed by fine mapping of homozygous mutant versus heterozygous sibling females using SSLP and SNP ( single nucleotide polymorphism ) markers , we localized the mgn mutation to a 2 . 02 Mb interval ( Figure 4F ) based on the Sanger Institute Ensembl zebrafish genome sequence assembly ( Zv7/danRer5 ) . Based on data from 976 meioses and 27 recombination events , we then calculated that the molecular lesion would be in the zebrafish macf1 gene . To date , a full-length zebrafish macf1 transcript has not been sequenced , however , the longest predicted zebrafish macf1 transcript is 23 , 975 base pairs ( bp ) ( NCBI RefSeq accession: XM_001920059 . 1 ) . Additionally , while macf1 is highly conserved , the frequency of multiple , alternatively spliced isoforms in other organisms [21] suggested that prediction of ovarian transcript sequences would be difficult . We therefore chose to utilize increasingly affordable next-generation sequencing technologies to sequence the entire 298 , 606 bp predicted macf1 genomic interval ( Zv7/danRer5 , Ensembl 52 genebuild ) as well as approximately 17 kb of sequence 5′ to the predicted gene . To do this economically , we had to enrich for genomic sequence specific to our target macf1 region . This was accomplished using a genomic sequence capture and release method . This method involves designing 22–27mer long region-specific oligonucleotides ( oligos ) that are hybridized to genomic DNA , extended using biotinylated nucleotides , and then affinity purified using streptavidin coated magnetic beads to capture the genomic template DNA ( modification of method in [22] ) . The captured DNA is then dissociated from the beads , sheared , and cloned into a library for massively parallel sequencing ( Figure 4A ) . We designed oligos approximately every 8 kb spanning a 319 . 8 kb region , performed the capture and sequencing , and then analyzed a 315 . 6 kb region that included all of macf1 and ∼17 kb of sequence 5′ to the gene . We obtained a 243-fold average enrichment for this region , as determined from the sequence analysis . The sequencing results also showed a 31-fold average read depth for unique sequence , covering 99 . 9% of our targeted region ( Figure 4B ) . Data were obtained for the entire predicted coding sequence with the exception of three small gaps . The genomic intervals surrounding these gaps were then amplified by PCR and sequenced by Sanger sequencing . A four bp and 12 bp gap were not confirmed by Sanger sequencing and no base changes were found in these intervals . Sanger sequencing did , however , confirm a gap of 31 bp . The 31 bp deletion causes a frame-shift at amino acid codon 5315 of the longest predicted zebrafish Macf1 protein isoform ( NCBI RefSeq accession: XP_001920094 . 1 ) , resulting in a premature stop codon and subsequent truncation of the predicted protein ( Figure 4C ) . One additional base change in predicted coding sequence was found by massively parallel sequencing but was not confirmed by Sanger sequencing . Based on the mgn mutant phenotype , we expected macf1 to be expressed during stage I of oogenesis . In situ hybridization experiments showed that , consistent with a function during early oogenesis , macf1 mRNA is expressed in stage I oocytes and additionally , has a restricted localization pattern during later oogenesis ( Figure 4D ) . In contrast , in situ hybridization analysis showed that in mgn mutant oocytes , macf1 mRNA levels appeared to be significantly reduced ( Figure 4E ) . This suggests that the transcript is subject to nonsense-mediated decay and is consistent with the presence of a premature stop codon in macf1 caused by the deletion in mgn mutants . To confirm that the deletion in macf1 is the only change in coding sequence in the interval to which the mgn mutation maps , we further narrowed the interval and then sequenced the remaining candidate genes . We initially attempted to identify SNPs to narrow the interval by PCR amplification and sequencing of non-coding regions , but we were unsuccessful in identifying polymorphisms . Therefore , we performed a second sequence capture experiment with wild-type and mutant fish from a new map cross to attempt to more efficiently identify any rare SNPs in the region and examine additional genes in the interval . For the second sequence capture , we designed oligos to a 613 . 9 kb region that consisted of the macf1 gene plus approximately 86 kb of sequence 5′ to macf1 and 230 kb of sequence 3′ to macf1 ( Figure 4F ) . We massively parallel sequenced the captured genomic DNA from a wild-type ( AB strain ) fish and a mutant ( TU strain ) fish and analyzed 616 kb of targeted sequence and surrounding sequence that was captured . For the wild-type sample , we obtained a 72-fold average enrichment and a 29-fold average read depth for unique sequence , covering 94% of our targeted region ( Figure 4G ) . For the mutant sample , we obtained a 43-fold average enrichment and a 17-fold average read depth for unique sequence , covering 97% of our targeted region ( Figure 4G ) . Using these data , we identified a SNP ( called SNP36 ) 14 . 35 kb from the 5′ end of macf1 . No genes are known or predicted to exist in this 14 . 35 kb region . We were unable to identify useful SNPs 3′ to the macf1 gene using the sequence capture data; therefore , we used a previously identified SSLP marker ( z53477 ) to narrow the interval 3′ to macf1 . We identified one recombinant fish from 311 meioses at SNP36 and 10 recombinants from 578 meioses at z53477 . The interval between z53477 and SNP36 contains part of si:dkey-190l1 . 2 ( similar to vertebrate gamma-aminobutyric acid B receptor , 2 ) , as well as the full sequence of 13 genes . Ten genes including macf1 are present in the second sequence capture interval ( sla1 ( Src-like-adaptor 1 ) , wisp1a ( novel protein similar to vertebrate WNT1 inducible signaling pathway protein 1 ) , ndrg1 ( N-myc downstream regulated gene 1 ) , si:rp71–45k5 . 2 ( novel protein ) , zgc:113424 ( novel forkhead domain-containing protein ) , si:rp71–45k5 . 3 ( novel protein ) , si:rp71–45k5 . 4 ( novel protein similar to vertebrate proteasome subunit , alpha type 2 ) , zgc:91910 ( novel protein ) , bmp8 ( bone morphogenetic protein 8 ) and macf1 ) . Examination of the coding and predicted coding sequences revealed that there were no sequence changes in any of the genes in the interval . Three genes in the mgn interval that were not included in the captured region were cloned from ovarian cDNA ( si:dkey-190l1 . 1 ( elongator complex protein 2 ) , si:ch211–254e15 . 2 ( UPF0436 protein C9orf6 homolog ) , si:ch211–254e15 . 1 ( novel protein similar to vertebrate catenin , alpha-like 1 ) ) and were sequenced by Sanger sequencing . No sequence changes were found in the predicted coding regions of these genes . Based on these data , we conclude that the mgn mutant phenotype is caused by the deletion in macf1 . macf1 belongs to the highly conserved spectraplakin family of proteins . Spectraplakins are multifunctional cytoskeletal linker proteins characterized by an N-terminal actin binding domain consisting of Calponin homology domains , a globular plakin domain that mediates interactions with various types of cell junctions , a plakin repeat domain that can interact with intermediate filaments , a spectrin repeat domain that allows dimerization , and finally a C-terminal microtubule binding domain that consists of an EF-hand and GAS2 domain [23] , [24] . The mgn deletion creates a stop codon in the spectrin repeat domain that is predicted to result in a loss of part of the spectrin domain and all of the microtubule binding domain ( Figure 4H ) . Thus , the truncated protein is not expected to bind to the microtubule cytoskeleton . It is well established that spectraplakins such as macf1 are able to interact with both the actin and microtubule cytoskeletons ( reviewed in [21] , [24] ) . To determine if the cytoskeleton is affected in mgn mutant oocytes , we first used an antibody to acetylated tubulin to visualize stable microtubules . We found that in wild-type stage I oocytes , stable microtubules were present throughout the oocyte but did not appear to be attached to a microtubule organizing center ( MTOC ) ( Figure 5A and 5A′; n = 13/13 ) . In zebrafish as in other organisms , the centrosome , which serves as an MTOC , is inactivated early in oogenesis . Stable microtubules were also present in mgn mutant oocytes , but their presence was significantly reduced at the periphery of the oocyte ( Figure 5B and 5B′; n = 23/24 ) . We next examined the actin cytoskeleton in wild-type and mgn mutant oocytes . In both wild-type and mutant oocytes , we found actin filaments localized to the nucleus and at the oocyte cortex ( Figure 5C and 5D , n = 15 wild-type and n = 13 mutant oocytes ) . Thus , the mutation in macf1 causes a loss of stable microtubule localization to the periphery of the oocyte , but it does not appear to affect the actin cytoskeleton . These results suggest that , consistent with loss of the predicted Macf1 microtubule binding domain , the defects seen in mgn mutants may be caused by a loss of tethering of stable microtubules to the oocyte cortex . Given the defect in cortical localization of RNAs and stable microtubules , we investigated if mgn mutant oocytes also exhibit defects in the localization of organelles by labeling ER and mitochondria with the membrane dye , DiOC6 . In addition to cytoplasmic ER and mitochondria , DiOC6 labels the aggregate of ER and mitochondria that , along with germ plasm mRNAs and other organelles , composes the Balbiani body ( reviewed in [1] , [2] ) . DiOC6 staining revealed that in wild-type oocytes , ER and mitochondria are found throughout the oocyte and in addition , are concentrated within the Balbiani body , which translocates from the nucleus to the oocyte cortex during mid to late stage I ( Figure 6A , n = 80/80 ) . In mutant oocytes , ER and mitochondria , as well as the Balbiani body itself , are typically absent from the periphery of the oocyte ( Figure 6B and 6C , n = 46/64 ) . In many oocytes , ER and mitochondria were tightly concentrated around the nucleus of the oocyte , as well as absent from the periphery ( Figure 6C; n = 24/64 ) . The defect in the peripheral localization of ER and mitochondria persists into stage II of oogenesis when in wild-type oocytes , the Balbiani body has disassembled and ER and mitochondria are localized throughout the oocyte ( Figure 6D ) . In mgn mutants , DiOC6 staining was concentrated in the middle of stage II mutant oocytes ( Figure 6E ) , consistent with a function for macf1 in the localization of ER and mitochondria to peripheral regions of the oocyte . The lack of ER and mitochondria localization to the periphery of mgn mutant oocytes is strikingly similar to the absence of acetylated tubulin in these regions ( Figure 5A and 5B ) . To confirm the absence of ER and mitochondria at the oocyte periphery observed with DiOC6 staining , we examined stage II oocytes by transmission electron microscopy ( TEM ) . TEM revealed a uniform distribution of ER and mitochondria throughout wild-type oocytes ( compare cytoplasmic and peripheral domains in Figure 7A′ and 7A″; n = 4 ) . In mgn mutant oocytes , the central cytoplasmic domain was similar to that of wild-type oocytes ( Figure 7B′; n = 7 ) ; however , in peripheral regions , very little ER and no mitochondria were present ( Figure 7B″ ) . These TEM results are consistent with our observations from the DiOC6 staining and further support a role for macf1 in the peripheral localization of organelles . In addition to the frequent loss of cortical Balbiani body localization , DiOC6 staining revealed a difference in Balbiani body size in mutant compared to wild-type oocytes . We found that in early stage I mutant oocytes , the Balbiani body was comparable in size to that of wild-type oocytes ( Figure 6F; 50–70 micron diameter ) ; however , in mid to late stage I oocytes ( 70–110 micron diameter ) the Balbiani body of mutant oocytes was significantly larger than in wild-type oocytes ( Figure 6F , average Balbiani body size = 12 microns versus 18 microns in 70–90 micron wild-type and mutant oocytes , respectively , and 18 microns versus 29 microns in 90–110 micron wild-type and mutant oocytes , respectively ) . Thus , the Bb is similarly sized in early stage I mgn mutant and wild-type oocytes but becomes abnormally large beginning in mid stage I of oogenesis in mgn mutants . The asymmetric localization of the nucleus in mutant stage I and II oocytes that we observed by histology was also apparent in DiOC6 stained oocytes ( Figure 6B ) . In 70–90 micron wild-type oocytes , the average distance between the centroid of the nucleus and the centroid of the oocyte was 6 . 3 microns ( n = 15 ) , whereas in 70–90 micron mutant oocytes , it was 12 microns ( n = 13 ) . Thus , there is a significant ( p<0 . 05 ) difference in the position of the nucleus in mutant mid stage I oocytes . It is not yet clear if the mislocalized nucleus causes the defect in Balbiani body size , is a consequence of it , or is an independent defect of mgn mutant oocytes .
The mechanisms guiding the establishment and maintenance of polarity in the vertebrate oocyte are not well understood . Here we have identified a novel function for macf1 in formation of the animal-vegetal axis during zebrafish oogenesis . macf1 belongs to the spectraplakin family of cytoskeletal linker proteins , which are highly conserved amongst species . Mutations in the Drosophila homolog of macf1 , short stop ( shot ) , were first identified in screens for neuromuscular specificity and axon guidance [25] , [26] and for integrin-mediated adhesion [27]–[29] . Subsequent work has defined several additional roles for shot in processes that include oocyte determination and tracheal cell fusion [30] , [31] . The diverse functions of shot have been attributed to the presence of multiple isoforms , and recent work has focused on defining tissue-specific domain requirements [32] . The function of macf1 in vertebrates has largely been characterized in cell culture experiments , since the mouse macf1 knockout results in embryonic lethality [33] , [34] . As in Drosophila , vertebrate macf1 has diverse functions . In Hela cells , for example , macf1 is required for the cortical localization of CLASP2 , a microtubule tip binding protein [35] . During mouse embryogenesis , macf1 is required for Wnt signaling , where it functions in translocating Axin to the cell membrane [34] . Our isolation of the zebrafish mgn mutant provides a unique opportunity to now study the role of macf1 during vertebrate oogenesis . We found that mgn is required during zebrafish oogenesis for formation of the animal-vegetal axis of the oocyte and egg . The only other vertebrate gene known to be required for this process is the zebrafish bucky ball ( buc ) gene , which encodes a novel protein [15] , [17] . In buc mutants , a Balbiani body does not form and RNAs are mislocalized , leading to a defect in animal-vegetal polarity of the oocyte and egg [15] , [17] . Similar to buc mutants , mgn mutant females produce eggs that display abnormal animal-vegetal polarity . In mgn mutant oocytes , however , a Balbiani body forms during early stage I as in wild-type oocytes , but it becomes abnormally large and typically does not localize to the oocyte cortex during mid stage I as in wild-type . Interestingly , a null mutation in Kinesin heavy chain or P-element insertions in the kinesin-associated mitochondrial adaptor , milton , which cause its overexpression , result in overgrowth of the Balbiani body in the Drosophila ovary [36] . Although similar to mgn mutants , the Balbiani body defects in these milton mutants affect the localization of mitochondria to the Balbiani body without disrupting RNA localization and are largely resolved during late oogenesis when the microtubule cytoskeleton reorganizes [36] . Thus , mgn is the only gene known to be required for regulating the size of the Balbiani body , as well as for localization of the Balbiani body and its associated mRNAs to the oocyte cortex . Macf1 may be directly required for these processes through binding to cytoskeletal elements within the Balbiani body . Alternatively , misregulation of Balbiani body size and localization in mgn mutants may be caused by a disruption of the stable microtubule cytoskeleton that prevents localization of the Balbiani body to the vegetal cortex of the oocyte . Analysis of mid stage I mgn mutant oocytes revealed an asymmetric localization of the nucleus that is not seen in wild-type oocytes at this stage and which persists into stage II . It is possible that this mislocalization of the oocyte nucleus is a consequence of the Balbiani body enlargement; however , it is also plausible that the nuclear mislocalization represents an independent defect in nuclear anchoring . Establishment and maintenance of a centralized nucleus is an active process , and it is well established that actin filaments , microtubules , and intermediate filaments play roles in nuclear migration and anchoring in various cell types ( reviewed in [37]–[39] ) . Nuclear anchoring is mediated by a family of proteins , all of which contain KASH ( Klarsicht , ANC-1 , Syne homology ) domains at their C-termini . The KASH domains target the proteins to the outer nuclear envelope , while the N-termini of these proteins tether them to the cytoskeleton ( reviewed in [40] , [41] ) . Interestingly , the KASH domain of Nesprin-3a interacts with the actin-binding domain of MACF1 in coimmunoprecipitation experiments in vitro [42] . It has recently been reported that vertebrate Macf1 isoform-3 localizes to the outer nuclear envelope and that the N-terminus associates with Nesprin-3 in COS-7 cells [43] . These data suggest that Macf1 may play a role in tethering the nucleus to the microtubule cytoskeleton . Such a function for Macf1 would be consistent with the defect in nuclear positioning that we observe in mgn mutant oocytes . In addition to the abnormally large Balbiani body and the nuclear positioning defect , DiOC6 staining and TEM revealed that the cytoplasmic ER and mitochondria that surround the Balbiani body were absent from the periphery of the oocyte . It has long been recognized that microtubules are essential for establishing and maintaining the structure of the ER ( reviewed in [44] ) and for localizing mitochondria to various cellular domains ( reviewed in [45] , [46] ) . Knockdown of the Kinesin-binding protein Kinectin in cultured cells results in a collapse of ER and mitochondria to the center of the cell [47] , demonstrating the involvement of motor proteins in the maintenance of ER architecture and mitochondrial localization . In addition , in cultured mouse cells mutant for the conventional kinesin heavy chain , kif5b , mitochondria display a perinuclear distribution and an abnormal absence from the cell periphery [48] . Several non-motor proteins have also been implicated in mediating ER-microtubule interactions ( reviewed in [44] , [49] ) . Climp-63 , for example , is an integral ER membrane protein that binds to microtubules , anchoring the ER to the cytoskeleton . Overexpression of a mutant Climp-63 lacking the microtubule binding domain results in displacement of endogenous Climp-63 and clustering of ER around the nucleus [50] . These phenotypes are similar to those of mgn mutant oocytes , in which ER and mitochondria are frequently concentrated around the nucleus and are absent from the periphery . This suggests that Macf1 may be required to maintain ER architecture and to localize mitochondria in the oocyte by positioning stable microtubules at the oocyte periphery . Alternatively , Macf1 may bind directly to ER , mitochondria and microtubules . We found that mgn is required to localize stable microtubules to peripheral regions in zebrafish oocytes . Microtubule-cortex interactions are mediated by a growing class of microtubule-associated proteins called microtubule plus-end-tracking proteins ( +TIPS ) . +TIPS mediate cortical interactions through association of microtubule ends with the cortical actin cytoskeleton and with the plasma membrane ( reviewed in [51] , [52] ) . Macf1 ( Acf7 ) has been identified as a +TIP in mouse endodermal cells where it functions to coordinate the actin and microtubule cytoskeletons in the cytoplasm and at the cell cortex and to maintain polarity in response to wound healing [33] . In Drosophila tendon cells , Macf1 ( Shot ) organizes the microtubule network at the muscle-tendon junction by forming a complex with the +TIPs EB1 and APC [53] , which regulate cortical targeting of microtubules ( reviewed in [51] , [54] ) . Recent studies have shown that Macf proteins bind to EB1 directly and exhibit plus end tracking in vivo [55] . These data are consistent with a function for Macf1 in cortical localization of microtubules . Together , the loss of stable microtubules at the oocyte periphery and the absence of peripheral ER and mitochondria in mgn mutant oocytes suggest that in the zebrafish oocyte , ER and mitochondria associate with stable microtubules and that Macf1 may function to tether the stable microtubule cytoskeleton to the cortical actin cytoskeleton . In the absence of functional Macf1 , the microtubule cytoskeleton is disrupted and the ER and mitochondria collapse into the middle of the cell . To identify the molecular lesion associated with the mgn phenotype , we identified a candidate gene using positional cloning methods and then used a novel targeted sequence capture technique combined with massively parallel sequencing to sequence the candidate gene and exclude nine neighboring genes in the interval . Identification of mutations by traditional positional cloning techniques is time consuming and can be hindered by a variety of factors including limited genome annotation and complex genetic loci , which make prediction of transcripts difficult . As a result , there is increasing interest in utilizing next-generation sequencing technologies to identify mutations by sequencing large regions of genomic DNA . It was recently demonstrated that a point mutation in the encore gene in Drosophila could be identified by whole genome resequencing [56] . Although extremely efficient , this method would be prohibitively costly in an organism with a large genome , such as zebrafish or mouse . A more cost-effective approach was recently used to identify a mutation in the mouse Megf8 gene [57] . This approach utilized positional cloning techniques to identify a 2 . 2 Mb interval containing the mutation . The authors then constructed a BAC contig spanning this region and sequenced 15 BACs with massively parallel sequencing to identify a point mutation in Megf8 . While this method is effective and more cost-efficient than whole genome resequencing , the construction of BAC contigs remains labor intensive and time consuming . We sought to utilize an efficient and cost-effective method for region-specific enrichment of genomic DNA to allow for massively parallel resequencing of targeted genomic intervals . In recent years , several methods have been developed for genomic enrichment [58]–[66] . A method for in-solution hybrid selection was recently reported based on hybridizing long biotinylated RNA sequences to DNA that has been randomly sheared and amplified [66] . Oligonucleotide tiling arrays were recently used to target and identify mutations in the mouse kit gene , human neurofibromin gene , and in seven human autosomal recessive ataxia-associated genes [61]–[63] . In addition , array-based genomic sequence capture of a 40 Mb linkage interval combined with massively parallel sequencing was used to identify TSPAN12 as the mutated gene in patients with familial exudative vitreoretinopathy [64] . Array based capture of a 2 . 9 Mb interval was also used to identify mutations in taperin ( C9orf75 ) as the cause of nonsyndromic deafness DFNB79 [65] . We used an in-solution method to capture a targeted region based on enzymatic extension of a hybridized oligo to a targeted genomic region using biotin-labeled nucleotides based on a previously established method [22] . This region-specific extraction ( RSE ) method reliably and efficiently extracts long genomic DNA fragments with streptavidin-coated microparticles [22] , [73] . Because RSE does not require shearing the genomic DNA prior to capture , as the previously described methods do , it retrieves large regions of genomic DNA with a very small number of oligos ( each spaced about 8 kb apart ) , making experimental design simple and cost-efficient . Using this method , we were able to achieve 99 . 9% coverage of our 300 kb targeted region and 97% of our 600 kb targeted region , and we demonstrated that this method is effective in the identification of significantly larger deletions than had previously been shown [61] . Based on the excellent read depth and coverage that we obtained , our method is equally effective at identifying point mutations and SNPs , as we did here . We used a combination of targeted genomic DNA capture and massively parallel sequencing to identify a deletion in zebrafish macf1 that affects Balbiani body function and causes a defect in oocyte polarity . Although conditional targeting of mouse macf1 ( acf7 ) in skin epidermis [67] and brain [68] has recently been reported , the function of vertebrate macf1 has primarily been studied in cell culture due to the lethality of the macf1 ( acf7 ) knockout in mice . Until our identification of the mgn mutant , a requirement for vertebrate macf1 had not been examined in the oocyte . We have shown that macf1 is required for cortical localization of the Balbiani body , RNAs , microtubules , and organelles in the zebrafish oocyte and that disruption of macf1 function affects polarity of the oocyte . Further studies will be required to define the molecular function of macf1 during oogenesis and will be particularly valuable for understanding the mechanism by which cell polarity is regulated in the ovary .
The animal work in this study was approved by the Institutional Review Board of the University of Pennsylvania School of Medicine . Phenotype characterization was carried out using the p6cv allele of mgn [19] . Oocytes were staged according to Selman et al . [20] . Ovaries were dissected from euthanized females and fixed overnight at 4°C in 4% paraformaldehyde . Following washes in PBS , ovaries were processed for histology or dehydrated in MeOH prior to in situ hybridization . For histological analysis , ovaries were embedded in JB-4 Plus plastic resin ( Polysciences ) and 5 micron sections were cut using a microtome . Sectioned ovaries were stained for 20 minutes with hematoxylin ( Sigma-Aldrich ) , washed in distilled water , stained for 20 minutes with eosin Y ( Polysciences ) , washed in distilled water , and cleared with 50% EtOH . Stained sections were coated with Permount ( Fisher ) and then coverslipped . Whole mount in situ hybridization was performed as previously described [69] . Following staining , oocytes were embedded in JB-4 Plus Plastic resin and sectioned as described above . The vasa DIG labeled probe was generated using pTY27 [70] and the dazl probe was made using pCRII zdazl ( gift from Kunio Inoue ) . Genomic DNA was pooled from 18 mutant and 24 sibling ( wild-type and heterozygous ) females . Using these pools , mgn was mapped to chromosome 19 using SSLP markers that cover the genome as described [71] . The SSLP markers z31313 and z24515 , which flank the mgn mutation , were used to genotype individual fish . SNP and SSLP markers were generated for further fine mapping . Fine mapping was performed using the SNP marker SNP3 ( 5′-CGT AGG CGT TGC ATA ACT GA-3′ and 5′-GCA AGC AAT CAT ACG CAC AT-3′ ) and either z53070 or CAmarker1 ( 5′-TTG AAG GGT CAC GTT TGA CA-3′ and 5′-CAA GGG TGA AGG GTG AAG AG-3′ ) , which are approximately 7 . 7 kb apart . 976 meioses were analyzed , and 16 recombinants at SNP3 and 11 recombinants at z53070/CAmarker1 were used to narrow the interval to a 2 . 7 cM ( 2 . 02 Mb based on Ensembl Zv7/danRer5 ) interval . For preparation of genomic DNA , fish were euthanized and then flash frozen in liquid nitrogen . 100 mg of tissue from each fish was homogenized and genomic DNA was isolated using the Qiagen Genomic-tip 100/G protocol ( Qiagen ) . Genomic DNA was enriched from two mgn mutant females by region specific extraction ( RSE ) [22] , [72] for two overlapping regions: 1 ) A 319 . 8 kb interval on chromosome 19 ( 32 , 557 , 347–32 , 877 , 103; Zv7/danRer5 ) where the molecular lesion associated with the mgn phenotype was predicted to be based on meiotic mapping; and 2 ) A 613 . 9 kb interval on chromosome 19 ( 32 , 326 , 938–32 , 942 , 969; Zv7/danRer5 ) that extended the 319 . 8 kb region on both sides of macf1 to examine flanking genes and identify SNPs . Genomic DNA from a wild-type ( +/+ ) AB fish was also enriched for the 613 . 9 kb interval to identify SNPs to narrow the interval . RSE was carried out by Generation Biotech at CHOP/UPenn on a Qiagen BioRobot EZ1 . For this capture technology , oligonucleotide primers are hybridized to targeted areas of the genome by exploiting sequence elements that are unique to the region of interest . The bound oligos are extended with biotinylated nucleotides to label the targeted DNA segments . Streptavidin coated magnetic microparticles are then added to the reaction mix to isolate the targeted DNA along with flanking regions [22] , [72] , [73] . The 30 microliter RSE reaction mix consisted of a pre-mixed set of targeting oligos ( 41 primers for the 319 . 8 kb region , 77 primers for the 613 . 9 kb region; oligo design and sequences are below ) combined with 600 ng of genomic DNA . The genomic DNA was denatured and an automated capture performed , followed by washing and elution in pre-loaded reagent cartridges . After RSE , the enriched DNA from each sample was removed from the microparticles by heating the solution at 80°C for 15 minutes to disrupt the biotin-streptavidin complex [22] . The microparticles were magnetically collected and the eluate , which contained the enriched material , was retained . The samples were then tested by quantitative PCR ( QuantiTect ) at three and six loci in the 319 . 8 kb and 613 . 9 kb target regions , respectively , and at the beta-actin 2 locus on chromosome 3 as off-target . These quantitative assays showed enrichment of the target region with low amounts of off-target material in each sample . Specifically , for each 25 microliter reaction , 8 microliters of sample was combined with 1X Qiagen Quantitect Probe PCR master mix ( Cat # 204345 ) , 0 . 4 micromoles each of forward and reverse primers ( Integrated DNA Technologies , IDT , Iowa ) and 0 . 2 micromoles probe ( IDT ) . Six 1:3 serially diluted zebrafish genomic DNA standards were run in duplicate for each locus as well as a single negative control . Forty cycles of 95°C for 15 seconds , 60°C for 1 min were run after the initial denaturation at 95°C for 15 minutes . Fluorescence was collected at 60°C . The primer sequences are: A custom software program was used to identify the oligos for the enrichment . It utilizes and integrates several free , open source or publicly available software solutions to automatically generate appropriate oligo sets at user-defined intervals , spanning genomic regions of interest . In this experiment , 41 capture oligos were used to target the 319 . 8 kb region ( Table 1 ) and 77 oligos for the 613 . 9 kb region ( Table 2 ) . The oligos were designed to target unique zebrafish sequence spanning the region of interest at approximately 8 kb intervals . Through the functionalities embedded in the software program , the UCSC web browser ( http://genome . ucsc . edu ) was first accessed to retrieve repeat- and SNP-masked DNA sequences for the target region in order to exclude these positions as capture points . The program is designed in such a way that the user enters a large genomic region ( currently up to 1 Mb ) , which is then parsed into smaller regions in which the oligos are designed . After generating the oligos , the user is able to redesign small regions in the event that the oligo selection was poor or missing for that smaller region . The preliminary oligo set was then checked via Blat against a local Blat server to ensure that the selected sequences only match to one region . Currently the local Blat server supports human , mouse and zebrafish genomes; more genomes can be added . In a final step , the software program was used to test for cross-reactivity to ensure minimal homo- and hetero-dimerization and hairpin formation . This step ensures that all selected oligos will work well together in a multiplexed format . The GC content for the final oligo set was between 48–52% and the Tm range was 57–61°C . Enrichment values were calculated using the following formula: ( Reads that map to the region examined/total number of reads ) / ( size of interval examined/genome size ) . Enrichment was calculated based on an estimated genome size of 1 . 6 Gb . For the first sequence capture , enrichment of mgn mutant DNA = ( 280884/5861492 ) / ( 315606/1 . 6×109 ) = 243 . For the second sequence capture , enrichment of wild-type DNA = ( 430216/15480724 ) / ( 616031/1 . 6×109 ) = 72 . For the second capture , enrichment of mgn mutant DNA = ( 259817/15639239 ) / ( 616031/1 . 6×109 ) = 43 . Ovaries were dissected from euthanized females and treated with 3 mg/ml collagenase for 10 minutes in Medium 199 ( Invitrogen ) to improve post-fixation morphology ( the majority of follicle cells were still present ) . For acetylated tubulin staining , ovaries were then fixed in 4% paraformaldehyde for 2 hours at room temperature , washed , dehydrated in MeOH and then stained essentially as described [74] . Anti-acetylated tubulin antibody ( Sigma ) was used at a dilution of 1:350 . For rhodamine phalloidin staining , dissected ovaries were fixed overnight at 4°C in Actin Stabilizing Buffer and stained as described [75] . Dissected ovaries were fixed for DiOC6 staining in 4% paraformaldehyde overnight at 4°C . Following washes in PBS , ovaries were stained in 5 micrograms/ml DiOC6 ( Calbiochem ) for 2 hours at room temperature or overnight at 4°C . Confocal microscopy was performed using a Zeiss LSM 510 Laser Scanning Inverted Microscope . Images of acetylated microtubule staining were median filtered using ImageJ to decrease background . Oocyte and Balbiani body sizes were calculated using measurements made with ImageJ software . Because fixed oocytes are not uniformly round , Balbiani body and oocyte diameters were determined by measuring the areas of each from single confocal sections and then calculating the diameters of circles based on the areas . For each oocyte , the optical section with the largest oocyte area was used to measure oocyte size , and the section with the largest Balbiani body area was used to measure Balbiani body size . For calculations of nucleus position , the centroids of the oocyte and the nucleus were determined from single confocal sections using ImageJ . For these studies , Excel was used to calculate statistical significance , which was determined by performing a two-sample unequal variance t-test where alpha = 0 . 05 . Ovaries from mgn mutant or heterozygous siblings were dissected and fixed in 2 . 5% glutaraldehyde+ 2% paraformaldehyde in 0 . 1 M Sodium Cacodylate overnight at 4°C . Fixed samples were rinsed in 0 . 1 M Sodium Cacodylate buffer , post-fixed with 2% osmium tetroxide , dehydrated in graded ethanol , and embedded in Epon . 70 nm thin sections were stained with uranyl acetate and bismuth sub-nitrite [76] , [77] . Stained sections were examined with a JEOL JEM 1010 electron microscope and imaged with a Hamamatsu CCD camera and AMT 12-HR software . Reagents and supplies were purchased from Electron Microscopy Sciences , Fort Washington , PA .
|
How the axes of the embryo are established is an important question in developmental biology . In many organisms , the axes of the embryo are established during oogenesis through the generation of a polarized egg . Very little is known regarding the mechanisms of polarity establishment and maintenance in vertebrate oocytes and eggs . We have identified a zebrafish mutant called magellan , which displays a defect in egg polarity . The gene disrupted in the magellan mutant encodes the cytoskeletal linker protein microtubule actin crosslinking factor 1 ( macf1 ) . In vertebrates , it can take years to identify the molecular nature of a mutation . We used a new technique to identify the magellan mutation , which allowed us to rapidly isolate genomic DNA linked to the mutation and sequence it . Our results describe an important new function for macf1 in polarizing the oocyte and egg and demonstrate the feasibility of this new technique for the efficient identification of mutations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics",
"developmental",
"biology"
] |
2010
|
Microtubule Actin Crosslinking Factor 1 Regulates the Balbiani Body and Animal-Vegetal Polarity of the Zebrafish Oocyte
|
The protozoan pathogen Trypanosoma brucei is transmitted between mammals by tsetse flies . The first compartment colonised by trypanosomes after a blood meal is the fly midgut lumen . Trypanosomes present in the lumen—designated as early procyclic forms—express the stage-specific surface glycoproteins EP and GPEET procyclin . When the trypanosomes establish a mature infection and colonise the ectoperitrophic space , GPEET is down-regulated , and EP becomes the major surface protein of late procyclic forms . A few years ago , it was discovered that procyclic form trypanosomes exhibit social motility ( SoMo ) when inoculated on a semi-solid surface . We demonstrate that SoMo is a feature of early procyclic forms , and that late procyclic forms are invariably SoMo-negative . In addition , we show that , apart from GPEET , other markers are differentially expressed in these two life-cycle stages , both in culture and in tsetse flies , indicating that they have different biological properties and should be considered distinct stages of the life cycle . Differentially expressed genes include two closely related adenylate cyclases , both hexokinases and calflagins . These findings link the phenomenon of SoMo in vitro to the parasite forms found during the first 4–7 days of a midgut infection . We postulate that ordered group movement on plates reflects the migration of parasites from the midgut lumen into the ectoperitrophic space within the tsetse fly . Moreover , the process can be uncoupled from colonisation of the salivary glands . Although they are the major surface proteins of procyclic forms , EP and GPEET are not essential for SoMo , nor , as shown previously , are they required for near normal colonisation of the fly midgut .
Various sub-species of the protozoan parasite Trypanosoma brucei cause sleeping sickness in humans and Nagana in domestic animals . Irrespective of their mammalian host range , all these parasites are dependent on tsetse flies for their transmission . Two features enable trypanosomes to establish chronic infections in the mammalian host - their ability to evade the immune response by periodic switching of their variant surface glycoprotein ( VSG ) coat ( reviewed in [1] ) and a quorum-sensing mechanism that drives the differentiation of proliferating slender bloodstream forms to non-dividing stumpy forms , thus limiting the parasitaemia [2] , [3] . Stumpy-inducing factor ( SIF ) is a small molecule ( <500 Da ) produced by the slender forms; its chemical identity is not known . Stumpy forms are pre-adapted for further differentiation and , following ingestion by the tsetse fly , differentiate into early procyclic forms in the lumen of the insect midgut [4] . In addition to changes in morphology and metabolism , differentiation involves the replacement of the VSG coat by two insect-specific coat proteins , GPEET and EP procyclin . At the beginning of tsetse infection procyclic forms can have two fates: they can be eliminated by the fly or they can migrate across/around the peritrophic matrix and colonise the ectoperitrophic space [5] . Once the infection is established , it is characterised by late procyclic forms that express high levels of EP , but are negative for GPEET . GPEET is not required for migration to the ectoperitrophic space , since deletion mutants can establish infections at normal rates [6] . Early and late procyclic forms can be maintained in axenic culture . Addition of glycerol to the culture medium prolongs the expression of GPEET; once glycerol is removed , the cells undergo a transient growth arrest and GPEET is repressed within a few days [4] , [7] . Different trypanosome stocks vary in the relative amounts of GPEET or EP that they express in culture [8] . In contrast to what is observed in tsetse , GPEET-negative cells can revert to being GPEET-positive in culture , for example in response to glucose depletion [9] or by an unknown mechanism that is independent of glycerol [10] . To complete the cycle in the fly , parasites must migrate from the midgut , via the proventriculus , to the salivary glands . This migration constitutes a major bottleneck in the life cycle [11] . Once they reach the salivary glands trypanosomes attach to the epithelia and proliferate as epimastigote forms , finally giving rise to infectious metacyclic forms that can infect a new mammalian host . Unicellular organisms can function as multicellular communities that exchange signals with each other and move in a coordinated manner . This is particularly well described for bacteria , which can form biofilms , communicate by quorum sensing and exhibit adventurous or social motility ( SoMo ) [12]–[15] . These types of concerted behaviour have implications for virulence and present potential targets for new classes of antimicrobial drugs . In contrast to what is known about social behaviour in prokaryotes , there is considerably less information on social interactions between unicellular eukaryotes . While several species of fungi are capable of forming biofilms [16] , studies of swarming motility have focused almost exclusively on the free-living social amoeba Dicytostelium discoideum [17] , [18] . In general , unicellular parasites tend to be studied as isolated entities or as organisms that need to perceive and interact with their hosts , with relatively little attention being paid to how they communicate with each other [19] . Procyclic forms of T . b . brucei exhibit SoMo when plated on a semi-solid surface , in a manner reminiscent of swarming bacteria [20] . Parasites first grow at the site of inoculation , and then form radial protrusions or “fingers” that extend outwards . Independent communities are able to sense each other and reorganise group movement to prevent contact . Migration on plates is abolished if the trypanosomes have a dysfunctional flagellum [20] or other motility defects [21] . It has been hypothesised that the social motility observed on plates might reflect one of the migration steps within the fly vector , either from the midgut lumen to the ectoperitrophic space , or from the ectoperitophic space to the salivary glands [20] . By using a series of mutants that had previously been characterised in tsetse , we show that SoMo is unrelated to the parasites' ability to establish salivary gland infections . Instead , it is a property of the early procyclic form , which is found in the first few days after transmission of bloodstream forms to the tsetse fly . We have also identified several new markers in addition to GPEET that are differentially expressed in early and late procyclic culture forms , and verified their differential expression in tsetse flies . Taken together , this confirms that early and late procyclic forms are distinct life-cycle stages with specific expression profiles and characteristics and links SoMo to an early event in the colonisation of the tsetse midgut .
As a first step we optimised the plating protocol for the fly-transmissible strain AnTat 1 . 1 . The main differences from the previously published protocol [20] are that we used SDM79 rather than SM as the medium and cells were not preincubated with ethanol before plating . In addition , low melting temperature agarose was replaced by normal agarose , rendering the plates more robust . While establishing the SoMo assay we observed that the time-point when radial protrusions formed differed between experiments . To test if the cell density influenced the assay , different numbers of cells were pipetted onto the plates ( Figure 1 ) . When 8×105 cells were plated in a volume of 5 µl , fingers were already visible after 24 hours . Cells plated at a density of 4×105 or 2×105 cells in 5 µl showed SoMo after 48 or 72 hours , respectively . It was reported previously that the doubling time of trypanosomes on plates is 24 h [20] . This suggests that the cells reach a threshold number of approximately 1 . 6×106 before migration starts . We observed that when communities were plated on their own , the radial projections always grew in a clockwise direction ( Figure 1 , 72 h ) . This directionality was overridden , however , when cells sensed and avoided neighbouring communities ( Figure 2 ) . When we tested a variety of mutants , the high frequency of clones that were SoMo-negative , coupled with the observation that some addback mutants gave inconsistent results , made us suspect that a factor unrelated to the genotype might be influencing the outcome . We have shown previously that culture conditions can influence GPEET expression [22] . When we monitored the expression of GPEET , we found that 3 cultures that were SoMo-positive were all GPEET-positive and conversely , 4 cultures that were SoMo-negative were all negative for GPEET . We then systematically examined SoMo of early and late procyclic forms . For these experiments we derived early procyclic forms from bloodstream forms and let them differentiate into late procyclic forms by removal of glycerol . When these cultures were seeded onto plates containing glycerol , both early and late procyclic forms grew and formed colonies at the inoculation site , but only the former produced migrating fingers ( Figure 2 ) . It has previously been shown that glycerol alone does not trigger the reversion of late to early procyclic forms in liquid culture [4] . Nevertheless , to be sure that the status of the cells had not changed on the plates , a “community lift” was performed . This entails placing a nitrocellulose filter on the plate; when the filter is removed , the cells adhere to it and can be labelled with antibodies . Incubation of the filter with antibodies against GPEET and EP confirmed that the early procyclic forms were positive for both , as expected , and that most cells in the colony of late procyclic forms were negative for GPEET . Some GPEET-positive cells can always be detected in cultures without glycerol [7]; these are visible as a narrow ring at the edge of the colony in Figure 2 . It is not clear if the few early procyclic forms actively migrate to the border of the colony or if cells at the edge are more likely to revert to expressing GPEET . Although late procyclic forms do not show SoMo , they are recognised by early procyclic forms , which react by changing their direction of migration ( Figure 2 ) . GPEET is the major surface protein of early procyclic forms . To test if it was required for SoMo we used the ΔGPEET deletion mutant previously generated in our laboratory [6] . Since these cells lack a marker for early procyclic forms , we once again took bloodstream forms and triggered them to differentiate to procyclic forms . In common with its wild-type parent , ΔGPEET was SoMo-positive as long as it was cultured in the presence of glycerol and became SoMo-negative after being transferred to glycerol-free medium ( Figure 3A ) . In order to track the differentiation status of ΔGPEET , it was transformed with a reporter construct in which the GFP coding region is fused to the GPEET 3′ untranslated region [23] . This regulatory sequence ensures that expression of GFP mimics that of GPEET , and indicates whether or not a cell is still an early procyclic form . A community lift using an anti-GFP antibody revealed once again that only the early procyclic forms migrate while the late , GFP-negative cells stay at the point of inoculation ( Figure 3B ) . In addition to migrating , ΔGPEET cells are still capable of recognising and reacting to other trypanosome communities ( Figure 3A ) . In culture , early and late procyclic forms are morphologically indistinguishable . Since GPEET was the only known marker for early procyclic forms at the beginning of this study , we used SILAC to identify additional proteins that were differentially expressed between GPEET-positive and GPEET-negative cells . Two independent experiments identified a limited number of candidates that were significantly different in at least one experiment ( ≥2-fold; Figures 4 and 5; Table S2 ) . Of the differentially regulated proteins , three examples were encoded by related genes . These were the calflagins ( Tb-44 , Tb-24 and Tb-17 ) , the two hexokinases ( HK1 and HK2 ) and three adenylate cyclases . The members of a protein family could not be identified unequivocally as they contained identical peptides that are randomly assigned during mapping . Lacking antibodies that discriminated between isoforms , we analysed the transcripts for unique signatures . In the case of the adenylate cyclases ( Tb927 . 5 . 285b , Tb927 . 5 . 320 and Tb 927 . 5 . 330 - here designated AC330 , AC320 and AC285b ) differences in their 3′ untranslated regions , allowed AC330 to be distinguished from AC320/285b by Northern blot analysis ( Figure 6A ) . Both were differentially expressed , with AC330 up 9-fold in early procyclic forms and AC320/285b up 6 . 25-fold in late procyclic forms . Thus , the changes in protein levels detected by SILAC are probably an under-estimate for the individual proteins . Since an antiserum was available against calflagins , we monitored expression of these proteins by immunofluorescence . This revealed that calflagin-positive cells were always also positive for GPEET ( Figure 6B ) . In addition , we performed quantitative RT-PCR to measure transcript levels in early and late procyclic forms ( Figure 6C ) . This confirmed the differential expression at the level of mRNA for the adenylate cyclases and HK1/HK2 . HK1 mRNA was expressed 4-fold more in early procyclic forms while HK2 was up-regulated 2-fold in late procyclic forms . In contrast to what was observed by immunofluorescence and SILAC , calflagin transcripts were down-regulated only 2-fold in late procyclic forms , suggesting that there is an additional level of regulation . Finally , we tested the mRNA levels of a set of putative pteridine transporters ( PPT: Tb927 . 1 . 2850 , Tb927 . 1 . 2880 ) , which we have observed to be up-regulated ( at least transiently ) during differentiation of early to late procyclic forms; these were increased 4 . 6-fold in late procyclic forms . In summary , although no other gene is as tightly regulated as GPEET , we have identified several additional differentially regulated transcripts/proteins in early and late procyclic forms . It has been shown previously that expression of the GPEET transcript and protein in the fly mirrors that of cells differentiating from early to late procyclic forms in culture [4] , [22] , [24] . To test if the new markers that we identified were similarly regulated in vivo , tsetse flies were infected and trypanosomes were harvested 3 and 12 days post infection . Figure 7A shows the co-expression of GPEET and calflagin in early procyclic forms isolated from fly midguts at day 3 and the repression of both proteins by day 12 . Quantitative RT-PCR ( Figure 7B ) showed the same profiles that were observed in culture , with GPEET , AC330 and HK1 being more highly expressed in early procyclic forms and AC320 , HK2 and PPT being more highly expressed in late procyclic forms . Taken together , these data convincingly show that early procyclic forms in culture are equivalent to the procyclic forms early in infection and late procyclic forms correspond to those in established infections . Despite the lack of SoMo by late procyclic forms , it is possible that it plays a role in migration of proventricular forms across the cardia to the tsetse salivary glands . To test this hypothesis we used a series of deletion mutants with defects in salivary gland infection . Our previous studies have implicated at least two proteins in the establishment of mature salivary gland infections , mitogen-activated kinase kinase 1 ( MKK1; [25] ) and the surface protein PSSA-2 [26] . Parasites lacking MKK1 were completely unable to establish salivary gland infections and parasites lacking PSSA-2 showed reductions in the prevalence and intensity of infections . A procyclin null mutant , lacking all EP and GPEET genes ( Δproc ) , also showed a defect in colonisation of the salivary glands [6] . ΔMKK1 and ΔPSSA-2 infect the midgut at normal rates and intensities [25] , [26] , while Δproc establishes heavy infections at about half the rate of its wild-type parent [6] . MKK1 AND PSSA-2 knockouts were plated as early and late procyclic forms; in the case of Δproc only early procyclic forms , derived directly from bloodstream forms , were tested ( Figure 8 ) . In all cases , the early procyclic forms were positive for SoMo and were also able to sense and avoid the communities of late procyclic forms on the same plate .
Early procyclic forms - defined as GPEET-positive cells - are detected in the midgut of tsetse flies in the first week after uptake of bloodstream form trypanosomes [4] , while establishment of a persistent infection correlates with differentiation to late ( GPEET-negative ) procyclic forms . We have discovered that SoMo is a property of early procyclic forms and that late procyclic forms are consistently SoMo-negative . There are several indications that SoMo reflects the migration of trypanosomes from the midgut lumen to the ectoperitrophic space in the first days of fly infection rather than the subsequent migration from the ectoperitrophic space to the salivary glands . First and foremost , SoMo is restricted to early procyclic forms whereas late procyclic forms , which are forerunners of the forms migrating to the salivary glands , are SoMo-negative . Second , the timing of the switch from early to late procyclic forms [4] correlates with the appearance of parasites in the ectoperitrophic space [5] . Third , SoMo is independent of GPEET , as is colonisation of the ectoperitrophic space [4] . Furthermore , three mutants ( Δproc , ΔPSSA-2 and ΔMKK1 ) that show normal colonisation of the midgut , but defects in colonisation of the salivary glands [6] , [25] , [26] are SoMo-positive as early procyclic forms . While it might be argued that these mutants have other defects , such as an inability to penetrate the proventriculus or to differentiate to epimastigote forms , in no case does the mutation impair SoMo by early procyclic forms . Social interactions between bacteria are known to involve outer membrane proteins [27] . Despite being the major surface glycoproteins of procyclic forms , and present in several million copies , neither GPEET nor EP procyclin is required by trypanosomes for SoMo . It is known , however , that procyclin null mutants export free GPI anchors to their surface [28] , and these might compensate for the loss of procyclins . The insect-stage specific transmembrane protein PSSA-2 [26] , which shows increased expression in late procyclic forms ( Table 2 ) , is also dispensable for SoMo . In this study we identified additional proteins and transcripts that are differentially expressed in these two life-cycle stages , both in culture and in the fly . Like GPEET , calflagins are expressed by early procyclic forms , but are down-regulated in late procyclic forms . It was recently shown by Emmer and coworkers that calflagins are expressed by bloodstream and procyclic forms [29] . However , when Kolev et al . induced differentiation from procyclic to epimastigote and metacyclic forms by overexpression of RBP6 [30] , the procyclic forms in their cultures were heterogeneous with respect to calflagin expression , suggesting that they were a mixed population of early and late forms . Calflagins were not detected in epimastigotes but were re-expressed by metacyclic forms in culture [30] and in the fly salivary glands [31] . Guided by SILAC , we also identified two pairs of proteins , HK1/HK2 and AC330/AC320 , whose transcripts are reciprocally expressed in early and late procyclic forms . This is similar to the situation that is seen with GPEET and EP3 [4] , [24] , [32] . Given that we only detected about 1300 proteins by mass spectrometry in the two experiments ( Table S2 ) , we do not claim that the list of differentially regulated proteins is complete , and indeed we suspect that there might be other sets of proteins that are reciprocally regulated without a discernible net change . Moreover , there might also be post-translational modifications or non-peptide moieties that are stage-specific . For example , the activity of the kinase that phosphorylates GPEET is restricted to early procyclic forms [33] . It is also known that early procyclic forms of T . congolense preferentially express PRS , a protease-resistant surface glycoconjugate [34] , [35] , although an equivalent molecule has not been reported for T . brucei . Our experiments show that the parasites need to reach a threshold concentration on plates before they start to migrate . Although the number of parasites in the midgut lumen is significantly lower during the early days of infection [5] , [36] , the three-dimensional structure of the midgut and host-derived factors might contribute to the response . Moreover , local accumulation of parasites , for example at the peritrophic matrix , could condition the micro-environment in a manner conducive to SoMo . In addition to migrating , early procyclic forms have the capacity to recognise and be repelled by communities of early or late procyclic forms . At present we can only speculate about the significance of such repellents , but one possibility is that they are used by late procyclic forms to prevent or reduce superinfection by a second strain of trypanosomes . Mixed infections of tsetse can be detected in the field [37] , [38] , but there is no information on whether flies acquire the parasites simultaneously or sequentially from infected mammals . Although the differentiation of early to late procyclic forms in the tsetse fly is irreversible , it should always be borne in mind that trypanosomes can switch between these two life-cycle stages in liquid culture [22] . The fact that these can change over time has important implications for the interpretation of results . In particular , before it can be concluded that a specific gene is required for SoMo , it is essential to determine whether the parasites are early or late procyclic forms . In conclusion , our findings add further credence to the designation of early and late procyclic forms as two distinct life cycle stages with different biological properties . Since early procyclic forms are only detected in the first week of tsetse fly infection [4] , [32] , this strongly suggests that SoMo reflects an early event in the colonisation of the insect host . It also implies that genes that are important for SoMo will also play a role in the establishment of a midgut infection . Using the SoMo assay as a surrogate for fly experiments would enable many more laboratories to examine this aspect of parasite transmission . In addition , related parasites such as the South American trypanosomes and Leishmania , which are also transmitted by insects , may be amenable to such studies .
No vertebrate animals were used in this study . Bloodstream form trypanosomes were obtained from frozen stabilates stored in liquid nitrogen . Antibodies were obtained from cell culture supernatants or from pre-existing sources of serum . The pleomorphic strain AnTat 1 . 1 [39] , [40] and genetically manipulated derivatives of it were used in this study . The deletion mutants ΔPSSA-2 [26] , ΔMKK1 [25] , ΔGPEET and Δproc [6] have all been described previously . Procyclic forms were cultured in SDM79 [41] supplemented with 10% heat inactivated foetal bovine serum ( iFBS ) . The medium for early procyclic forms was also supplemented with 20 mM glycerol [4] . Pupae of Glossina morsitans morsitans were obtained from the Department of Entomology , Slovak Academy of Science ( Bratislava ) . Teneral flies were infected with early procyclic forms during the first blood meal as described [42] . Flies were dissected and total RNA was isolated from midguts using standard procedures [43] . Fifty midguts were collected 3 days post infection for RNA isolation and immunofluorescence analysis . Approximately 10–15 infected midguts were collected for analysis at days 12 or 13 post infection . Bloodstream forms obtained from frozen stabilates ( 500 µl blood plus 500 µl HMI-9 ) were centrifuged and resuspended in SDM79 supplemented with 10% iFBS and 20 mM glycerol . Differentiation to early procyclic forms was induced by the addition of 6 mM cis-aconitate and a temperature shift from 37° to 27° [44] . Early procyclic forms were cultured in the same medium [4] . To trigger differentiation to late procyclic forms , early procyclic forms were transferred to SDM79 , 10% iFBS without glycerol , as described previously [4] . Stable transformation of procyclic form trypanosomes was performed as described [26] . To generate the ΔGPEET/GFP-GPEET cell line , which expresses GFP under the control of the GPEET promoter and 3′ UTR , ΔGPEET [6] was stably transformed with the plasmid pCorleone-GFP/GPEET-blast [23] . When linearised with Spe I , this plasmid integrates upstream of a procyclin locus . The protocol to produce plates was adapted from [20] . Plates were always used within 24 h . Briefly , 36 ml SDM79 supplemented with 10% iFBS were pre-warmed to 42°; for plates containing glycerol , 400 µl of a 2M glycerol stock was added . 4 ml agarose ( Promega V3125; 4% w/v in water ) was added to the pre-warmed medium and the resulting 0 . 4% agarose medium was immediately poured into Petri dishes with a diameter of 85 mm , 10 ml per dish . The open plates were then air-dried for 1 hour in a laminar flow cabinet . Cells from an exponentially growing culture were centrifuged briefly and resuspended in the residual medium at a density of 3–4×107 cells ml−1 . Five µl were spotted onto the surface of the agarose , the Petri dish was sealed with Parafilm and incubated at 27° . All experiments were performed at least twice and there were no incongruent results . It should be noted , however , that the number of spokes produced by a given clone can vary between experiments . A nitrocellulose filter ( Whatman Protran BA 85 ) was laid carefully on top of the cells on the agarose plate and incubated for 5 minutes at room temperature . The filter was then removed and air-dried for 15 minutes . The membrane was blocked in PBS containing 5% ( w/v ) defatted milk for 1 h at 4° , after which the primary antibodies were added at the appropriate dilution and incubated for 1 h at room temperature . The following primary antibodies were used: TBRP1/247 mouse α-EP 1∶500 [45] , K1 rabbit α-GPEET 1∶1000 [42] and mouse α-GFP ( Roche , 1∶2000 ) . After incubation with the primary antibodies the membrane was washed 3 times in TBS Tween , then incubated with secondary antibodies ( in PBS 5% milk ) for 1 h at room temperature . The following secondary antibodies were used at a dilution of 1∶10000: goat α-mouse IRDye 800CW ( LI-COR Biosciences ) and goat α-rabbit IRDye 680LT ( LI-COR Biosciences ) . The membrane was washed 3 times in TBS Tween and then scanned on a LI-COR Odyssey Infrared Imager model 9120 , using Odyssey Application Software , Version 3 . 0 . 21 . Images from plates were made with a Nikon MH-56 digital camera . To quantify the intensity of the community lifts a grey scale image of the membranes was exported from the Odyssey Application Software and analysed with ImageJ 1 . 46r . Seven individual areas were analysed for each value . The values were subtracted from 255 to obtain a maximum intensity of 255 and a minimum intensity of 1 . The graphs were generated with Prism6 . Late procyclic forms were derived from early procyclic forms by removal of glycerol in two independent experiments . Pairs of early and late procyclic forms were adapted to SDM80 supplemented with 10% dialysed foetal bovine serum . The medium for early procyclic forms was supplemented with 20 mM glycerol . SILAC and mass spectrometry analyses were performed as described [46] at the Mass Spectrometry and Proteomics Facility , Department of Clinical Research , University of Bern . The isolation of RNA from early and late procyclic culture forms and northern blot analysis were performed as described [6] . Purified RNA was subjected to DNAse treatment prior to cDNA synthesis . Reverse transcription was performed using an Omniscript RT kit ( Quiagen , Switzerland ) according to the manufacturer's instructions with random hexamers as primers . PCR primers are shown in Table S1 . qPCR was performed using MESA GREEN qPCR MasterMix Plus for SYBR Assay ( Eurogentec ) in the ABI Prism 7000 Sequence Detection System ( Applied Biosystems ) . Specificity of the reactions was confirmed by agarose gel electrophoresis and melting temperature analysis . The data were analysed using 7000 System SDS software v1 . 2 ( Applied Biosytems ) . Two biological replicates were analysed independently . Within an experiment , technical triplicates were run in parallel . Cells were washed twice with PBS and spread on a coverslip to let them settle down for 10 minutes . The cells were fixed with 4% paraformaldehyde and 0 . 1% glutaraldehyde in PBS for 15 minutes , then permeabilised with 0 . 2% Triton X-100 and blocked with 2% BSA/PBS . The primary antiserum , rabbit K1 anti-GPEET was diluted 1∶1000 [42] and the calflagin mouse antiserum ( a gift from David Engman ) , was diluted 1∶500 [29] . The secondary antibodies Alexa Fluor 488 goat anti-rabbit and Cy3 goat anti-mouse ( Invitrogen ) were diluted 1∶1000 in 2% BSA/PBS . Images were taken with a Leica DFC360FX monochrome CCD ( charge-coupled-device ) camera mounted on a Leica DM5500 B microscope with a 100× oil immersion objective and analysed using LAS AF software ( Leica ) .
|
African trypanosomes , single-celled parasites that cause human sleeping sickness and Nagana in animals , are transmitted by tsetse flies . Bloodstream form trypanosomes ingested by tsetse differentiate into procyclic forms in the midgut lumen of the insect . Successful transmission to a new mammalian host requires at least two migrations within the fly: one from the midgut lumen to the ectoperitrophic space , and a subsequent migration from the ectoperitrophic space to the salivary glands . Procyclic forms can exhibit social motility , a form of coordinated movement , on semi-solid surfaces . While social motility in bacteria is linked to virulence , the biological significance for trypanosomes is unknown . We demonstrate that social motility is a property of early procyclic forms , which are equivalent to the forms present during the first week of fly infection . In contrast , late procyclic forms characteristic for established infections are deficient for social motility . Our findings link social motility to a biological process , confirm that early and late procyclic forms are distinct life-cycle stages and imply that genes essential for social motility will be of key importance in fly transmission . We suggest that using the social motility assay as a surrogate for fly experiments should enable many more laboratories to examine this aspect of parasite transmission .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"motility",
"developmental",
"biology",
"cell",
"biology",
"parasitic",
"cell",
"cycles",
"cell",
"migration",
"life",
"cycles",
"parasitology",
"biology",
"and",
"life",
"sciences",
"cell",
"processes",
"microbiology",
"molecular",
"cell",
"biology",
"cell",
"aggregation",
"molecular",
"biology",
"cell",
"differentiation",
"parasitic",
"life",
"cycles"
] |
2014
|
Social Motility of African Trypanosomes Is a Property of a Distinct Life-Cycle Stage That Occurs Early in Tsetse Fly Transmission
|
Aedes albopictus is an invasive species which continues expanding its geographic range and involvement in mosquito-borne diseases such as chikungunya and dengue . Host selection patterns by invasive mosquitoes are critically important because they increase endemic disease transmission and drive outbreaks of exotic pathogens . Traditionally , Ae . albopictus has been characterized as an opportunistic feeder , primarily feeding on mammalian hosts but occasionally acquiring blood from avian sources as well . However , limited information is available on their feeding patterns in temperate regions of their expanded range . Because of the increasing expansion and abundance of Ae . albopictus and the escalating diagnoses of exotic pathogens in travelers returning from endemic areas , we investigated the host feeding patterns of this species in newly invaded areas to further shed light on its role in disease ecology and assess the public health threat of an exotic arbovirus outbreak . We identified the vertebrate source of 165 blood meals in Ae . albopictus collected between 2008 and 2011 from urban and suburban areas in northeastern USA . We used a network of Biogents Sentinel traps , which enhance Ae . albopictus capture counts , to conduct our collections of blooded mosquitoes . We also analyzed blooded Culex mosquitoes collected alongside Ae . albopictus in order to examine the composition of the community of blood sources . We found no evidence of bias since as expected Culex blood meals were predominantly from birds ( n = 149 , 93 . 7% ) with only a small proportion feeding on mammals ( n = 10 , 6 . 3% ) . In contrast , Aedes albopictus fed exclusively on mammalian hosts with over 90% of their blood meals derived from humans ( n = 96 , 58 . 2% ) and domesticated pets ( n = 38 , 23 . 0% cats; and n = 24 , 14 . 6% dogs ) . Aedes albopictus fed from humans significantly more often in suburban than in urban areas ( χ2 , p = 0 . 004 ) and cat-derived blood meals were greater in urban habitats ( χ2 , p = 0 . 022 ) . Avian-derived blood meals were not detected in any of the Ae . albopictus tested . The high mammalian affinity of Ae . albopictus suggests that this species will be an efficient vector of mammal- and human-driven zoonoses such as La Crosse , dengue , and chikungunya viruses . The lack of blood meals obtained from birds by Ae . albopictus suggest that this species may have limited exposure to endemic avian zoonoses such as St . Louis encephalitis and West Nile virus , which already circulate in the USA . However , growing populations of Ae . albopictus in major metropolitan urban and suburban centers , make a large autochthonous outbreak of an arbovirus such as chikungunya or dengue viruses a clear and present danger . Given the difficulties of Ae . albopictus suppression , we recommend that public health practitioners and policy makers install proactive measures for the imminent mitigation of an exotic pathogen outbreak .
Understanding the blood feeding patterns of mosquitoes is of paramount importance in determining their vector status in the maintenance and epidemic transmission of arboviruses . Blood feeding patterns of mosquito vectors provide insight into the ecological transmission cycles of pathogens and lead to more efficient disease and vector control measures for the benefit of animal and human health . For invasive mosquitoes with expanding geographic ranges , such as Aedes albopictus ( Skuse ) , the specific blood-hosts impact endemic diseases and can lead to the epidemic transmission of exotic pathogens . The Asian tiger mosquito , Ae . albopictus , has dispersed extensively from its native tropical range in Southeast Asia and is now found on every continent except Antarctica [1] , [2] . The last decade has seen a dramatic expansion of Ae . albopictus into temperate regions of Europe and North America [3]–[5] . In many parts of its expanded range , this species is implicated as a significant vector of emerging and re-emerging arboviruses such as dengue ( DENV ) and chikungunya ( CHIKV ) . Although historically not an important vector of CHIKV , Ae . albopictus has become the principal driver of recent epidemics in Asia and islands in the Indian Ocean because of a mutation in the virus envelope protein enhanced transmission efficiency by this species [6] , [7] . Autochthonous transmission of CHIKV has also been recorded in temperate regions of Italy and France [8] , [9] where invasive Ae . albopictus have become abundant [3] . Aedes albopictus was also the sole vector in local epidemics of dengue in Hawai'i and other regions [10] , [11] and is a competent laboratory vector for at least 22 arboviruses [12] . Due to the widespread and increasing distribution of Ae . albopictus in temperate regions and the escalating diagnoses of exotic pathogens in travelers returning from endemic or epidemic areas [13] , [14] , the risk of an outbreak in a new area is no longer hypothetical . Furthermore , because this species thrives in artificial containers found in close association with human peridomestic environments , it is essential to fully investigate the host feeding patterns of Ae . albopictus in order to completely understand its role in disease ecology and public health significance . Surprisingly , given the vector potential and medical importance of Ae . albopictus , few studies have been conducted to investigate the host feeding patterns of this species in its native and expanding geographic range . This is likely because adult Ae . albopictus are a difficult species to collect efficiently in traps , and blood fed specimens are especially rare . From the few studies that have been conducted , the precise host feeding preferences of Ae . albopictus seem to vary considerably ( Table 1 ) . The species has been generally reported to feed on a wide range of mammals including humans , but will also feed on avian hosts at various proportions , and has even been incriminated to feed on amphibians and reptiles [15]–[34] . It has thus been considered an opportunistic feeder and a classic bridge vector candidate between zoonotic arboviruses and humans . However , caution should be taken in labeling Ae . albopictus as an efficient bridge vector because the large variation in the feeding plasticity of this species questions the exact role that it may play as an enzootic or epidemic vector of arboviruses . For example , in its native tropical range , Ae . albopictus feeds exclusively on humans in Indonesia [35] , whereas in Singapore it feeds on humans , oxen , and dogs [15] . Additionally , studies conducted in Thailand [36] have reported that Ae . albopictus feed on humans , swine , buffalo , dogs , and chickens , while more recent investigations [26] report that Ae . albopictus feeds only on humans , with a few ( <6% ) double-host blood meals between humans and swine/cat/dog . In temperate Japan , Ae . albopictus primarily feed on mammals , with a high propensity for humans , but also on birds and amphibians/reptiles [29] , [30] ( Table 1 ) . In temperate locations of the expanding range of Ae . albopictus , the host preference of this species is also variable . Studies conducted at a tire dump in Missouri , USA , reported that Ae . albopictus will feed on birds ( 17% ) but prefer mammals ( 64% ) , with 8 . 2% of those mammalian feedings obtained from humans [19] . A follow up study conducted in other tire yards and surrounding vegetation of rural and urban habitats in Missouri , Florida , Indiana , Illinois , and Louisiana , USA , concluded that Ae . albopictus showed a strong preference for mammals ( >94% ) , with up to 8% human-derived blood meals , while also detecting avian ( 1% ) and reptilian ( 5% ) blood meals [20] . An additional study in suburban landscapes of North Carolina , USA , reported that Ae . albopictus feeds predominately on mammalian hosts ( 83% ) , but also on birds ( 7% ) , amphibians ( 2% ) , and reptiles ( 2% ) [27] . In Europe , Italian populations of Ae . albopictus rarely feed on birds in urban settings , while 99% of specimens have been reported to feed on mammals , with 90% of those mammalian blood meals being derived from humans [31] . The same investigators report that in suburban settings of Italy , 7% of Ae . albopictus had fed on avian species , while the vast majority of the blood meals were mammalian-derived ( 95% ) , with 43% containing human blood [31] . Finally , in urban zones of Spain , Ae . albopictus obtained blood meals exclusively from humans ( 100% ) [32] ( Table 1 ) . Although it is apparent that Ae . albopictus feeds predominantly on mammals , the degree of mammalophagic or anthropophagic host feeding preferences of this species appear location specific . Because of the rapidly expanding range of Ae . albopictus , its abundance in metropolitan centers , and its close association with humans in peridomestic habits , combined with the emergence and resurgence of exotic pathogens for which Ae . albopictus is a capable vector , it is clear that assessing its host feeding preferences in newly invaded areas is critical to elucidate disease transmission cycles and develop strategies to reduce the local risk of an exotic arbovirus outbreak . However , the collection of Aedes ( Stegomyia ) spp . , such as Ae . albopictus , has been difficult because standard vector surveillance traps are generally placed 1 . 5 m above the ground , are operated overnight , and utilize light as an attractant [37] . Since Ae . albopictus is diurnal and not attracted to light , host-seeks near the ground surface , and utilizes visual , in addition to olfactory cues for host location [18] , [21] , [38] these traps are not an effective way to collect this species . Consequently , most blood meal analyses to date were performed on specimens collected from areas where their densities are very high , such as tire yards and tire dumps ( Table 1 ) . The creation of newly developed vector surveillance traps , such as the Biogents Sentinel ( BGS ) trap , have only recently allowed the collection of large number of Ae . albopictus specimens from typical urban and suburban areas for ecological studies [39] . These traps simulate convection currents created by human body heat , utilize lures which mimic human odors , are operated during the day , placed at the ground level , and utilize contrasting black and white markings that provide additional visual cues that may be attractive to Ae . albopictus [37]–[41] . We investigated the host feeding patterns of Ae . albopictus in temperate North America , near the northernmost boundary of established populations in the eastern United States [4] , [5] . We used an extensive network of BGS traps , which enhance Ae . albopictus capture counts , to conduct a multi-year collection of blooded mosquitoes ( 2008–2011 ) in urban and suburban sites as part of a larger area-wide project aimed at managing the Asian tiger mosquito [42] , [43] . Additionally , we assayed blood meals from Culex mosquitoes collected in the same traps , locations , and dates as Ae . albopictus to determine the diversity of different blood meal sources obtained from the two vectors . We discuss the implications of our results on established and expanding populations of Ae . albopictus and the imminent outbreaks of exotic diseases such as chikungunya or dengue fevers in North America .
All studies were conducted within the jurisdictions of the authors' respective governance domain by professional mosquito control personnel . All entomological surveys and collections made on private lands or in private residences were conducted after acquisition of oral or written consent from residents . No specific permits were required for the mosquito collections . These studies did not involve endangered or protected species . All collections were conducted within two counties ( Mercer and Monmouth ) located in central New Jersey , USA . Mercer County ( 40°13′N , 74°44′W ) is highly urban , with 364 , 883 residents [44] and a population density of 630 . 2 inhabitants per square kilometer . Mercer County and the low-income City of Trenton , where the studies were conducted , have a population density of 4 , 286 . 5/km2 ( USCB 2009a ) . The City of Trenton contains typical dense inner city housing , often built as adjoining row homes or duplexes [45] . Monmouth County ( 40°44′N , 74°17′W ) is defined as primarily suburban and is located in east-central New Jersey with a population of 630 , 380 [46] . The boroughs on the Raritan Bayshore , within Monmouth County , where the studies were conducted , have an average population of 1 , 907 . 4/km2 [46] . The Raritan Bayshore primarily contains middle income coastal suburban homes which are often interspersed with forest and green space remnants [42] . Within each county , three predefined ∼1 , 000-parcel sites ( a parcel is a combination of a house and its associated yard space ) , ranging in area from 1 km2 ( Mercer ) to 2 km2 ( Monmouth ) were chosen for our investigations . Although individual parcel sizes within the study sites in Mercer County were smaller ( 199 . 5±18 . 3 m2 ) than those in Monmouth County ( 571 . 1±31 . 2 m2 ) , the number of residents within Mercer sites ( 19 , 494 ) were larger than within Monmouth sites ( 12 , 743 ) . Every site , within each county , was previously selected to contain similar socioeconomic parameters , geography , human population density , and mosquito abundance . For a detailed description about site selection and the parameters of each individual site , please refer to [42] , [43] . Mosquitoes were sampled on a weekly basis during 2008–2011 using a network of Biogents Sentinel ( BGS ) traps ( Biogents AG , Regensburg , Germany ) . Specific details of surveillance protocols are outlined elsewhere [40]–[43] , [47]; but briefly , trap locations were chosen by overlaying a grid of specific distance intervals . We used a 175–200 m distance between BGS traps for each site in Mercer County and 200–400 m distances in Monmouth County because of the larger site areas and limiting number of traps in inventory . These distances were based on current knowledge of Ae . albopictus flight range [21] and the available resources within each county . A total of 36 to 51 BGS traps , depending on the year , were deployed weekly in Mercer County , while 55 to 57 traps were deployed in Monmouth County . Each BGS trap was placed in residential backyards ( near vegetation or shade ) of each parcel selected , and was operated for 24 hours prior to collection . Each week , traps were placed in the same location within the backyards . The BGS trap was used with a solid BG-lure ( Biogents AG , Regensburg , Germany ) containing ammonia , lactic acid and fatty acids , components known to be attractive to Ae . albopictus [37] . Although the BGS trap was designed to capture host seeking ( unfed ) Aedes ( Stegomyia ) mosquitoes [39] , the trap also captures other species such as Culex mosquitoes [37] , [42] in addition to occasionally collecting female mosquitoes in varying gonotrophic stages ( unengorged , blood fed , black blooded , and gravid ) . An unengorged or unfed mosquito does not contain visible evidence of blood in the abdomen , while a blood fed mosquito displays a distended abdomen with reddish blood clearly visible . A black blooded specimen has digested most of the blood meal and retains only a small portion of dark red or black blood visible near the ventral anterior of the abdomen , corresponding with Sella stage VI [48] . Gravid specimens have completely digested blood meals and contain visible eggs ready for oviposition . Collections were placed on dry ice immediately and transported to the laboratory for identification and pooling . Species identification , enumeration , and gonotrophic stage determination was conducted under a dissecting microscope using a chill table to maintain a cold chain . Specimens were stored at −80°C for subsequent blood meal determination . Abdomens of blooded Ae . albopictus were dissected over a chill table and then extracted using a Qiagen DNeasy Blood and Tissue Kit ( Qiagen Sciences , Germantown , MD , USA ) . Specimens with very small blood remnants or those deemed poorly preserved ( desiccated ) , were not utilized for DNA extraction because those samples rarely yield useful data [49] . To avoid contamination , forceps were flamed between extractions . To save time and reagents , we used a strategy that allows rapid identification of human-derived blood meals and mixes between human and non-human mammals [49] . This technique identifies human-derived blood meals based on the size of the PCR product on a gel without the need for extensive sequencing , thus drastically reducing costs . A mix between human and non-human blood is detected as two bands , and only the non-human band must be excised from the gel and purified with a QIAquick Gel Extraction Kit ( Qiagen , Valencia , CA , USA ) prior to sequencing [49] . Samples that did not amplify with the above assay were also tested with previously established primers designed for birds [50] , reptiles/amphibians [51] , and an additional primer set for mammals [52] . Approximately half of the specimens were tested with all bloodmeal identification methods above to legitimize the use of the rapid-assay [49] . To test for contamination , negative controls were employed in all reactions . The negative controls consisted of the PCR master mix with sterile water . Except for the short human-only band obtained with the Egizi et al . assay [49] , and when the non-human band was excised from the agarose gel ( see above ) , all PCR products were cleaned with Exo-Sap-IT ( USB Products , Cleveland , OH , USA ) , cycle-sequenced with the forward primer of each pair , and run on capillary automated sequencers . Sequences were BLASTed in GenBank ( http://www . ncbi . nlm . nih . gov/blast/Blast . cgi ) to compare with sequences of known species . Only matches of >98% similarity were identified as the source of the blood meal [53] . A large number of blooded Culex mosquitoes , consisting primarily of Culex pipiens pipiens L . and Culex restuans Theobald , were also collected by the BGS traps . Because of the difficulty in accurate morphological identification of field-collected specimens due to age or damage [54]–[56] these specimens are often pooled as Culex spp . After using a molecular assay to identify all Culex mosquitoes to species [57] , we tested blood fed Culex specimens from both counties collected in the same traps , locations , and dates as Ae . albopictus . Culex p . pipiens and Cx . restuans were the only Culex species collected in the BGS traps , and were assayed from Mercer County during 2009–2011 and from Monmouth County during 2008 and 2011 . Blooded Culex specimens were extracted as described above for Ae . albopictus , amplified with the BM primer pair [58] , then cleaned , sequenced , and identified as above . The BM primer pair targets a wide range of species , including mammals , birds , and reptiles , but it inadvertently amplifies in Ae . albopictus [49] and therefore cannot be used to identify blood meals in that species . Spatial differences in the proportion of Ae . albopictus feeding on selected host species between the counties was compared by using Pearsons χ2 analysis for trend . All analyses were performed using IBM SPSS Statistics 21 ( IBM , Armonk , NY , USA ) . Confidence intervals surrounding the estimated proportion of blood meals taken from a given species were calculated using the formula 95% CI = ±1 . 96× ( square root p ( 1−p ) /n ) , where p = the proportion of blood meals from a given source , and n = the total number of blood meals identified [59] .
Our BGS trap surveillance during the active mosquito seasons of 2008–2011 collected 73 , 828 Ae . albopictus females in Mercer and Monmouth Counties ( Table S1 ) . A total of 33 , 392 Ae . albopictus were collected in Mercer County , 187 ( 0 . 56% ) of which were visually determined to contain blood ( blood fed or black blooded , hereafter “blooded” ) ; while 40 , 436 Ae . albopictus were collected in Monmouth County , with 219 ( 0 . 54% ) containing blood . In Mercer County , the number and proportion of blooded Ae . albopictus collected during each month was as follows: May ( n = 1 , 1 . 25% of monthly total ) , June ( 13 , 0 . 82% ) , July ( 23 , 0 . 42% ) , August ( 70 , 0 . 57% ) , September ( 61 , 0 . 57% ) , and October ( 19 , 0 . 60% ) . Blooded Ae . albopictus in Monmouth County were collected during May ( n = 4 , 1 . 24% of monthly total ) , June ( 25 , 1 . 11% ) , July ( 65 , 0 . 99% ) , August ( 72 , 0 . 45% ) , September ( 37 , 0 . 33% ) , and October ( 16 ( 0 . 56% ) . We also captured 14 , 989 Culex mosquitoes ( Cx . p . pipiens , Cx . restuans , and Cx . spp . ) from both counties ( Table S2 ) . The BGS trap is highly specific for capturing host seeking Ae . albopictus females , as apparent by the nearly 74 , 000 specimens of this species that were captured versus the 15 , 000 specimens of Culex mosquitoes ( Tables S1 , S2 ) . Interestingly , BGS traps were also capable of capturing blooded Ae . albopictus and Culex mosquitoes , as evidenced by the collection of over 406 blooded Ae . albopictus and 745 blooded Culex ( Tables S1 , S2 ) . Of the 406 blooded Ae . albopictus collected , 117 individuals were too desiccated and therefore only 289 specimens were suitable for dissection . Subsequently , the blood meal origin of 165 ( 57 . 10% ) specimens was successfully determined ( Table S1 , 2 ) . In Mercer County , 125 were tested for host blood meal origination with a successful identification from 86 ( 68 . 80% ) specimens ( Table 2 ) . In Monmouth County , 164 Ae . albopictus were tested , with a successful host determination from 79 ( 48 . 17% ) of those specimens ( Table 2 ) . Aedes albopictus fed exclusively on mammalian hosts in Mercer and Monmouth Counties , with over 84% of all identified blood meals stemming from humans ( 52 . 12% ) , cats ( 20 . 61% ) , or dogs ( 11 . 52% ) ( Table 2 ) . Blood meals were also detected from opossums ( 4 . 24% ) , gray squirrels ( 3 . 64% ) , cottontail rabbits ( 1 . 21% ) , and a white-footed mouse ( 0 . 61% ) . A small percentage ( 6 . 06% ) of double blood meals ( from two different host species ) were detected in Ae . albopictus ( 4 . 65% of total in Mercer and 7 . 60% of total in Monmouth ) , and all included human blood ( human+dog , n = 5; human+cat , n = 4; human+deer , n = 1 ) . The number of Ae . albopictus feeding on humans was significantly higher in suburban Monmouth ( 62% ) than in urban Mercer ( 43% ) County locations ( χ2 = 8 . 151; df = 1; p = 0 . 004 ) , but significantly more Ae . albopictus fed on cats in Mercer than in Monmouth County ( χ2 = 5 . 256; df = 1; p = 0 . 022 ) . No significant difference was observed in the number of Ae . albopictus feeding on dogs between the two counties . No avian-derived blood meals were detected in any of the Ae . albopictus specimens tested . Human- and cat-derived blood meals in Ae . albopictus were detected every month of our studies , while dog-derived blood meals were absent during May ( Figure 1 ) . Only 2 . 08% of all human-derived blood meals were detected in May , while the vast majority was detected during the month of August ( 38 . 54% ) . Four contiguous months ( July , August , September , and October ) accounted for over 87% of all blood meal collections ( Figure 1 ) . We collected 745 blooded Culex ( 349 Cx . p . pipiens , 181 Cx . restuans , 215 Cx . spp . ) mosquitoes during 2008–2011 , and tested a subsample of 198 individuals identified as Cx . p . pipiens or Cx . restuans for blood meal source determination ( Table 3 ) . We selected 198 specimens to approximate the number of blood meals identified from Ae . albopictus and chose specimens from the same dates and traps as feasible . We were able to identify the blood meal source of 159 ( 80 . 30% ) samples . Blooded Cx . p . pipiens were collected during April ( n = 1 , 0 . 79% ) , May ( 19 , 15 . 08% ) , June ( 37 , 29 . 37% ) , July ( 26 ( 20 . 63% ) , August ( 19 , 15 . 08% ) , September ( 21 , 16 . 67% ) , and October ( 3 , 2 . 38% ) . Blooded Cx . restuans were collected during May ( n = 10 , 30 . 30% ) , June ( 12 , 36 . 36% ) , July ( 6 , 18 . 18% ) , August ( 2 , 6 . 06% ) , September ( 2 , 6 . 06% ) , and October ( 1 , 3 . 03% ) . In Mercer County , specimens were tested from 2009–2011 and resulted in successful host determination from 61 Cx . p . pipiens ( n = 74 , 82 . 43% ) and 7 Cx . restuans ( n = 7 , 100% ) . In Monmouth County , the blood meal hosts of 65 Cx . p . pipiens ( n = 80 , 81 . 25% ) and 26 Cx . restuans ( n = 37 , 70 . 27% ) were determined from 2008 and 2011 ( Table 3 ) . Culex mosquitoes were predominately ornithophagic ( n = 149 , 93 . 71% ) with only a small proportion feeding on mammalian hosts ( n = 10 , 6 . 29% ) ( Table 3 ) . In Mercer County , the avian blood meal hosts of Cx . p . pipiens included 16 avian species ( 88 . 52% ) , while mammalian blood meals were obtained from only three species ( 11 . 48% ) . Mammalian blood was not detected in Cx . restuans from Mercer County , whereas avian blood meals were derived from four species ( Table 3 ) . In Monmouth County , avian hosts of Cx . p . pipiens included 12 species ( 95 . 39% ) , while mammalian blood meals were obtained from only two species ( 4 . 62% ) . No mammalian blood was detected in Cx . restuans from Monmouth County and avian-derived blood meals were obtained from ten species ( Table 3 ) .
Recent decades have witnessed a dramatic global expansion of Ae . albopictus into temperate areas and an increase in locally acquired autochthonous cases of tropical diseases such as DENV and CHIKV [9] , [11] , [67] . Because of the increasing abundance of Ae . albopictus and the escalating diagnoses of exotic pathogens in travelers returning from endemic or epidemic areas [14] , the risk of a tropical disease outbreak in a new area is no longer speculative . We have shown that in urban and suburban areas of temperate northeastern USA , invasive populations of Ae . albopictus fed exclusively on mammalian hosts and that a large proportion ( 50–60% ) fed on human hosts . Although we did not detect any avian-derived blood meals from Ae . albopictus during our investigations , the species has been traditionally classified as an opportunistic feeder whose host preference is greatly dependent on the abundance of available local hosts [18] , [21] . Our studies indicate that Ae . albopictus may play a greater role in anthroponoses disease cycles , such as DENV and CHIKV , and a lesser role in zoonoses involving an avian animal reservoir . However , we cannot rule out the possibility that Ae . albopictus may occasionally act as a bridge vector for endemic pathogens such as St . Louis encephalitis virus and WNV by feeding on infected hosts when their abundance is great . Nonetheless , the large and growing populations of Ae . albopictus in major metropolitan urban and suburban centers , make a large autochthonous outbreak of an arbovirus such as CHIKV or DENV a clear and present danger . This may be particularly imminent in the case of CHIKV , as the virus is explosively spreading in the Caribbean region of the western hemisphere for the first time [68] . Given the difficulty in successful suppression of Ae . albopictus in areas where it has become firmly established [5] , [43] , we strongly recommend further ecological investigations on this species and caution public health practitioners and policy makers to install proactive measures for the imminent mitigation of an exotic pathogen outbreak .
|
Aedes albopictus is one of the most invasive and aggressive disease vectors in the world . The range of this species is currently still expanding , particularly into highly dense human population centers in temperate areas in the USA and Europe , raising the public health threat of emerging and re-emerging diseases such as chikungunya and dengue . The prominence of Ae . albopictus as a major vector was exposed during the global pandemic of chikungunya virus , primarily because of a virus adaptation which enhanced the transmission efficiency by this mosquito species and also because of the first locally-transmitted cases of chikungunya virus in temperate Europe . Blood feeding patterns by mosquitoes are a critical component of virus proliferation and determine the degree and intensity of disease epidemics , particularly in newly invaded areas . We examined the blood meal sources of invasive Ae . albopictus in the northernmost boundary of their range in temperate North America and found that the species fed exclusively on mammalian hosts , with over 90% of their blood meals derived from humans and their associated pets ( cats and dogs ) . The high mammalian affinity of Ae . albopictus suggests that this species may be an efficient vector of mammal-driven zoonoses and human-driven anthroponoses such as dengue and chikungunya viruses in this region .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"infectious",
"disease",
"epidemiology",
"tropical",
"diseases",
"animals",
"parasitology",
"global",
"health",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"zoology",
"veterinary",
"science",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"disease",
"ecology",
"veterinary",
"epidemiology",
"epidemiology",
"dengue",
"fever",
"disease",
"vectors",
"insects",
"arthropoda",
"mosquitoes",
"west",
"nile",
"fever",
"entomology",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"organisms"
] |
2014
|
Comparative Host Feeding Patterns of the Asian Tiger Mosquito, Aedes albopictus, in Urban and Suburban Northeastern USA and Implications for Disease Transmission
|
Multiple sclerosis ( MS ) is an inflammatory disease of the central nervous system characterized by myelin loss and neuronal dysfunction . Although the majority of patients do not present familial aggregation , Mendelian forms have been described . We performed whole-exome sequencing analysis in 132 patients from 34 multi-incident families , which nominated likely pathogenic variants for MS in 12 genes of the innate immune system that regulate the transcription and activation of inflammatory mediators . Rare missense or nonsense variants were identified in genes of the fibrinolysis and complement pathways ( PLAU , MASP1 , C2 ) , inflammasome assembly ( NLRP12 ) , Wnt signaling ( UBR2 , CTNNA3 , NFATC2 , RNF213 ) , nuclear receptor complexes ( NCOA3 ) , and cation channels and exchangers ( KCNG4 , SLC24A6 , SLC8B1 ) . These genes suggest a disruption of interconnected immunological and pro-inflammatory pathways as the initial event in the pathophysiology of familial MS , and provide the molecular and biological rationale for the chronic inflammation , demyelination and neurodegeneration observed in MS patients .
Multiple sclerosis ( MS ) is a common autoimmune disease of the central nervous system ( CNS ) affecting over two million people worldwide [1] . Although described as early as the 14th century , it was Jean-Martin Charcot in 1868 who recognised MS as a distinct entity , and provided the first detailed description of its clinical and pathological features [2] . Knowledge of the biological processes involved in the onset of MS have advanced greatly , and an increasing number of disease-modifying treatments ( DMTs ) have been approved since the 1990s; however , a cure has remained elusive [3 , 4] . A better understanding of the molecular mechanisms orchestrating the disruption of biological processes in MS patients is critical for the development of efficacious treatments that address the causes of MS and its progression , enhance remyelination , and prevent axonal loss and disability [5] . Large scale genome-wide association studies ( GWAS ) have already identified more than 200 genes that can moderately affect the individual’s susceptibility to the disease [6] . Given the large size of these case-control studies , risk variants that remain undiscovered to date are expected to be individually rare . Thus , we implemented high-throughput second generation sequencing technologies in multi-incident MS families for the identification of rare disease-causing variants . Although the majority of patients do not present a family history of MS , the prevalence of familial aggregation has been estimated at 12 . 6% globally [7] . In these families , rare variants co-segregating with MS are likely to account for the highest attributable risk towards the disease; however , additional genetic and environmental factors are expected to play a significant role in the presentation of clinical symptoms , level of disability , disease progression , penetrance and onset age [8 , 9] . The application of whole-exome sequencing ( WES ) in MS families has already nominated pathogenic mutations in NR1H3 , P2RX4/P2RX7 , NLRP1 and GALR2 [9–12] . Although only one of these discoveries has been replicated [13] , mutations responsible for Mendelian forms of MS highlight the molecular mechanisms underlying the cause of disease , and provide the means for the generation of new cellular and animal models of MS based on human genetic etiology [9] . The comprehensive characterization of the biological pathways disrupted in these models will nominate targets for pharmaceutical intervention trials and precision medicine approaches . In addition , genetic screening for these pathogenic variants will enable the identification of at risk individuals , provide confirmation of diagnosis , and facilitate the prediction of disease prognosis and treatment efficacy [14 , 15] . This is critical to improve quality of life for MS patients , as early diagnosis and selection of effective DMTs have been associated with improved patient outcomes , and reduced accumulation of irreversible neurological damage [16] .
Fibrinolysis is the process responsible for dissolving fibrin of blood clots and promote tissue repair and remodeling following vascular lesion . Plasmin is the primary fibrinolysin , and is the active enzyme from the proteolysis of plasminogen ( PLG ) by serine proteases , plasminogen activator tissue type ( PLAT ) or PLAU ( Fig 3 ) [18] . Components of the PLG activation system have been found to play a role in cardiovascular diseases , cancer proliferation , and inflammatory diseases , including sepsis , metabolic disease , and arthritis [19] . In MS , a rare genetic variant in PLG ( p . Gly420Asp ) was found to be over-represented in patients compared to healthy controls [20] . In addition , several neurological diseases , including MS , present abundant CNS deposition of fibrinogen resulting in microglial activation , axonal damage , and inhibition of oligodendrocyte differentiation and remyelination [21 , 22] . Interestingly , fibrin has been suggested as a promising therapeutic target for neurological diseases , as its depletion is protective against inflammatory demyelination in animal models [21] . The complement system consists of a large collection of plasma proteins that can be activated in a cascade-like fashion in response to invading pathogens and damaged host cells . Crosstalk between the fibrinolysis and complement systems has been well described , and includes plasmin which is capable of effectively cleaving complement components C3 and C5 into their active forms [23] . The activation of the complement leads to opsonisation of pathogens for phagocytosis , anaphylatoxin production to promote inflammation , and the assembly and deposition of the membrane attack complex which disrupts membrane integrity resulting in the death of targeted bacteria and infected or damaged cells ( Fig 3 ) [24 , 25] . The complement system has been linked to MS pathophysiology , with deposition of active complement components within brain plaques , peri-plaques and adjacent white matter regions [26 , 27] . In addition , complement components play a role in microglial activation , neuroinflammation , and synaptic loss in neurodegenerative diseases [28 , 29] . An inflammasome is a cytosolic protein complex that is critical for secretion of interleukin ( IL ) -1β and IL-18 , initiating an inflammatory cascade and inducing pyroptosis . Although the majority of studies support a central role for inflammasomes in the innate immune response , a role in T-cell biology has also been suggested . The assembly of the inflammasome is activated by pattern-recognition receptors ( PRR ) sensing pathogen-associated molecular patterns ( PAMPs ) and danger-associated molecular patterns ( DAMPs ) , or changes in intracellular cation concentrations [49] . Several PRR sensor molecules can activate inflammasome complex formation , and include nucleotide-binding domain ( NOD or NACHT ) -leucine rich repeat ( LRR ) -pyrin domain ( PYD ) -containing proteins ( NLRPs ) and NATCH-LRR-caspase activation and recruitment domain ( CARD ) -containing proteins ( NLRCs ) [50] . Each NOD-like receptor ( NLR ) is activated by unique stimuli and promote the formation of a specific inflammasome . The assembly of the inflammasome complex serve as a scaffold for the recruitment of the apoptosis-associated speck-like protein containing a CARD ( ASC ) adaptor , encoded by PYCARD , and oligomerization of the inactive zymogen pro-caspase-1 , initiating its autoproteolytic cleavage and activation ( Fig 3 ) . Caspase-1 then cleaves cytokine precursor pro-IL-1β and pro-IL-18 into their biologically active forms which are secreted and trigger a potent inflammatory response [51 , 52] . A subgroup of NLRs , including NLRP12 , NLRC3 and NOD2 are capable of enhancing or attenuating inflammatory signaling cascades by modulating diverse signaling pathways , including the NF-κB and extracellular signal-regulated kinase ( ERK ) pathways , which regulate the expression of inflammasome components , cytokines and chemokines [49 , 53] . Mutations in several inflammasome components can cause autoinflammatory syndromes . Activating mutations in NLRP3 cause cryopyrin-associated periodic syndromes ( CAPS ) , which is characterized by systemic inflammation with fever and blood neutrophilia [52] . Interestingly some low penetrance NLRP3 CAPS mutations have been described in patients diagnosed with MS , suggesting a role for the inflammasome in the onset of autoimmune diseases [54] . This is further supported by studies showing an increased expression of NLRP3 inflammasome-related genes in RRMS patients compared to controls [55] . In addition , the study of multi-incident MS families have nominated pathogenic mutations in NLRP1 , and purinergic receptors P2RX4/P2RX7 which initiate inflammasome formation by modifying intracellular calcium and potassium concentrations [10 , 11] . Rare missense variants in NLRP5 and NLRP9 have also been found to correlate with disease course and severity in MS patients [8 , 56] . The Wnt signal transduction pathway regulates multiple biological processes including cell proliferation , migration , polarity , differentiation and axon outgrowth . Wnt proteins have also been shown to regulate effector T-cell development , regulatory T-cell activation and dendritic-cell maturation , and to play an important role in the expression of inflammatory mediators during bacterial infections [64 , 65] . At least three Wnt-dependent pathways have been proposed; one canonical Wnt/β-catenin pathway , and two non-canonical pathways , which include the Wnt/Ca2+ pathway activated through the nuclear factor of activated T-cells ( NFAT ) . Activation of the canonical pathway is initiated through binding of Wnt ligands to Frizzled ( FZD ) receptors , causing an accumulation of β-catenin in the cytosol and subsequent translocation to the nucleus , where it forms an active transcription factor complex with T-cell factor/lymphoid enhancer factor ( TCF/LEF ) [64 , 66] . Activation of the non-canonical Wnt/Ca2+ pathway by FZD receptors triggers calcium release from intracellular stores . Increased intracellular calcium concentrations in turn activates , calcineurin , which dephosphorylates NFAT unmasking the nuclear localization sequence , facilitating nuclear translocation and activation of Wnt/Ca2+ target genes ( Fig 3 ) [67] . Wnt signaling in microglial cells induces a strong pro-inflammatory response through the activation of the canonical Wnt/β-catenin pathway . This activation mediates increased expression of cytokines , including IL-6 , IL-1α and IL-15 , chemokines such as C-X-C motif chemokine ligand 2 ( CXCL2 ) , CXCL11 , and C-C motif chemokine ligand 7 ( CCL7 ) , innate immune response components including complement C3 , and inflammasome components NLRP3 and NOD2 [68] . The Wnt/β-catenin pathway is also a major key signaling mechanism for myelinating processes , as well as oligodendrocyte development and differentiation [69–71] . The canonical Wnt pathway is also required for angiogenesis in the CNS , maturation of the blood-brain-barrier ( BBB ) , and reduced immune cell infiltration [72 , 73] . Activation of the Wnt/Ca2+ pathway regulates cytokine production in T-cells , but it is also integral for T-cell proliferation , differentiation , and activation [74] . Activation of both , canonical and non-canonical Wnt pathways , is protective against neurotoxic injury and has been found to be deregulated in degenerative and inflammatory CNS disorders [75] . In the EAE model of MS , activation of the canonical Wnt signaling pathway promotes neurogenesis and repair , whereas its inhibition results in exacerbated clinical scores [73 , 76] . In contrast , activation of the Wnt/Ca2+ pathway in EAE mice triggers an amplified pro-inflammatory response [77] . Nuclear receptors are ligand-activated transcription factors that play integral roles in many physiological processes by directly regulating gene expression . These processes include metabolism , immunity , homeostasis , cell proliferation and development , amongst others [102] . In general , nuclear receptors bind to promoter-specific DNA sequences and interact with co-repressor complexes to inhibit gene expression . Ligand-induced activation of nuclear receptors triggers the dissociation of inhibitory complexes , and the recruitment of nuclear receptor co-activator complex components that promote gene transcription [102] . Amongst others , the nuclear receptor family of transcription factors include vitamin D receptor ( VDR ) , peroxisome proliferator activated receptors ( PPARs ) , and liver X receptors ( LXRs ) , which have been shown to play important roles in the pathophysiology of MS . VDR is expressed in immune cells , and modulates the innate and adaptive immune responses . In addition , Vitamin D insufficiency is common in MS patients , and was found to correlate with disease activity , disability , and progression [103 , 104] . The activation of PPAR and LXR have been shown to inhibit canonical and non-canonical Wnt pathways , and the NF-κB signaling pathway; resulting in dysregulated inflammatory response and impaired remyelination [69 , 105–107] . These findings are supported by studies in the EAE model of MS , which showed that PPAR and LXR-deficient mice presented an exacerbated clinical phenotype , higher cytokine production , and more severe demyelination compared to wild-type or untreated animals [107–110] . In addition , LXR-α which is encoded by NR1H3 , was found to harbor a rare p . Arg415Gln mutation co-segregating with MS in two multi-incident families , and common alleles resulting in increased disease susceptibility [9] . Although this association was initially controversial [96 , 111] , it has now been independently replicated [6 , 112] . Compared to the extracellular medium , most mammalian cells have low concentrations of sodium and calcium , and large concentrations of potassium ions . This cation imbalance is regulated by membrane permeability and ion exchangers , which are critical to maintain cellular homeostasis . In cells of the innate and adaptive immune systems , ion channels and ion transporters modulate membrane potentials and regulate several physiological functions , including gene expression , apoptosis , proliferation , and migration [116] . Oscillations in intracellular calcium concentrations , due to an intricate interplay between calcium , potassium , sodium and chloride channels in the plasma membrane as well as intracellular organelles , regulate the function of many enzymes and transcription factors implicated in lymphocyte development , innate and adaptive immune responses , and autoimmunity [116] . A role for ion transporters in the pathophysiology of MS is supported by upregulation of calcium and potassium channels in MS patients , triggering apoptotic signals , demyelination and neuronal degeneration [117] . In addition , significant associations with MS risk , and pathogenic mutations have been described in P2RX4 and P2RX7 , non-selective cation channels activated by extracellular ATP [10 , 118] , and a missense variant in calcium voltage-gated channel subunit alpha1 H ( CACNA1H ) was found nominally associated with MS clinical course [8] . Small molecules capable of modulating voltage-gated calcium , sodium , and potassium channels have been developed to treat pain , stroke , migraine , epilepsy , cancer , and autoimmune disorders amongst others; and are thought to provide a good basis for the development of novel MS treatments [119–121] .
The existence of Mendelian forms of MS has been a recurrent topic of controversy , despite the evidence for familial aggregation , and the measurable increased disease risk for blood-relatives of MS patients [7 , 143] . In this study we present the genetic characterization of 34 multi-incident MS families , which have nominated pathogenic variants in 12 genes . Therefore , our data support the existence of Mendelian forms of MS , which can be attributed to a single rare variant of major effect that is largely responsible for the onset of MS and its transmission across generations . However , it should be noted that replication of our findings is warranted as the extremely low MAF observed for these variants , and the relatively low number of carriers within families , precludes sufficient statistical power for meaningful linkage and association analysis . A monogenic cause for MS could not be identified for 22 families . This was not an unexpected outcome given that complex diseases frequently are genetically heterogeneous , even within families [144 , 145] . In these families , pathogenic variants might have been overlooked given that WES technologies are not only unable to assess variants in non-coding regulatory regions , but also do not efficiently capture and sequence all coding exons , and are largely unsuited for the identification of copy number variations and rearrangements which may be responsible for the onset of disease [146] . It is also plausible that our reduced penetrance and phenocopy frequency thresholds are overly stringent , resulting in the exclusion of disease-relevant variants . The genes harboring rare disease-causing variants for familial MS , herein or previously described [9–12] , play critical roles in cellular cation homeostasis , and the regulation of transcription and activation of inflammatory mediators; suggesting a disruption of the innate immune system as the common underlying biological mechanism for the initiation of MS symptoms ( Fig 3 ) . Variants in PLAU , MASP1 , and C2 , as well as risk alleles in PLG and PLAU [6 , 20] , suggests a disruption in the fibrinolysis and complement cascade in response to microbial threads or cellular debris as a trigger for MS . In addition , PLAU activation increases angiogenesis , which has been associated with MS severity , and sustains the inflammatory response by providing oxygen and nutrients to the sites of inflammation [147 , 148] . Complement genes are necessary for the generation of anaphylatoxins C3a and C5a , opsonisation of pathogens , and formation of the membrane attack complex [39] . Inhibition of the complement system has been shown to reduce the expression of inflammatory mediators , and promote the activation of anti-inflammatory pathways , including the LXR and PPAR nuclear receptor pathways , halting neuroinflammation in the chronic relapsing EAE model [149] . Anaphylatoxins C3a and C5a bind to their corresponding membrane-bound receptors ( C3aR , C5aR1 and C5aR2 ) on the surface of monocytes and macrophages regulating the aggregation of the inflammasome ( Fig 3 ) . This complex regulatory mechanism activates or inhibits inflammasome formation in distinct cell types [150 , 151] , and is mediated through the mobilization of calcium and potassium cations from the extracellular space and intracellular stores to the cytoplasm [26] . Activation of C5aR directly promote the influx of extracellular calcium and release from intracellular stores , whereas activation of C3aR triggers the efflux of intracellular ATP which activates purinergic receptors that mediate calcium influx and potassium efflux through the plasma membrane [152–154] . These include purinergic receptors P2RX4 and P2RX7 in which digenic mutations for familial MS and risk alleles have been described [10 , 118] . Regulation of the inflammasome has also been observed in response to increased intracellular calcium concentrations due to sublytic deposition of the membrane attack complex , or decreased expression of NLRP3 and increased expression of NLRP12 in response to C1q ( complement component 1 , q subcomponent ) ; indicating that numerous elements of the fibrinolysis and complement cascades are capable of regulating the inflammatory response ( Fig 3 ) [150 , 155] . A disruption of cellular cation homeostasis in the pathophysiology of MS is further supported by disease-causing variants for multi-incident MS families in potassium channel KCNG4 and cation exchangers SLC24A1 and SLC8B1 ( Fig 2 ) . The activation of the inflammasome has been proposed as a mechanism of autoimmunity in MS patients [156] , a hypothesis that is supported by rare variants in inflammasome components NLRP1 , NLRP3 , NLRP5 and NLRP9 which were identified in MS families , or found to correlate with disease course and severity [8 , 11 , 54 , 56] . In this study we describe two missense substitutions in NLRP12 , a NOD-like receptor family member that negatively regulates inflammation and NF-κB signaling , while promoting T-cell activation and differentiation [53] . Mutations in NLRP12 that cause increased secretion of IL-1β have been described in patients with FCAS , and supports a role for NLRP12 in the onset of autoimmune diseases [52] . Interestingly , the NLRP12 p . Leu972His substitution identified in MS patients seems to have the opposite effect ( Fig 5C ) , which may explain why the p . Arg352Cys substitution associated with FCAS had similar frequencies in MS patients and controls , and failed to co-segregate with disease in families . Activation of purinergic receptors and increased cytosolic calcium concentrations also regulate Wnt signaling pathways by inhibiting glycogen synthase kinase-3-β ( GSK3β ) and activating the phosphatase activity of calcineurin [90 , 91 , 157] . GSK3β inhibits both the canonical and non-canonical Wnt signaling pathways by phosphorylating β-catenin and NFAT , thus promoting their nuclear export and degradation [67 , 87 , 158] . Interestingly , inhibition of GSK3β has been shown to accelerate myelin debris clearance and axonal remyelination [159] . Calcium-bound calcineurin dephosphorylates NFAT which translocates to the nucleus and activates the transcription of the Wnt/Ca2+ target genes , including several cytokines , chemokines , and PLAU ( Fig 3 ) [160] . Activated calcineurin also blocks NF-κB and MAPK pathways , inhibiting toll-like receptor signaling in response to pathogens and cellular damage , thus providing another means of innate immune regulation [161] . NFATC2 , is one of five NFAT transcription factors , and one of the four genes in the Wnt signaling pathway found to harbor disease-causing variants for familial MS ( Fig 2 ) . In the Wnt/Ca2+ pathway , we also identified rare missense variants in RNF213 , which targets NFATC2 for proteasomal degradation [97] . As previously described , mutations in RNF213 are associated with MMD , a progressive cerebral angiopathy which may lead to cerebral infarction , but also quasi-MMD which encompasses various clinical entities including autoimmune disease and atherosclerosis [100] . RNF213 mutations cause these phenotypes through the disruption of cerebral blood flow and reduction of angiogenesis [162] . It is unclear whether KCNG4 and SLC24A1 also play a key role in the activation of the Wnt/Ca2+ signaling pathway; however , rapid calcium influx via plasma membrane channels , which is buffered by mitochondrial calcium uptake and slow release through SLC8B1 , has been proposed as the mechanism for sustained activation of NFATC2 [139 , 163] . Interestingly , mitochondrial dysfunction and impaired calcium sequestration amplify NLRP3 inflammasome signaling [164 , 165] . In the canonical Wnt/β-catenin pathway , which is upregulated in response to demyelinating events [159] , we identified pathogenic variants in UBR2 and CTNNA3 ( Fig 2 ) . The activation of the canonical Wnt pathway modulates the immune response by initiating a pro-inflammatory signaling cascade , which includes several cytokines and chemokines , and complement and inflammasome components ( Fig 3 ) [68 , 69] . UBR2 has been shown to regulate the activation of the canonical Wnt pathway upstream of β-catenin; and although its mechanism of action still remains to be resolved , depletion of UBR2 leads to reduced expression of β-catenin target genes [78] . In contrast , CTNNA3 as well as CTNNA1 and CTNNA2 , the other two members of the α-catenin protein family ( Fig 4 ) , inhibit the Wnt/β-catenin pathway [81] . These proteins also play an important role in cell-cell adhesion in ependymal cells , and thus variants identified in MS families could not only disrupt the Wnt signaling pathway , but also BBB integrity [72 , 80] . In oligodendrocyte and glial cells , the expression of major components of the canonical Wnt signaling pathway , including β-catenin , is regulated by oxysterols and LXRs [107 , 166] . Oxysterols , which can modulate the innate and adaptive immune response , bind LXRs activating the nuclear receptor complex and promoting the initiation of transcription [167] . Genetic association and familial mutations for MS have been described in components of the oxysterol synthesis pathways and nuclear receptor complex [6 , 9 , 168 , 169] . In addition , a missense variant in nuclear receptor co-activator NCOA3 , causing increased expression of inflammatory mediators in microglial cells ( Fig 5D ) , was identified in three multi-incident MS families ( Fig 2 ) . Although NCOA3 is considered a co-activator that directly binds to nuclear receptors promoting transcriptional activities , it can also serve as a co-activator for NF-κB , enhancing the expression of target genes and maintaining the immune response [170 , 171] . Moreover , molecules acting as NR1H3 ( LXR-α ) agonists have been shown to inhibit NLRP3 inflammasome by downregulating the expression of its components , and NF-κB signaling by suppressing the phosphorylation of IκB ( Fig 3 ) [172]; thus providing additional links between these pro-inflammatory pathways . The identification of rare variants co-segregating with disease in families and genetic associations in components of the nuclear receptor complex and oxysterol synthesis pathways , suggest an important role for these genes in the pathophysiology of MS by regulating not only the synthesis of inflammatory mediators , but also neuronal development , oligodendrocyte differentiation and myelin synthesis [114 , 115 , 166 , 173 , 174] . Although replication of our findings in additional multi-incident MS families is necessary to confirm a pathogenic role for these genes and rare variants , they suggest disruption of innate immunity , inflammation , angiogenesis and cation homeostasis as critical processes in the onset of Mendelian forms of MS . Although these genes provide a mechanistic insight into the etiology of disease , it should be noted that not all family members harboring the nominated disease-causing variants developed MS ( Fig 2 ) . Therefore , despite the highly susceptible genetic background created by these variants , additional genetic , epigenetic or environmental factors are likely required to trigger the onset of MS . Although the variants identified in these families are rare , they provide the means for the development of cellular and animal models based on human genetic etiology . Models in which to further characterize the biological pathways disrupted in MS patients , and develop and assess the efficacy of novel therapeutic options tackling the pathophysiological processes of MS . In addition , we envision gene screening being used as a tool for disease confirmation , and accurate risk assessment in healthy family members of MS patients . Following the confirmation of pathogenicity in additional MS families , and with the knowledge gained from characterizing newly developed models of MS based on the identified variants , we foresee the development of personalized treatments for MS patients , and preventative strategies for at risk individuals . These may include PPAR and LXR agonists for MS patients and unaffected family members harboring substitutions in NCOA3 , and calcium channel blockers for those with variants in SLC24A1 or SLC8B1 . In conclusion , the implementation of WES in multi-incident MS families have nominated pathogenic variants in 12 genes , which highlight innate immunity and inflammatory response as critical processes leading to the onset of MS . A global effort towards the analysis of additional MS families , and the characterization of the biological processes disrupted by these variants , is necessary to expand our knowledge and understanding of the molecular and biological mechanisms underlying the genesis of MS . This gained knowledge is essential to drive the development of personalized medicine approaches with the potential to improve treatment efficacy and patient prognosis .
A total of 33 multi-incident MS families of European descent from Canada and 1 from Germany were selected for this study . DNA was available for 191 family members diagnosed with MS , 423 unaffected family members and 48 married-in individuals . In each family , DNA was available for at least 4 MS patients ( mean = 5 . 46 , SD ± 1 . 58 , range = 4–11 ) . Additional samples from Canada , and Italy were available for the replication of all nominated variants . NCOA3 p . Arg485Cys was additionally genotyped in cohorts from Spain and Austria . The Canadian cohort was collected through the longitudinal Canadian Collaborative Project on the Genetic Susceptibility to Multiple Sclerosis ( CCPGSMS ) , and consists of 13 , 870 samples ( 2 , 502 MS probands which include 2 , 039 with a family history of MS , 2 , 390 additional family members diagnosed with MS , 7 , 903 family members free of MS symptoms , and 1 , 075 unrelated healthy controls ) [175 , 176] . The male to female ratio for MS probands and unrelated controls was 1:2 . 76 and 1:0 . 96 , respectively; and with a mean age at onset for MS patients of 30 . 8 years ( SD ± 9 . 6 ) . The Spanish cohort consisted of 3 , 200 MS patients and 2 , 803 healthy controls , with a male to female ratio of 1:1 . 88 and 1:1 . 49 , respectively , and a mean age at MS onset of 30 . 7 years ( SD ± 11 . 7 ) . The Austrian cohort consisted of 552 MS patients with a male to female ratio of 1:2 . 4 and a mean age at MS onset of 31 . 4 years ( SD ± 9 . 7 ) . The Italian cohort included 46 MS patients and 32 healthy relatives from 15 multi-incident families recruited as part of the InTegrative Analysis of famiLies with MultIple Sclerosis of ItaliAN Origins ( ITALIANO ) multicenter study . The large majority of CCPGSMS probands self-report European descent ( 98 . 0% ) , and the remainder reported Asian ancestry ( 1 . 6% ) , African ancestry ( 0 . 3% ) or First Nations ( 0 . 1% ) . Samples from European cohorts are of Caucasian ancestry . All patients were diagnosed with MS according to Poser or McDonald criteria [177 , 178] . The ethical review boards at each institution approved the study [University of British Columbia ethical review board ( H08-01669 ) ; Medical University of Vienna ethics committee ( EK Nr:2195/2016 ) ; San Raffaele Ethical Committee ( NEUFAM ) ; Comité Ético de Investigación de Euskadi ( CEIC_E300911 ) ; Fondo de Investigaciones Sanitarias , Instituto de Salud Carlos III—Fondo Europeo de Desarrollo Regional ( FIS PI13/00879 and PI16/01259 ) ; Hospital Regional Universitario de Málaga ( CTS7670/11 , sample collection: C-36-003 ) ; and Hospital Virgen Macarena de Sevilla ( PI13/01527 and 2254 ) ] , and all participants provided written informed consent . WES data for Canadian samples was generated on an Ion Torrent Proton ( Thermo Fisher Scientific ) system with a 100× minimum average sequencing depth . The Ion Torrent Server v4 was used to map reads to NCBI Build 37 . 1 reference genome using the Torrent Mapping Alignment Program ( TMAP ) and to identify variants differing from the reference . Sequences with a mapping Phred quality score under 20 , fewer than five reads or over 95% strand bias were excluded from further analysis [9 , 169] . German and Italian samples were sequenced on a HiSeq 2500 ( Illumina ) , and the raw sequences were aligned against the human reference genome ( hg19 ) with BWA and processed with a GATK best practices pipeline using Unified Genotyper variant caller . Annotation of variants was performed with ANNOVAR [179] . WES data for 132 MS patients from 34 families was generated for the identification of pathogenic variants ( S4 Table ) . Heterozygote non-silent variants identified in WES data from all patients in a single family , and with a MAF below 1% in public ( ExAC ) or proprietary databases of variants [17] , were genotyped in all family members to validate WES genotype calls and assess segregation with disease , and Canadian MS probands and healthy controls to assess population frequencies , as previously described [9] . To account for reduced penetrance and the presence of phenocopies , variants were deemed to segregate with disease when found in at least 75% of individuals diagnosed with MS and no more than one unaffected family member , excluding unaffected parents of MS patients . When a variant segregating with disease could not be found , additional affected family members for whom DNA was available were analyzed by WES , and rare non-silent variants identified in all but one MS patient were assessed for segregation with disease . Additional variants in each gene of interest were identified by mining WES data from 426 MS patients from Canada , 15 probands from multi-incident MS families from Italy , 100 healthy controls from Canada , and 955 multi-ethnic diseased controls . Missense or nonsense variants identified exclusively in MS patients , and with a MAF below 1% in public databases of variants [17] were assessed for segregation within families . All variants deemed to co-segregate with disease were genotyped using Sequenom MassArray iPLEX platform or TaqMan genotyping probes ( Tables 1 & S5 ) . For every additional patient identified harboring a variant of interest , all blood-related family members for whom DNA was available were genotyped using Sanger sequencing to confirm genotype calls and assess segregation with disease as previously described [169 , 180] . Haplotype analysis were performed using microsatellite markers spanning each locus of interest . Primer sequences are available at the National Centre for Biotechnology and Information ( https://www . ncbi . nlm . nih . gov/probe ) . PCR reactions were performed under standard conditions with one primer pair for each marker labeled with a fluorescent tag . PCR products were run on an ABI 3730xl ( Applied Biosystems ) and analyzed using GeneMapper 4 . 0 . Marker sizes were normalized to those reported in the Centre d’Etude du Polymorphisme Humain ( CEPH ) database ( http://www . cephb . fr/ ) . Plasmids containing FLAG-tagged full length wild-type , p . Gln475 or p . His972 human NLRP12 were kindly donated by Dr . Beckley Davis ( Franklin & Marshall College , USA ) . Full length cDNA encoding wild-type NCOA3 was PCR amplified from total human brain cDNA , and the p . Arg485Cys substitution introduced by fusion PCR . After restriction digestion , PCR products were inserted into pcDNA4-myc-his A ( pZ ) between KpnI and XhoI . 1–2μg of empty vector or expression vectors for NLRP12 wild-type ( WT ) , NLRP12 p . Gln475 ( L475Q ) , NLRP12 p . His972 ( L972H ) , NCOA3 wild-type ( WT ) , or NCOA3 p . Cys485 ( R485C ) was transfected into mouse microglial cell line BV2 using polyethylenimine ( PEI ) . Twenty-four hours after transfection , whole cell lysates were subjected to Western blot for Caspase-1/p10 ( Santa Cruz , Cat# sc-56036 , RRID: AB_781816 ) , FLAG-NLRP12 ( Sigma , Cat# F7425 , RRID: AB_439687 ) and β-actin ( Sigma , Cat# A5316 , RRID: AB_476743 ) , or iNOS ( Cell Signaling Technology , Cat# 13120 , RRID: AB_2687529 ) , NCOA3 ( Cell Signaling Technology , Cat# 2126S , RRID: AB_823642 ) and β-actin . Activation of NF-κB was assessed in HEK293 cells transfected with 200ng of a pcDNA4 vector containing the coding region for either wild-type , p . Gln475 or p . His972 NLRP12 , together with 200ng of a reporter PGL3 plasmid with a response element for NF-κB . Cells were co-transfected with a vector to express p65 in order to induce NF-κB activation . 15ng of a vector containing Renilla Luciferase was transfected into cells as an internal control . Twenty-four hours post-transfection cells were lysed , and luciferase activity in cell lysates was measured using a luciferase assay kit ( Promega , Cat# E1500 ) . All experiments were performed at least in triplicate , and protein bands quantified with Quantity One ( Bio-Rad ) . One-way ANOVA and Tukey’s Honest Significant Difference ( HSD ) post hoc test were used to identify statistically significant differences between groups .
|
Although the majority of patients diagnosed with multiple sclerosis do not have a family history of disease , 13% report having a close relative also diagnosed with multiple sclerosis . In these families , the cause of multiple sclerosis can be largely attributed to a single genetic variant that is transmitted through generations . In this study we analyzed DNA from 132 patients from 34 families , resulting in the identification of 12 rare genetic variants that are largely responsible for the onset of multiple sclerosis in these families . These variants are located in genes implicated in specific immunological pathways , and suggest the biological mechanisms that trigger the onset of multiple sclerosis . These genes and variants provide the means for the generation of cellular and animal models of human disease , and highlight biological targets for the development of novel treatments .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"inflammatory",
"diseases",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"cytokines",
"multiple",
"sclerosis",
"neurodegenerative",
"diseases",
"immunology",
"data",
"mining",
"developmental",
"biology",
"demyelinating",
"disorders",
"clinical",
"medicine",
"inflammasomes",
"molecular",
"development",
"molecular",
"biology",
"techniques",
"information",
"technology",
"genotyping",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"computer",
"and",
"information",
"sciences",
"proteins",
"molecular",
"biology",
"immune",
"system",
"biochemistry",
"clinical",
"immunology",
"heredity",
"neurology",
"physiology",
"autoimmune",
"diseases",
"genetics",
"biology",
"and",
"life",
"sciences",
"human",
"genetics"
] |
2019
|
Exome sequencing in multiple sclerosis families identifies 12 candidate genes and nominates biological pathways for the genesis of disease
|
Kashin-Beck disease , a syndrome characterized by short stature , skeletal deformities , and arthropathy of multiple joints , is highly prevalent in specific regions of Asia . The disease has been postulated to result from a combination of different environmental factors , including contamination of barley by mold mycotoxins , iodine deficiency , presence of humic substances in drinking water , and , importantly , deficiency of selenium . This multifunctional trace element , in the form of selenocysteine , is essential for normal selenoprotein function , including attenuation of excessive oxidative stress , and for the control of redox-sensitive molecules involved in cell growth and differentiation . To investigate the effects of skeletal selenoprotein deficiency , a Cre recombinase transgenic mouse line was used to trigger Trsp gene deletions in osteo-chondroprogenitors . Trsp encodes selenocysteine tRNA[Ser]Sec , required for the incorporation of selenocysteine residues into selenoproteins . The mutant mice exhibited growth retardation , epiphyseal growth plate abnormalities , and delayed skeletal ossification , as well as marked chondronecrosis of articular , auricular , and tracheal cartilages . Phenotypically , the mice thus replicated a number of the pathological features of Kashin-Beck disease , supporting the notion that selenium deficiency is important to the development of this syndrome .
Kashin-Beck disease , an environmentally-induced musculoskeletal syndrome , is prevalent in millions of individuals residing within specific regions of Tibet , China , Siberia , and North Korea [1]–[3] . The disease first becomes evident in childhood with affected individuals exhibiting short stature , joint and limb deformities , and radiographic evidence of delayed skeletal ossification [4] , [5]; features attributed to impaired epiphyseal growth and chondronecrosis [1] , [3] , [4] , [6] . With age , severe secondary osteoarthritis of multiple joints becomes evident . Although several factors have been implicated in the pathogenesis of this disease , deficiency of dietary selenium intake , and hence , profoundly low serum selenium levels represent one of the most salient features of Kashin-Beck disease [1] , [7] , [8] . This raised the possibility that deficiencies in one or more selenoproteins might play key etiological roles in this musculoskeletal disorder . Selenium is an essential dietary micronutrient that is associated with various organic molecules , including the 21st amino acid , selenocysteine ( Sec ) , that is required for the function of a class of proteins known as the selenoproteins [9] . There are 25 human and 24 murine genes that encode for selenoproteins , approximately one third of which function as antioxidants , protecting cells against macromolecule damage caused by oxidative and/or nitrosative stress [10] . Other members of the selenoprotein family include the deiodinases , including , for example , deiodinase 2 ( D2 ) that catalyzes conversion of thyroxine ( T4 ) into the active tri-iodothyronine ( T3 ) form; the latter has many roles in growth and development , including the regulation of epiphyseal growth plate differentiation [11]–[13] . Selenoprotein synthesis requires that Sec residues be inserted into the growing polypeptide chains via specific UGA codons present within the coding regions of selenoprotein mRNAs . Although UGA specifies a ‘stop’ codon in the universal genetic code , it can also code for Sec with the participation of a group of proteins that recognize the Sec insertion sequence ( SECIS ) element located in the 3′ untranslated region of selenoprotein mRNAs [14] . UGA is recognized by Sec tRNA ( designated Sec tRNA[Ser]Sec ) which is encoded by Trsp ( present in one copy per haploid genome ) . Sec tRNA[Ser]Sec is responsible for the expression of all selenoproteins and thus it provides a unique tool for studying the role of selenium within this class of proteins . The targeted deletion of Trsp is embryonic lethal owing to a loss of activity of selenoproteins that are essential for normal growth and development [15] . Hence , investigation of the tissue-specific consequences of selenoprotein loss in vivo must be carried out via a conditional mutagenesis approach . Cre-loxP technology has been successfully used to study the consequences of floxed Trsp gene ( Trspfl/fl ) excision in cardiac muscle , hepatocytes , lymphocytes , mammary epithelium , and neurons , with pathological changes attributable to selenoprotein deficiency being demonstrated in all these tissues [9] . In addition to dietary deficiency of selenium , deficiency of iodine , ingestion of fungal mycotoxins from contaminated stored food , high humic acid levels in the drinking water , or some combination of these components have been proposed to have roles in the genesis of Kashin-Beck disease [1] , [7] , [8] , [16] . The relative contributions of the various environmental factors implicated in Kashin-Beck disease are not known but an emphasis has been placed on the role of selenium deficiency [1] , [17] , [18] . Induction of dietary selenium and/or iodine deficiency in rodents , however , fails to replicate many of the clinical features of Kashin-Beck disease , while showing only modest effects on skeletal growth and bone volume [17] , [18] . This could be due to the difficulty in attaining the levels of selenium deficiency observed in Tibetan children which are below 27 ng/ml in 90% of affected individuals . Values in the 5 ng/ml range are observed in a third of affected individuals [1] . In contrast , rats maintained on selenium deficient diets only exhibited serum levels in the 30 ng/ml range [17] . We hypothesized that interfering with cartilage selenoprotein synthesis , and hence mimicking the effects of severe selenium deprivation in this tissue , might recapitulate features of Kashin-Beck disease , including delayed endochondral bone development , impaired ossification , chondronecrosis , and dwarfism . To test this hypothesis , we deleted the Trsp gene in cells that give rise to the skeleton . This was accomplished by generating Trspfl/fl mice expressing the Cre recombinase under the control of the Col2a1 gene promoter in order to obtain Trsp deletions in osteo-chondroprogenitors [19] , [20] .
Previous studies have shown that the generalized knockout of Trsp is lethal at the embryonic stage [15] , [21] . To produce viable individuals , we crossed a mouse with floxed Trsp alleles with a transgenic line expressing the Cre recombinase under the control of the Col2a1 promoter . Col2a1-Cre mice exhibited Cre-mediated gene excision activity starting at approximately 9 days post coitum ( dpc ) in notochord and cranial mesenchyme ( sites where the endogenous Col2a1 gene begins to be expressed ) , with Cre reporter activity being evident in all cartilage primordia ( ribs , long bones , spine , basicranium ) by 15 dpc [19] . Furthermore , using the same β-galactosidase-based Cre reporter strain as Ovchinnikov et al . [19] , we detected Cre activity not only in mature cartilage and endochondrally derived bone , but also within regions of the cranium resulting from intra-membranous ossification [20] . These results suggested that osteo-chondroprogenitors were the targets for Cre-mediated genomic alterations in Col2a1-Cre transgenic mice . Compared to littermate controls , Col2a1-Cre; Trspfl/fl mice demonstrated marked reduction in skeletal and cartilage growth . As early as 1 wk after birth , the mutant mice began to exhibit dwarfism , marked auricular hypoplasia , shortened snouts , decreased head size with frontal bossing , smaller limbs and shorter tails ( Figure 1A and 1B ) . Males and females were equally affected . Interestingly , Col2a1-Cre; Trspfl/fl mice were indistinguishable from control mice within the first few days after birth . Indeed , mouse lengths ( nose to tail base ) measured on post-natal day 1 , showed no difference between the two groups ( Figure 1C ) . In contrast , by 3 . 5–4 weeks of age , the differences in length of the two groups had become significant ( Figure 1C ) , indicating that an additional factor was involved , such as exposure to ambient oxygen . The 3 . 5–4 wk time point was selected for analyses owing to the high incidence of death ( or need for euthanasia ) in 4–5 wk old Col2a1-Cre; Trspfl/fl mice ( Figure 1D ) . Moribund animals demonstrated marked rib cage indrawing , suggestive of inspiratory respiratory distress . Radiographs of littermate control and Col2a1-Cre; Trspfl/fl mice ( Figure 2A ) demonstrated the smaller skeletons of the mutant mice , and the rounding of the cranium . The spaces between the vertebral bodies , due to the intervertebral discs , were narrowed in the mutant mice as compared with the littermate controls ( Figure 2A ) . The lumbar vertebrae of the Col2a1-Cre; Trspfl/fl mice were smaller , with irregular outlines and were unevenly ossified . These findings were confirmed by microcomputed tomography ( micro-CT ) imaging of the lumbar vertebrae ( Figure 2B ) . Micro-CT also demonstrated the reduced cranial size and abnormal cranial shape in Col2a1-Cre; Trspfl/fl mice ( Figure 2C and 2D ) , with snout shortening , narrowing of the jaw , and rounding of the calvaria . This pattern is consistent with reduction in the growth of the chondrocranial component of the skull . Indeed , other mutations that perturb chondrocranial growth exhibit similar , albeit less extreme , effects on the shape of the mouse skull [22] , [23] . The more porous nature of the frontal bones in the mutant , evident in the micro-CT image of the cranium ( Figure 2C ) , may have been reflective of impaired intramembranous bone development secondary to Col2a1-Cre mediated deletions of Trsp in osteoblasts [20] . Micro-CT confirmed that smaller knees were a feature of the mutant mice ( Figure 2E ) . Quantitative micro-CT of femoral cortical bone demonstrated that the Col2a1-Cre; Trspfl/fl long bones were impaired in their mineralization ( p = 0 . 005 ) when compared to the bones of littermate controls ( Figure 2F ) . Interestingly , despite the reduced size of the knees , the width of the radiolucent gap corresponding to the tibial epiphyseal growth plates in was not correspondingly reduced in the mutant mice . Histological sections of mutant mice tibiae revealed increased epiphyseal growth plate width despite reduced bone length . The epiphyses were also characterized by a disorganized primary spongiosa in comparison to littermate controls ( Figure 3A and 3B ) . Indeed , there was an approximate 45% increase in tibial growth plate width in the mutant mice ( data not shown ) that was due to increases in both the proliferative and hypertrophic chondrocyte zones ( Figure 3C and 3D ) . This increased width was present despite the finding of decreased cell proliferation ( as shown by 5-bromo-2-deoxyuridine , BrdU , incorporation ) and only modest increases in TUNEL ( terminal deoxynucleotidyl transferase dUTP nick end-labeling ) positive apoptotic cells in the mutant growth plates ( Figure 3E–3G ) . We speculate that accumulation of cells in the proliferative zone may have been the result of a slowing of the hypertrophic differentiation process , and that hypertrophic zone expansion may have been due to impaired osteoblastic invasion of this layer . Areas of chondronecrosis , a key pathological feature of Kashin-Beck disease [3] , [4] , [6] were observed in cartilaginous tissues , including the articular cartilage , ears , and tracheal rings . The articular cartilage of a littermate control knee shown in Figure 4A exemplifies the way that cartilage tissue sections should appear when stained with hematoxylin , fast green and safranin O , with the red stain indicating the presence of the glycosaminoglycans . In contrast , areas of chondronecrosis were observed in the articular cartilage of Col2a1-Cre; Trspfl/fl knees ( Figure 4B and 4C ) , as indicated through the loss of chondrocytes and glycosaminoglycan staining . These areas of chondronecrosis were flanked by sparse TUNEL positive apoptotic cells ( Figure 4D ) . Chondronecrosis was confined primarily to the femoral condylar and tibial plateau articular cartilage , but were not evident within the tibial epiphyseal growth plates . Massive chondronecrosis was evident in the hypoplastic ears of the mutant mice ( Figure 5A and 5B ) , again these areas contained small numbers of TUNEL positive chondrocytes ( Figure 5C and 5D ) . Neutrophil infiltrates , a feature of necrotic cell death , surrounded some of the chondronecrotic areas in the ears ( data not shown ) , and these necrotic regions were often flanked by clusters of proliferating cells ( Figure 5E and 5F ) , possibly reflective of cartilage regeneration . Lastly , premature death in Col2a1-Cre; Trspfl/fl mice appeared to result from respiratory distress , and could be explained by tracheal narrowing and/or collapse upon inspiration due to loss of tracheal ring integrity due to the tracheomalacia . In keeping with this idea , we found marked hypoplasia and chondronecrosis ( Figure 6A and 6B ) , along with scattered TUNEL positive apoptotic cells ( Figure 6C and 6D ) , in Col2a1-Cre; Trspfl/fl tracheal cartilages . To the best of our knowledge , there are no reports of an increase frequency of tracheomalacia in newborns within Kashin-Beck disease endemic areas . To assess whether the targeted removal of Trsp in chondrocytes resulted in a down-regulation of selenoprotein expression , immunoblot analysis of chondrocytes for one of the major housekeeping selenoproteins , thioredoxin reductase ( TR1 ) , was performed ( Figure 7A and 7B ) . This revealed a decrease in TR1 levels in the chondrocytes of knockout mice as compared to control mice . In contrast , TR1 levels were virtually unchanged in liver lysates , which served as the control tissue . The reason a complete loss in TR1 expression was not observed in the chondrocyte preparation may have been due to contaminating cell types such as fibroblasts in the lysate preparations [24] , and/or to incomplete excision of floxed Trsp by Cre recombinase in chondrocytes [15] . Genomic PCR revealed that Trsp excisions were present in the cartilage-rich tissues ( xiphoid process , tail tip ) of Trspfl/fl Col2a1-Cre mice; however , the appearance of the un-excised band in these tissues was again likely due to the presence of non-chondrocytic cells in the tissue samples ( Figure 7C ) . There was a faint excision band present in the liver DNA samples , suggesting that there was some low level of ‘leakiness’ of the Cre transgene in this tissue . In summary , Trsp deletions were present in the cartilage-rich tissues of the mutant mice , and this was accompanied by a decrease in the expression level of an important selenoprotein , TR1 . The chondronecrosis and impaired skeletogenesis seen in Col2a1-Cre; Trspfl/fl mice potentially results from chondrocyte and osteoblast deficiencies in a range of different anti-oxidant selenoproteins , including specific glutathione peroxidases ( such as glutathione peroxidases 1 and 4 ) and , as we have seen , thioredoxin reductases ( such as TR1 ) [9] . The antioxidant selenoproteins constitute an important line of defense against cellular damage , necrosis , and apoptosis brought about by excessive levels of reactive oxygen and nitrogen species [25] . The thioredoxin reductases , for example , are required for normal thioredoxin function , and as such support cell proliferation and the activities of specific redox-sensitive molecules , such as specific transcription factors or their regulators [26]–[28] . The idea that the post-natal onset of runting in Col2a1-Cre; Trspfl/fl mice may have been due to exposure to ambient oxygen , and the ensuing increase in oxidative stress , would be consistent with anti-oxidant defenses of chondrocytes and osteoblasts having been compromised . It was also possible that intracellular T3 deficiency , resulting from decreased selenoprotein deiodinase activity , may have contributed to the phenotype of Col2a1-Cre; Trspfl/fl mice . The thyroid secretes T4 into the circulation which must then be converted within target tissues into the active T3 form by intracellular D2 [29] , [30] . T3 , a hormone utilized by all tissues , is also required for normal growth plate development [11] , and nuclear receptors for this ligand have been shown to be expressed in skeletal cells [31] . Although diminished D2 deiodinase activity could theoretically lead to a tissue-specific deficiency of T3 , genetic deletion of D2 , either alone or in combination with D1 , yielded no evidence of impaired skeletal growth [32] . This is in keeping with recent data indicating that neither D1 or D2 are expressed in rodent chondrocytes , and that D2 is only found in mature osteoblasts [33] . In contrast , D3 , a deiodinase that down-regulates intracellular T3 levels in the skeleton ( to avoid accelerated bone maturation ) , is expressed in young rodent chondrocytes and osteoblasts . Thus , mice lacking D3 exhibited generalized growth retardation attributed to perinatal thyrotoxicosis and that was subsequently compounded by severe hypothyroidism starting around the time of weaning [34] . In view of these results , we hypothesize that osteo-chondroprogenitor specific deficiency of the D3 selenoprotein in Col2a1-Cre; Trspfl/fl mice may have allowed the accumulation of abnormally high concentrations of intracellular T3 during the perinatal period . Raised T3 levels would not only act to accelerate bone maturation , but would also contribute to oxidative stress [35] , thus aggravating macromolecule damage in cells already impaired in their selenoprotein-based anti-oxidant defenses . It should be noted , however , that congenital thyroid hormone deficiency in humans has been associated with multiple skeletal effects , including short stature , vertebral and cranial abnormalities , as well as delayed endochondral bone formation and skeletal maturation [36] . These phenotypic features , which may also be due in part to neuro-endocrine axis abnormalities , are manifested primarily as delays in growth and development , and can be reversed by thyroid hormone administration . We are not aware of any evidence showing that human hypothyroidism is associated with chondronecrosis , and given the unremarkable phenotype of D1 and D2 deficient mice [32] , it is unlikely that T3 deficiency within osteo-chondroprogenitor cells would , by itself , be able to account for the skeletal phenotype of Col2a1-Cre; Trspfl/fl mice . Lastly , given that very low serum levels of selenium and iodine almost invariably coexist in individuals within Kashin-Beck disease endemic regions [1] , it raises the possibility that our dietary iodine-proficient Col2a1-Cre; Trspfl/fl mice may not entirely mimic the human syndrome . Indeed , it is conceivable that low iodine levels , and hence reduced thyroid hormone activity , might actually be protective in the presence of profound selenium deficiency . The topic of iodine and selenium supplementation in individuals residing in Kashin-Beck disease endemic areas and in those exhibiting the clinical manifestations of this disorder has been reviewed by Vanderpas [37] . In general , epidemiological studies have suggested that either selenium or iodine supplementation decrease the incidence and/or clinical severity of Kashin-Beck disease . However , selenium supplementation in established disease appeared not to have a beneficial clinical effect [38] . Regarding this latter finding , it could be argued that once clinical manifestations are present , some degree of permanent damage to joint and epiphyseal cartilage has already occurred . Clearly , selenium and iodine supplementation in Kashin-Beck endemic areas should be instituted during gestation and then continued until skeletal maturity has been reached so as to prevent growth plate and articular cartilage damage , and sequelae such as secondary osteoarthritis . We have shown that the Trsp gene , and hence , preservation of selenoprotein activity in osteo-chondroprogenitors , is essential to murine skeletogenesis and the maintenance of cartilage viability . Indeed , loss of Trsp was associated not only with defects in cartilage and bone development , but also severe chondronecrosis of auricular and tracheal cartilages . Our findings lend support to the idea of selenium deficiency being a key factor in the pathogenesis of the skeletal abnormalities of Kashin-Beck disease .
Mice with floxed Trsp alleles [15] , after being backcrossed ( N = 6 ) onto the C57BL/6J genetic background , were then interbred with mice expressing Cre recombinase under the control of a type II collagen ( Col2a1 ) gene promoter [19] . The latter were originally of a mixed background when purchased from The Jackson Laboratory ( Bar Harbor , ME ) but were subsequently fully ( N>10 ) backcrossed onto the C57BL/6J genetic background prior to their use in the experiments reported herein . Tail tip clippings from pups were taken immediately post-weaning and processed for genotyping for both the Trspfl/fl and the Cre transgene . Cartilage-specific excision of Trsp was verified using a previously PCR reaction [15] . Specific pathogen-free mice were maintained on standard mouse chow ( Pico-Vac Lab Mouse Diet #5062 , Brentwood , MO ) , and housed in a barrier facility in accordance with both University of Calgary Animal Care Committee and Canadian Council on Animal Care guidelines . For identification purposes , newborn pups were marked with dots on the base of the footpad according to the Ketchum Manufacturing's ( Brockville , ON , Canada ) tattooing protocol . Weights were determined every three to four days for up to three weeks , and length measurements ( from the tip of the nose to the base of the tail ) were taken at birth and again at 3 . 5 and 4 weeks of age . Micro-CT scanning ( vivaCT 40 , Scanco Medical , Basserdorf , Switzerland ) was performed at different locations on 4 wk old mice ( skull , spine , knees ) . Scanning was done at 10 µm isotropic resolution ( 55 kVp , 109 µA , 400 ms integration time , 2 , 000 projections on 360° , 2048 CCD detector array , and cone-beam reconstruction ) and images were taken at various orientations and cutplanes . X-ray images of the whole animal or of distinct regions were also taken using this instrument . The generalized shape images were obtained as described [39] , [40] . This refers to the mean shape as obtained through scaling , superimposition , and averaging of the volumetric image data for the entire sample . For the cortical analysis , a 1 . 5 mm section of the mid-shaft of the femur was scanned and evaluated using semi-automatically drawn counters and were thresholded and Gaussian filtered ( sigma = 1 . 2 , support = 1 ) , to form binarized images upon which measures of cortical thickness could be obtained [41] . Knees were fixed in 4% paraformaldehyde for 7 days then decalcified in 14% EDTA . Ear samples were fixed in 10% formalin , and lungs were inflated and fixed with 10% formalin delivered via a cannula inserted into the proximal trachea . Samples from knees , ears and trachea were embedded into 50∶50 paraffin blocks and cut into 4 µm sections . Histology stains included hematoxylin and eosin , and a triple-stain , consisting of hematoxylin , fast green and safranin-O . For detection of proliferating cells , mice were injected with 100 mg/kg BrdU 2 hrs prior to euthanasia; replicating cells were detected via staining with a BrdU Kit ( Zymed , Invitrogen Corporation , Carlsbad , CA ) ; alternatively , cell proliferation was assessed using an anti-PCNA antibody ( 1∶50 , Santa-Cruz Biotechnology , Santa Cruz , CA ) in combination with the Vectastain ABC Rabbit Kit , ( Burlingame , CA ) , using hematoxylin as the counter-stain . Apoptosis was measured using the TUNEL method ( Chemicon Apoptag kit , Millipore , Billerica , MA ) with methyl-green counterstaining . Quantification of proliferating and apoptotic cells was done by averaging positive cells from 2–4 serial sections within the proliferative zone of the growth plate of both the distal femurs and proximal tibiae of the mice . Rabbit antibodies raised against thioredoxin reductase 1 ( TR1 ) were prepared in the laboratory of DLH . Anti-rabbit-HRP-conjugated secondary antibody was obtained from Sigma ( St . Louis , MO ) ; NuPAGE 4–12% polyacrylamide gels and PVDF membranes were from Invitrogen ( Carlsbad , CA ) , and SuperSignal West Dura Extended Duration Substrate was from Thermo Scientific ( Rockford , IL ) . Chondrocyte samples were isolated from ribs of 10 day old mice through successive digestions with pronase ( 2 mg/ml ) and collagenase D ( 3 mg/ml ) . Chondrocyte and liver protein extracts prepared from control and knockout mice were subjected to electrophoresis on NuPAGE 4–12% polyacrylamide gels , transferred to PVDF membranes and immunoblotted with anti-TR1 antibody ( 1∶1000 dilution ) . Membranes were washed with 0 . 1% TBS-T ( 20 mM Tris/HCl , pH 7 . 5 , 150 mM NaCl and 0 . 1% Tween 20 ) and anti-rabbit-HRP-conjugated secondary antibody ( 1∶20000 ) was used . Following incubation with the secondary antibody , membranes were washed with 0 . 1% TBS-T , incubated in SuperSignal West Dura Extended Duration Substrate and exposed to X-ray film . Where applicable , samples were compared using a standard student's t-test .
|
Kashin-Beck disease ( KBD ) is a severe , chronic , and deforming musculoskeletal disease affecting millions of individuals in specific regions of Asia . Starting in childhood , the disorder leads to joint and limb deformities , short stature , and delayed skeletal development . Articular cartilage damage due to chondronecrosis and limb deformities then lead to secondary osteoarthritis and severe disability . Factors proposed to cause KBD include selenium deficiency , iodine deficiency , contamination of grain with toxic molds , and humic substances in well water . Soil and water deficiency in selenium ( and iodine ) are a consistent feature of KBD endemic areas , and affected individuals show profound deficiencies of these two elements . Thus far , there have been no convincing rodent models of KBD based on selenium ( and/or iodine ) deficiency achieved through dietary manipulation . Our manuscript describes a conditional gene mutation approach in mice that , in effect , mimics severe selenium deficiency , achieving this specifically within skeletal progenitor cells . By deleting selenocysteine tRNA ( required for normal selenoprotein activity ) in osteo-chondroprogenitors , we found that mice develop post-natal impairment of skeletal growth , dwarfism , delayed ossification , impaired endochondral bone formation , as well as severe chondronecrosis . Our mutant mouse supports the idea that selenium deficiency is key to the skeletal pathology of KBD .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"rheumatology/cartilage",
"biology",
"and",
"osteoarthritis",
"developmental",
"biology/aging",
"pathology/histopathology",
"genetics",
"and",
"genomics/disease",
"models",
"radiology",
"and",
"medical",
"imaging/computer",
"tomography",
"nutrition/deficiencies",
"cell",
"biology/gene",
"expression"
] |
2009
|
Osteo-Chondroprogenitor–Specific Deletion of the Selenocysteine tRNA Gene, Trsp, Leads to Chondronecrosis and Abnormal Skeletal Development: A Putative Model for Kashin-Beck Disease
|
Staphylococcus aureus is capable of infecting nearly every organ in the human body . In order to infiltrate and thrive in such diverse host tissues , staphylococci must possess remarkable flexibility in both metabolic and virulence programs . To investigate the genetic requirements for bacterial survival during invasive infection , we performed a transposon sequencing ( TnSeq ) analysis of S . aureus during experimental osteomyelitis . TnSeq identified 65 genes essential for staphylococcal survival in infected bone and an additional 148 mutants with compromised fitness in vivo . Among the loci essential for in vivo survival was SrrAB , a staphylococcal two-component system previously reported to coordinate hypoxic and nitrosative stress responses in vitro . Healthy bone is intrinsically hypoxic , and intravital oxygen monitoring revealed further decreases in skeletal oxygen concentrations upon S . aureus infection . The fitness of an srrAB mutant during osteomyelitis was significantly increased by depletion of neutrophils , suggesting that neutrophils impose hypoxic and/or nitrosative stresses on invading bacteria . To more globally evaluate staphylococcal responses to changing oxygenation , we examined quorum sensing and virulence factor production in staphylococci grown under aerobic or hypoxic conditions . Hypoxic growth resulted in a profound increase in quorum sensing-dependent toxin production , and a concomitant increase in cytotoxicity toward mammalian cells . Moreover , aerobic growth limited quorum sensing and cytotoxicity in an SrrAB-dependent manner , suggesting a mechanism by which S . aureus modulates quorum sensing and toxin production in response to environmental oxygenation . Collectively , our results demonstrate that bacterial hypoxic responses are key determinants of the staphylococcal-host interaction .
Staphylococcus aureus is a major human pathogen , capable of causing a variety of life-threatening , invasive diseases and infecting nearly every organ in the human body . Yet S . aureus also innocuously colonizes the skin and nares of one-third to one-half of the population [1] . These facts suggest a remarkable flexibility in terms of metabolic and virulence programs , allowing staphylococci to adapt to diverse and changing host environments during invasive infection , while also enabling a commensal lifestyle characterized by low virulence and immunotolerance . The mechanisms by which bacterial pathogens adapt to changing host environments are poorly understood , in part due to the technical difficulty in measuring adaptive responses in vivo . However , recent advances in high-throughput sequencing have enabled an unprecedented evaluation of the host-pathogen interface . Transposon sequencing ( TnSeq ) is a sensitive and high-throughput tool combining highly-saturated transposon mutant libraries with massively-parallel sequencing to calculate the fitness of all nonessential bacterial genes under a given selective pressure [2] . TnSeq has been successfully used to determine the bacterial genes required for survival in a number of different in vitro conditions and infection models [3–7] . More recently , a TnSeq library was generated in S . aureus and used to determine genes contributing to fitness in abscess and infection-related ecologies [8] . These studies illustrate the power of TnSeq analyses to determine the genetic requirements for bacterial adaptation to diverse host environments . One of the most common invasive disease manifestations of staphylococcal infection is osteomyelitis , and S . aureus is by far the most common pathogen causing musculoskeletal infection . Osteomyelitis causes enormous morbidity , including functional disabilities , the requirement for invasive procedures , and the propensity to evolve into chronic infection even with appropriate management [9 , 10] . Two factors contribute to the therapeutic recalcitrance of osteomyelitis . First , skeletal tissues are intrinsically hypoxic , and bacterial infection further disrupts the vascular architecture of bone [11–13] . Second , the human skeleton is constantly being remodeled through the opposing actions of osteoblasts and osteoclasts . The kinetics of bone remodeling are affected dramatically by bacterial infection through osteo-immunologic crosstalk [14–16] . Thus , pathogens invading the bone must adapt to hypoxia as well as constant shifts in the available host substrates for adhesion and nutrient acquisition . While these factors would seemingly create an inhospitable environment for bacterial proliferation , bone is one of the most common locations of metastatic infection following S . aureus bacteremia [17] . One mechanism by which bacterial pathogens adapt to potentially hostile host environments is through the actions of one or more two-component systems ( TCSs ) . Bacterial TCSs consist of a membrane bound histidine kinase sensor , which upon binding of its cognate ligand phosphorylates a response regulator . Response regulators most often function as transcriptional factors , and differentially coordinate changes in gene expression in response to a given stress . We therefore hypothesized that the ability of S . aureus to adapt to changes in available oxygen and shifts in substrate availability in inflamed skeletal tissues may rely on one or more TCSs , and that these responses would be key determinants of pathogenesis during osteomyelitis . In this study , we employed TnSeq analysis during acute murine osteomyelitis to determine the genetic requirements for S . aureus survival during invasive infection . A large number of S . aureus genes were identified as essential for growth within bone , some of which have previously been implicated in hypoxic responses . Intravital oxygen monitoring was utilized to define changes in tissue oxygenation during osteomyelitis . Finally , we evaluated the effects of changing oxygenation on S . aureus quorum sensing and virulence factor production . Collectively , these studies determine the staphylococcal genes essential for survival during invasive infection of bone , define shifts in tissue oxygenation during invasive infection , and interrogate the mechanisms by which S . aureus can modulate its virulence in response to changes in oxygen availability .
In order to characterize the genes required for invasive S . aureus infection , TnSeq analysis was performed during experimental osteomyelitis using a recently described S . aureus transposon insertion library in strain HG003 [8] . To identify potential bottlenecks in bacterial survival during osteomyelitis , a timecourse infection was first performed by inoculating murine femurs with strain HG003 . An inoculum of 5x106 CFU was chosen based on direct comparison with strain LAC , which has served as the wildtype strain in prior osteomyelitis analyses and is representative of the most common lineage ( USA300 ) of strains causing osteomyelitis in the United States [18] . At days 1 , 3 , 5 , 7 , and 12 post-infection , infected femurs were harvested and processed for CFU enumeration . After an initial period of replication from day 1 to day 3 post-infection , decreases in bacterial burdens were noted by days 5 and 12 ( S1 Fig ) . Day 5 was therefore chosen for TnSeq analysis of acute osteomyelitis , as it likely represents the first bottleneck encountered by invading bacteria . For TnSeq analysis of osteomyelitis , mice were infected with the TnSeq library by direct inoculation into the femur . Five days post-infection , femurs from infected mice were processed for genomic DNA extraction . One limitation of TnSeq analysis during invasive infection is the requirement for an outgrowth step after the recovery of bacteria from infected tissues . Although in vitro outgrowth could potentially confound fitness calculations , it is necessary to decrease host DNA contamination and allow for efficient sequencing of microbial DNA , and thus has become a standard practice during TnSeq analysis of invasive infection models [5 , 8 , 19–23] . We opted for a short outgrowth in liquid media to minimize any confounding effects on fitness calculation . For an in vitro comparator , an equivalent volume of the osteomyelitis inoculum was grown in vitro for 24 hours prior to collection and genomic DNA extraction . To determine mutants with compromised in vivo fitness , a “dval” was calculated for each gene in each condition ( inoculum , in vitro comparator , or osteomyelitis ) . A total of 65 genes were found to be essential for survival during osteomyelitis ( S1 Table ) but not in vitro growth , and mutations in an additional 148 genes resulted in significant in vivo compromise relative to the in vitro comparator ( S2 Table ) . Of the 213 genes identified by TnSeq , 39 essential and 73 compromised genes encode hypothetical proteins , respectively . Of the remaining 101 genes , 12 essential genes and 32 compromised genes have Kyoto Encyclopedia of Genes and Genomes ( KEGG ) identifiers . Thirty-two of the 44 genes with KEGG identifiers can broadly be categorized into metabolic pathways , with specific pathways represented including carbon metabolism ( 9 genes ) , amino acid biosynthesis ( 7 genes ) , and the TCA cycle ( 5 genes ) . In the TCA cycle , mutations in genes sucB ( SAOUHSC_01416 ) , sucC ( SAOUHSC_01216 ) , and sucD ( SAOUHSC_01218 ) , which encode enzymes responsible for the conversion of α-ketoglutarate to succinate , each resulted in compromised growth during osteomyelitis . Moreover , genes encoding enzymes in pathways that feed into the TCA cycle were also important for intraosseous growth , including pyruvate carboxylase ( SAOUHSC_01064 pyc ) , pyruvate dehydrogenase ( SAOUHSC_01040 pdhA ) , and a putative malic enzyme ( SAOUHSC_01810 ) . Mutations in 7 S . aureus genes encoding amino acid biosynthesis enzymes compromised bacterial growth during osteomyelitis , yet did not significantly impair growth in vitro . These genes encode enzymes in the biosynthetic pathways for tryptophan ( SAOUHSC_01369 trpC , SAOUHSC_01367 trpG , and SAOUHSC_01377 ) , cysteine ( SAOUHSC_00488 cysK ) , lysine ( SAOUHSC_01868 ) , leucine ( SAOUHSC_02288 leuD ) , and the conversion of serine to glycine ( SAOUHSC_02354 glyA ) . Mutations in 6 genes encoding components of purine and pyrimidine metabolic pathways resulted in significant in vivo compromise during osteomyelitis . Two of these genes ( SAOUHSC_02126 purB , SAOUHSC_02360 tdk ) were essential for staphylococcal survival in bone . A substantial portion of the oxidative phosphorylation pathway was also found to be necessary for staphylococcal growth during osteomyelitis . Four of the 12 essential genes with KEGG identifiers and 1 of the mutants with compromised growth are involved in oxidative phosphorylation , including components of quinol oxidase complexes ( SAOUHSC_01000 qoxC , SAOUHSC_01032 cydB ) , and 3 subunits of the F-type ATPase ( SAOUHSC_02340 atpC , SAOUHSC_02343 atpG , SAOUHSC_02346 atpH ) . Collectively , the results of TnSeq analysis during experimental osteomyelitis suggest broad adaptations in metabolism and energy production are required for staphylococcal survival during invasive infection of bone . In contrast to an abundance of genes encoding hypothetical proteins or metabolic pathways , relatively few genes encoding known or putative virulence factors were identified by TnSeq as important for staphylococcal survival in bone . Phosphatidylglycerol lysyltransferase , encoded by mprF ( SAOUHSC_01359 ) , catalyzes the modification of phosphatidylglycerol with L-lysine and contributes to bacterial defenses against neutrophils , cationic antimicrobial peptides , and certain antibiotics [24] . The mprF gene was essential for growth during osteomyelitis , suggesting that resistance to antimicrobial peptides and neutrophils are important components of staphylococcal survival in bone . A second virulence-associated gene identified by TnSeq as essential for S . aureus osteomyelitis was isdF ( SAOUHSC_01087 ) , which encodes a component of the iron-regulated surface determinant heme uptake system [25] . Interestingly , mutation of the ferric uptake regulator ( SAOUHSC_00615 fur ) gene also resulted in compromised intraosseous growth , illustrating a potential role for iron acquisition in the pathogenesis of staphylococcal osteomyelitis . Mutation in the genes encoding thermonuclease ( SAOUHSC_00818 nuc ) , a fibrinogen-binding protein ( SAOUHSC_01110 ) , the repressor of toxins ( SAOUHSC_01879 rot ) , and two serine proteases ( SAOUHSC_01935 splF , SAOUHSC_01938 splD ) also compromised the survival of S . aureus during osteomyelitis . The genes identified by TnSeq as critical for staphylococcal osteomyelitis encode diverse metabolic processes , hypothetical proteins , and select virulence factors . These results suggest that complex bacterial adaptations occur in response to invasive infection of bone . One mechanism by which bacterial pathogens sense and ultimately respond to host-imposed stresses is through TCSs . We therefore hypothesized that staphylococcal TCSs might coordinate the complex adaptations observed during osteomyelitis . Strikingly , TnSeq analysis identified only one S . aureus TCS as required for intraosseous survival . The staphylococcal respiratory response ( SrrAB ) system is involved in coordination of the staphylococcal response to hypoxia and other stresses [26] , and has been shown to directly regulate select virulence factors [27] . Both the histidine kinase ( srrB ) and the response regulator ( srrA ) components of the SrrAB locus were essential for staphylococcal survival in bone , implying that this TCS might be particularly important for coordination of the metabolic and virulence adaptations to intraosseous growth ( S1 Table ) . In total , these results reveal the power of TnSeq analysis to identify S . aureus genes required for invasive infection of bone . Among the mutants that exhibited decreased survival in the osteomyelitis model , we identified a single TCS , SrrAB , which coordinates responses to hypoxia and nitrosative stress in vitro [26] . Moreover , mutations in two additional genes regulated by SrrAB specifically under conditions of nitrosative stress , cydB and qoxC , also resulted in significantly decreased fitness during osteomyelitis ( S1 and S2 Tables ) [26] . Bone and bone marrow are intrinsically hypoxic , leading to the hypothesis that SrrAB contributes to osteomyelitis pathogenesis by sensing and responding to changes in environmental oxygen [11 , 28] . Because the SrrAB regulon was previously defined under conditions of nitrosative stress , we sought to further define the oxygen-dependent SrrAB regulon by performing global transcriptional analysis of a clinically relevant strain ( LAC ) in comparison to a mutant strain lacking srrAB expression ( ΔsrrA ) in both aerobic and hypoxic growth conditions . Inactivation of srrAB under aerobic conditions resulted in the differential regulation of 64 genes ( 39 transcripts increased in abundance and 25 decreased in abundance upon inactivation of srrAB ) ( S3 Table ) . Under hypoxic growth conditions , srrAB inactivation led to differential regulation of 78 genes ( 22 transcripts increased in abundance and 56 decreased in abundance ) ( S4 Table ) . Of the genes differentially regulated by SrrAB under aerobic or hypoxic conditions , only 16 were previously identified as members of the SrrAB regulon under nitrosative stress , suggesting that specific stresses might invoke different SrrAB-dependent transcriptional responses [26] . Moreover , by defining the SrrAB regulon under aerobic versus hypoxic conditions , we discovered that an additional 7 genes important for survival during osteomyelitis in the TnSeq dataset are also SrrAB-regulated ( S3 and S4 Tables ) . The requirement of multiple genes in the SrrAB regulon for survival during osteomyelitis suggests that the SrrAB TCS is an important orchestrator of S . aureus stress responses in inflamed skeletal tissues . Previous reports have demonstrated a significant defect in the growth of an srrAB mutant under anaerobic conditions but not under hypoxic growth conditions [26 , 27 , 29] . To confirm that SrrAB was not found to be essential in the TnSeq analysis simply because of a defect in growth , the ΔsrrA polar transposon mutant and mutations in known genes of the SrrAB regulon ( pflA , pflB , qoxA , and qoxC ) were analyzed in the LAC strain background . The growth rate of each mutant under aerobic or hypoxic conditions was monitored over time . Under aerobic and hypoxic growth conditions , ΔsrrA had an enhanced lag phase compared to WT but reached equivalent optical densities to WT by 8 hours ( S2 Fig ) . The ΔqoxA and ΔqoxC mutants were unable to reach maximal optical densities as previously reported due to disruption of the electron transport chain [26] . These results indicate that an srrAB mutant is not impaired for growth under hypoxia , further validating our TnSeq methods . To investigate the role of SrrAB in osteomyelitis in a clinically-relevant background without the potentially confounding influence of competition from other mutants in the TnSeq library , groups of mice were infected with either WT or ΔsrrA in the LAC background . At 5 or 14 days post-infection , femurs were either processed to quantify bacterial burdens or subjected to micro-computed tomography ( microCT ) imaging ( day 14 ) for quantification of cortical bone destruction . Inactivation of SrrAB resulted in a significant reduction in bacterial burdens in infected femurs at both 5 and 14 days post-infection ( Fig 1A ) . Moreover , murine femurs infected with ΔsrrA sustain significantly less cortical bone destruction than WT-infected femurs ( Fig 1B–1D ) . These results demonstrate that SrrAB is critical for S . aureus survival in infected bone and for induction of pathologic changes in bone remodeling during osteomyelitis . Furthermore , the data suggest that staphylococci encounter hypoxic and/or nitrosative stresses during osteomyelitis . Normal bone and bone marrow are intrinsically hypoxic , with a physiologic oxygen concentration range of 11 . 7 to 48 . 9 mmHg ( 1 . 5–6 . 4% O2 ) , compared to atmospheric oxygen at approximately 160 mmHg ( 21% O2 ) [11 , 28] . TnSeq analysis demonstrated that the hypoxia-responsive SrrAB TCS is essential for S . aureus survival in bone , suggesting that bacterial invasion and the resulting inflammation associated with osteomyelitis trigger further decreases in skeletal oxygen availability . In order to determine the oxygen concentrations of S . aureus infected murine femurs during osteomyelitis , an Oxylite monitor was used to record oxygen tensions at the inoculation site at various times post-infection . In uninfected mice , average pO2 in the intramedullary canal was 45 . 2 mmHg , ( Fig 2 ) consistent with previously reported bone marrow physoxia [28] . As infection progressed , the infectious focus became increasingly hypoxic , with an average oxygen tension of 14 . 3 mmHg at 10 days post-infection . This decreased oxygen tension was not due to the trauma induced by the inoculation procedure , as mock-infected bone showed an elevated mean pO2 of 77 . 5 mmHg by 4 days post-procedure ( Fig 2 ) . Collectively , these findings demonstrate that skeletal tissues become increasingly hypoxic during S . aureus osteomyelitis . Intravital pO2 monitoring revealed that skeletal tissues become increasingly hypoxic during osteomyelitis , with dramatically reduced oxygen tensions as early as 24 hours after infection . These data and the results of TnSeq analysis suggest that the srrAB promoter is active in vivo . To test the hypothesis that srrAB promoter activity increases with decreasing oxygen availability in infected skeletal tissues , a luminescent reporter construct was created in which expression of the luxABCDE operon is driven by the srrAB promoter . Mice were infected with WT S . aureus containing either this construct or a promoterless vector control , and at 1 hour or 24 hours post-infection infected femurs were explanted and immediately imaged for bioluminescence . No detectable luminescence above background was detected in femurs infected with WT bacteria containing the promoterless control plasmid at 1 hour or 24 hours after infection ( Fig 3 ) . In contrast , femurs infected with the PsrrAB-pAmiLux construct showed no detectable luminescence above background at 1 hour post-infection , but displayed strong luminescent signal at 24 hours after infection , corresponding to the induction of hypoxia in infected skeletal tissues ( Fig 3 ) . Collectively , these results demonstrate that the srrAB promoter is activated in vivo during infection of hypoxic skeletal tissues . Intravital oxygen monitoring revealed hypoxia of skeletal tissues upon infection with S . aureus , suggesting that inflammation triggers a reduction in skeletal oxygen concentrations . Neutrophils are a significant source of both oxidative and nitrosative stresses in vivo and contribute to formation of oxygen-limited abscesses in response to staphylococcal infection [30] . To test the hypothesis that SrrAB is necessary to resist hypoxic and/or nitrosative stresses imposed by neutrophils in vivo , mice were either rendered neutropenic with serial anti-Ly6G ( 1A8 ) monoclonal antibody injections or given an isotype control monoclonal antibody and subsequently infected with WT or ΔsrrA [31] . At 14 days post-infection , femurs were processed for enumeration of bacterial burdens . In mice treated with control antibody , a significant virulence defect was again observed in mice infected with the ΔsrrA mutant ( Fig 4 ) . However , in mice administered the anti-Ly6G ( 1A8 ) antibody , a significant increase in bacterial burdens was observed upon infection with ΔsrrA , such that bacterial burdens no longer differed significantly from non-neutropenic mice infected with WT ( Fig 4 ) . These results suggest that intraosseous survival requires SrrAB to resist hypoxic and/or nitrosative stresses produced by neutrophils in response to S . aureus osteomyelitis . The observation that S . aureus infection of murine skeletal tissues leads to dramatically reduced oxygen concentrations prompted further evaluations of how oxygenation impacts the production of staphylococcal virulence factors . We previously demonstrated that secreted toxins regulated by the accessory gene regulator ( agr ) locus are particularly important for the pathogenesis of S . aureus osteomyelitis [32 , 33] . The agr locus ( agrABCD ) encodes a quorum sensing system coupled to a TCS , and is responsible for growth phase-dependent regulation of a number of S . aureus virulence factors [34] . The response regulator of the agr locus , AgrA , directly regulates the production of alpha-type phenol soluble modulins ( PSMs ) , which contribute significantly to the pathology of S . aureus osteomyelitis [32 , 35] . Indeed , alpha-type PSMs were found to be the sole mediators of cytotoxicity in concentrated culture supernatant towards murine and human osteoblasts in vitro [32] . However , a recent report demonstrated that alpha-type PSM expression is directly linked to alpha toxin ( Hla ) expression [36] . To verify that PSMs are the sole mediators of cytotoxicity toward osteoblastic cells in S . aureus concentrated supernatants , strain LACΔpsmα1–4 ( Δpsm ) containing the overexpression vector pOS1-plgt driving hla expression in trans was tested for cytotoxicity towards osteoblastic cells ( S3 Fig ) . While WT supernatant displayed maximum cytotoxicity , Δpsm and Δpsm pOS1-plgt-hla did not show significantly different cytotoxicity from control . Deletion of hla in an erythromycin-resistant LAC background also failed to attenuate cytotoxicity ( S3 Fig ) . Moreover , targeted inactivation of RNAIII in LAC did not decrease cytotoxicity , further supporting the AgrA-regulated alpha-type PSMs as the sole secreted mediators of cytotoxicity toward osteoblastic cells ( S4 Fig ) . To determine the impact of culture oxygenation on S . aureus exotoxin production , concentrated supernatants were prepared from S . aureus grown either aerobically or under limited oxygenation . Incubation of several different mammalian cell lines or primary human osteoblasts with varying amounts of concentrated culture supernatant demonstrated dose-dependent cytotoxicity that significantly increased if the bacteria were cultured under lower oxygenation ( Fig 5 ) . This phenomenon was not strain dependent , as hypoxic growth also increased the cytotoxicity of strains MW2 and Newman towards osteoblastic cells ( S5 Fig ) . These results indicate that S . aureus virulence factor production is modulated in response to environmental oxygen levels . SrrAB regulates select virulence factors under microaerobic conditions in part by directly interacting with the agr P2 and P3 promoters [27 , 29] . This observation , combined with the role of SrrAB in responding to hypoxic stresses led to the hypothesis that SrrAB may regulate quorum sensing and virulence factor production in response to changes in oxygenation . To investigate the impact of SrrAB on PSM-mediated killing of osteoblasts , osteoblastic cells were incubated with varying amounts of culture supernatant from WT or ΔsrrA strains grown either aerobically or under hypoxia . Aerobically grown ΔsrrA supernatants demonstrated dose-dependent killing of murine osteoblasts that was significantly increased compared to aerobically grown WT supernatants , mimicking the effect of hypoxia on cytotoxicity ( Fig 6A ) . The cytotoxicity of aerobically grown ΔsrrA was diminished by expression of the srrAB locus in trans ( S6 Fig ) . These data suggest that SrrAB represses PSM-mediated cytotoxicity under aerobic conditions . Because SrrAB repressed PSM-mediated cytotoxicity under aerobic conditions , we hypothesized that SrrAB impacts quorum sensing in response to oxygenation . To test this hypothesis , the reporter plasmid pDB59 ( agrP3 promoter driving YFP expression ) was introduced into WT and ΔsrrA [37] . Aerobically grown WT cultures demonstrated significantly lower agrP3 activation compared to cultures grown under limited oxygenation ( Fig 6B ) . This decrease was partially SrrAB dependent , as aerobically grown ΔsrrA strains demonstrate a 2-fold higher expression of agrP3 than aerobic WT cultures ( Fig 6B ) . To further confirm these results , quantitative RT-PCR was conducted on aerobically and hypoxically grown cultures of WT and ΔsrrA . Transcription of agrA was increased relative to aerobically grown WT for both ΔsrrAB and hypoxically grown cultures . Hypoxically grown cultures also demonstrated significantly elevated levels of psmα and RNAIII transcripts ( Fig 6C ) . Inactivation of SrrAB resulted in an over 30-fold increase in psmα1–4 transcription and a near 20-fold increase in RNAIII expression under aerobic conditions . Under hypoxic conditions , inactivation of srrAB resulted in a 3000-fold and 160-fold increase in psmα1–4 and RNAIII transcript levels , respectively . Collectively , these data indicate that S . aureus quorum sensing and resultant cytotoxicity towards mammalian cells is modulated in an SrrAB-dependent manner in response to changing oxygen availability , and further define SrrAB as an important regulator of metabolic and virulence adaptations during invasive infection .
TnSeq analysis during experimental osteomyelitis revealed S . aureus genes essential for invasive infection . Among the mutants with reduced in vivo fitness was one TCS , SrrAB , which was previously characterized as a coordinator of hypoxic and nitrosative stress responses [26] . SrrAB was originally identified as a regulator of oxygen-dependent toxic shock syndrome toxin-1 ( TSST-1 ) expression , and was noted to have homology to the global respiratory regulator ResDE in Bacillus subtilis [29 , 38] . Subsequent analyses revealed that SrrA is capable of binding to the agr P2 and P3 promoter regions , and that overexpression of SrrAB reduces virulence in a rabbit endocarditis model [27] . These findings indicate a link between SrrAB and quorum sensing and suggest that oxygenation could impact staphylococcal virulence . Yet the specific signal ( s ) that activate SrrAB , and the mechanism by which this system modulates quorum sensing have yet to be determined . Our data suggest that aerobic growth of S . aureus limits quorum sensing and agr-dependent virulence factor production in a manner that is partially dependent on SrrAB . Conversely , hypoxic growth results in significantly increased cytotoxicity toward mammalian cells . Since equivalent bacterial densities were achieved under conditions of hypoxic and aerobic growth , these results imply that the output of quorum sensing can be functionally uncoupled from bacterial density by changes in culture oxygenation . Such uncoupling could be particularly advantageous for quenching of virulence factor production in environments with higher oxygen availability , such as during colonization of the skin or nares . Since inactivation of SrrAB under aerobic conditions failed to fully restore quorum sensing and cytotoxicity to the levels observed with hypoxic growth , it is likely that this phenomenon is a result of multiple factors . Additional studies are therefore needed to determine the SrrAB-dependent and SrrAB–independent mechanisms by which oxygenation regulates quorum sensing . To this end , it has previously been demonstrated that both the S . aureus autoinducing peptide ( AIP ) and AgrA can be functionally inactivated by oxidation [39 , 40] , suggesting a potential SrrAB-independent mechanism for modulation of quorum sensing by oxygen . Moreover , the S . aureus genome is known to encode other redox-sensitive regulators such as Rex , MgrA , SarA , and AirSR [41–45] . It is therefore possible that environmental oxygen is not a direct regulator of quorum sensing , but rather that a change in the redox status of the bacterial cell or oxidative damage triggers changes in virulence factor production . Nevertheless , our findings suggest that shifts in available oxygen , as well as the inherent differences in physiologic oxygen concentrations in various host tissues , could have a significant impact on staphylococcal virulence . Additionally , these data highlight the importance of in vitro culture conditions on the study of staphylococcal virulence . Global transcriptional analyses defined the SrrAB regulon of S . aureus under conditions of aerobic and hypoxic growth . Interestingly , although some overlap was noted with the previously reported SrrAB nitrosative stress regulon , we identified additional SrrAB-regulated genes under conditions of changing oxygenation [26] . Although these findings may relate to technical issues or strain-dependent differences in gene regulation , they suggest that SrrAB may integrate multiple environmental signals , or that oxidative and nitrosative stress trigger a common endogenous bacterial pathway that activates SrrAB . In order to begin defining the host components that trigger hypoxic or nitrosative stress responses in S . aureus , we examined the role of neutrophils during osteomyelitis . Neutrophils impose nitrosative and oxidative stress to invading pathogens through the respiratory burst , which generates reactive oxygen and nitrogen species . Moreover , neutrophils contribute to tissue hypoxia through abscess formation [30] . In support of the role of SrrAB in resisting such hypoxic and nitrosative stresses , we found that neutrophil depletion partially rescued the virulence defect of an srrA mutant . Additional studies are needed to parse out the effects of neutrophil-derived reactive oxygen and nitrogen species versus abscess-associated tissue hypoxia on the survival of S . aureus . Furthermore , it is likely that other innate and adaptive immune responses contribute to changes in tissue oxygenation and thus the redox status of invading pathogens . In addition to the genes encoding SrrAB and its targets , TnSeq analysis during osteomyelitis revealed a large number of staphylococcal mutants with compromised fitness in vivo . Many of the genes identified as essential or compromised during in vivo growth can be broadly classified as associated with metabolism . In contrast , very few prototypical virulence factors were identified as essential for osteomyelitis . The lack of traditional secreted virulence factors identified through TnSeq analysis is not surprising due to the nature of the technique . Infection with a pooled transposon library allows for mutants deficient in a particular gene to potentially co-opt bacterial factors from other mutants . Indeed , this phenomenon has been characterized for the exchange of metabolic intermediates in S . aureus , and it is conceivable that secreted virulence factors could be “shared” in a similar manner [33] . An additional limitation of TnSeq analysis is the requirement for a short outgrowth step in liquid media following harvest of the transposon library from infected bone . This outgrowth step could potentially confound our results by altering the fitness of mutants recovered from infected tissues . However , a brief outgrowth step is necessary to decrease the amount of murine DNA present in the femur homogenate and allow for effective sequencing of bacterial DNA , and is a common adjustment in TnSeq analyses of infected tissues [5 , 8 , 22] . TnSeq analysis of staphylococcal osteomyelitis paralleled a previous TnSeq analysis of staphylococcal growth in soft tissue abscesses [8] . In fact , 40 of the 65 genes identified as essential for growth during osteomyelitis were also essential for growth in murine abscesses . This observation is consistent with our previous data showing that osteomyelitis is characterized by exuberant abscess formation in the bone marrow [32] , and suggests common stresses are encountered by staphylococci in neutrophil-rich inflammatory lesions . However , 25 of the 65 genes essential for intraosseous survival were not found to be essential for abscess growth , and may reflect unique adaptations to colonization of skeletal tissues . Of the genes required for S . aureus survival during invasive infection , many encode hypothetical proteins or proteins without a previously characterized role in virulence . This observation highlights the power of TnSeq analysis as an unbiased evaluation of the genetic requirements for bacterial survival in host tissues . In summary , the results of this study elucidate bacterial survival strategies during invasive infection , link changes in environmental oxygen to staphylococcal quorum sensing and virulence , and provide a firm foundation to identify new targets for antimicrobial and vaccine design .
All experiments involving animals were reviewed and approved by the Institutional Animal Care and Use Committee of Vanderbilt University and performed according to NIH guidelines , the Animal Welfare Act , and US Federal law . The S . aureus TnSeq library in strain HG003 has been previously described [8] . All other experiments were conducted in an erythromycin-sensitive , tetracycline-sensitive derivative of the USA300 strain LAC ( AH1263 ) , which served as the wildtype ( WT ) unless otherwise noted [46] . Strain LACΔpsmα1–4 has been previously described [32 , 47] . Strains ΔsrrA , ΔqoxA , ΔqoxC , ΔpflA , and ΔpflB in the LAC background were created by bacteriophage phi-85-mediated transduction of erm-disrupted alleles from the respective JE2 strain mutants obtained from the NARSA transposon library [48] . Strains Δpsm pOS1-plgt and Δpsm pOS1-plgt-hla were provided by Dr . Juliane Bubeck-Wardenburg [36] . Construction of strain LAC ΔRNAIII is described below . Plasmid pDB59 ( agr-P3-YFP ) was electroporated into LAC or ΔsrrA for monitoring of agr-dependent quorum sensing [37] . All strains were grown in glass Erlenmeyer flasks at 37°C with orbital shaking at 180 rpm . All S . aureus strains were grown in Tryptic Soy Broth ( TSB ) , Brain-Heart Infusion ( BHI ) , or Roswell Park Memorial Institute medium ( RPMI ) supplemented with 1% casamino acids ( CA ) . Escherichia coli was grown in Luria Broth ( LB ) . Erythromycin and chloramphenicol were added to cultures at 10 μg ml-1 where indicated . Ampicillin was added to cultures at 100 μg ml-1 where indicated . Cadmium chloride was added to cultures at 0 . 1 mM where indicated . A 5:1 flask to volume ratio was utilized unless otherwise noted . For comparative growth analyses , overnight aerobic cultures were back-diluted 1:1000 into fresh TSB or BHI media and optical density at 600 nm ( OD600 ) was monitored over time . RNAIII including upstream and downstream flanking regions were amplified using primers 5'-GCATGCGTCGATATCGTAGCTGGGTCAG-3' and 5'-GAATTCGAAGTCACAAGTACTATAAGCTGCG-3' , and cloned into the HincII site of pUC18 [49] to create pGAW1 . To delete RNAIII , inverse PCR was performed with primers 5'-TTTGGGCCCTATATTAAAACATGCTAAAAG-3' and 5'-TTTCTCGAGGTAATGAAGAAGGGATGAGTT-3' amplifying RNAIII flanking regions and the remaining plasmid backbone of pGAW1 . The vector was religated after treatment with Polynucleotide Kinase ( New England Biolabs , MA ) and designated pGAW3 . To insert an antibiotic resistance cassette , pGAW3 was digested with ApaI and XhoI , religated with the ApaI-XhoI fragment from pJC1075 [50] ( cadCA , conferring resistance to cadmium ) and designated pGAW6 . The SphI- KpnI fragment from pGAW6 was cloned into the allelic replacement vector pJC1202 [50] using the same restriction sites and designated pGAW7 . Strain RN4220 was electroporated with plasmid pGAW7 and plated on GL agar containing 5 μg chloramphenicol ml–1 at 30°C . Allelic exchange was carried out as previously described [50] . Phage 80a was then used to transduce the mutation into LAC to generate LAC RNAIII::cad , herein designated LAC ΔRNAIII . To express srrAB in trans , the srrAB open reading frame was PCR amplified from genomic DNA of LAC using primers 5’-ATCTCGAGATGTCGAACGAAATACTTATCG-3’ and 5’- ATGGATCCTTCAATTTTATTCTGGTTTTGGTAG-3’ . The resulting srrAB amplicon was then cloned into the shuttle vector pOS1 under control of the lgt promoter [51] . As a control , wild type and ΔsrrA strain LAC were transformed with pOS1-lgt lacking an insert . To examine expression of srrAB in vivo the srrAB promoter was PCR amplified from genomic DNA of LAC using primers 5’-TACCCGGGTGTATTTATCACAAAGTTTGAGAAT-3’ and 5’-ATCGTCGACACAGGTCATACCTCCCAC-3’ . The resulting amplicon was then cloned into the shuttle vector pAmiLux , kindly provided by Dr . Julian Davies [52] . As a control , wild type stain LAC was transformed with pAmilux lacking an insert . Osteomyelitis was induced in 7- to 8-week old female C57BL/6J mice as previously reported [32] . An inoculum of 1x106 colony-forming units ( CFU ) in 2 μl PBS was delivered into murine femurs . For some experiments , mice were rendered neutropenic by serial intraperitoneal injections of an anti-Ly6G ( clone 1A8 ) monoclonal antibody ( BioXcell , West Lebanon , NH ) at days -3 , 0 , 4 , 7 , and 10 post-infection . As a control , mice received serial injections of an isotype control antibody ( rat IgG2a ) . At various times post-infection , mice were euthanized and the infected femur was removed and either processed for CFU enumeration or imaged by micro-computed tomography ( microCT ) . For CFU enumeration , femurs were homogenized and plated at limiting dilution on Tryptic Soy Agar ( TSA ) . Analysis of cortical bone destruction was determined by microCT imaging as previously described [32] . Differences in cortical bone destruction and bacterial burdens were analyzed using Student’s t test . Bioluminescent imaging was performed on infected femurs explanted into sterile multiwell plates at either 1 or 24 hours after infection with WT bacteria containing PsrrAB-pAmiLux or pAmiLux . Luminescence was measured in an IVIS 200 Imaging System ( Perkin Elmer , Akron , OH ) with an exposure time of 5 minutes , f-stop of 1 , and binning of 4 . All images were manually scaled to the same minimum and maximum values to exclude background and include the peak luminescent value . Intravital oxygen concentrations were measured in infected femurs using an Oxylite ( Oxford Optronix , United Kingdom ) oxygen and temperature monitor in conjunction with a flexible bare-fibre sensor . Mice were anesthetized with isoflurane and the surgical incision was re-opened . Oxygen readings were obtained by insertion of the sensor directly through the intramedullary canal and into the infectious focus . Measurements from the probe were recorded at least 5 minutes after probe placement to allow for temperature equilibration and stabilization of oxygen readings . The S . aureus TnSeq library in the HG003 background has been previously described [8] . In order to identify potential bottlenecks in the murine osteomyelitis model that could confound TnSeq analysis , groups of mice were first infected with strain HG003 using an inoculum of 5x106 CFU and then at various times post-infection the infected femurs were collected and processed for CFU enumeration . Day 5 was chosen as a timepoint for TnSeq analysis of acute osteomyelitis as it likely represents the first bottleneck encountered by invading bacteria . To prepare the TnSeq library for inoculation into murine femurs , an aliquot of the library containing 5x107 CFU/ml was thawed and inoculated into 100 ml of BHI media in a 500 ml Erlenmeyer flask . This culture was incubated at 37°C for 12 hours and then back-diluted 1:100 into fresh BHI at the same flask to volume ratio and grown an additional 3 hours . Bacterial cells were harvested by centrifugation and resuspended in PBS to a concentration of 7x106 CFU in 2 μl PBS . This inoculum dose failed to cause mortality or severe morbidity requiring euthanasia when administered to five wildtype mice by retro-orbital injection ( mice were monitored for a total of 4 days ) . Genomic DNA was prepared from the inoculum using a Qiagen DNeasy Kit with 40 μg ml-1 lysostaphin added to the lysis buffer . The inoculum was used to initiate experimental osteomyelitis in groups of mice as above . Another equivalent aliquot of the inoculum was seeded into a 50 ml BHI culture in a 250 ml Erlenmeyer flask . This culture was grown for 24 hours , after which time the bacterial cells were harvested and genomic DNA was prepared as above . This genomic DNA served as the in vitro comparator for TnSeq analysis . At 5 days post-infection , mice inoculated with the TnSeq library were euthanized , and the infected femurs were harvested and homogenized in 1 ml of PBS . 500 μl of this homogenate was archived by freezing at -80°C in 20% glycerol and the remaining 500 μl of the homogenate was seeded into 4ml of BHI media and cultured at 37°C and 180 rpm shaking for 5 . 5 hours . Bacteria were then collected by centrifugation and subjected to genomic DNA preparation as above . Recovered bacteria from 2 mice were pooled , and 3 biologically independent groups of mice were analyzed separately . Genomic DNA samples were subsequently prepared for sequencing on an Illumina HiSeq 2000 ( Tufts University Genomic Core Facility ) . Sequencing , data analysis , and fitness calculations were performed as previously reported [8] . Briefly , a “dval” was calculated for each gene in each condition ( inoculum , in vitro comparator , or osteomyelitis ) . The dval represents the observed number of mappable reads of insertions in a gene , divided by the number of mappable reads of insertions predicted for that gene based on its size relative to the genome and the total number of mappable reads obtained for that experiment . Genes with dval of ≤0 . 01 were considered “essential” in a given condition . Genes with dval of >0 . 01 but ≤0 . 1 were considered “compromised” in a given condition , whereas genes with dval >0 . 1 were considered “fit” . A dval ratio was calculated by dividing the dval of a given gene in osteomyelitis by the dval of the same gene during in vitro comparator growth . For genechip analysis , aerobic cultures of WT or ΔsrrA were prepared as follows . Three colonies of WT or ΔsrrA were inoculated into 10 ml of TSB in a 50 ml Erlenmeyer flask . This culture was grown overnight then back-diluted 1:1000 into 50 ml of TSB in a 250 ml flask . The back-diluted cultures were grown at 37°C and 180 rpm orbital shaking until OD600 reached 0 . 5 , at which time an equal volume of ice-cold 1:1 acetone:ethanol was added and the cultures were stored at -80°C until processed for RNA isolation . For comparison of RNA from aerobic versus hypoxic conditions , TSB cultures of WT or ΔsrrA were incubated overnight as above , back-diluted 1:1000 into 100 ml of TSB in a 500 ml flask and grown to an OD600 of 0 . 5 . Fifty milliliters of the culture was then placed into a tightly capped 50ml conical ( hypoxic condition ) and incubated for one hour at 37°C before mixture with acetone:ethanol and storage at -80°C . The remaining 50 ml of culture was moved to a 250 ml Erlenmeyer flask ( aerobic condition ) and incubated for one hour at 37°C before mixture with acetone:ethanol and storage at -80°C . For RNA isolation , bacterial cells were harvested by centrifugation and resuspended in LETS buffer ( 0 . 1 M LiCl , 10 mM EDTA , 10 mM Tris HCl , 1% SDS ) . The resuspended cells were disrupted in the presence of 0 . 5 mm RNAase-free zirconium oxide beads in a Bullet Blender ( Next Advance , Averill Park , NY , USA ) . Disrupted cells were heated at 55°C for 5 . 5 minutes and centrifuged for 10 minutes at 15 , 000 rpm . The upper phase was collected and transferred to a new tube before adding 1 ml of TRI-Reagent . After mixing , 200 μl of chloroform was added , and the resultant solution was mixed vigorously for 15 seconds . Samples were centrifuged at 15 , 000 rpm for 10 min , and the aqueous phase was transferred to a new tube . RNA was precipitated with isopropyl alcohol and subsequently washed with 70% ethanol before drying and resuspension in deionized water . RNA samples were subsequently treated with DNase I and re-purified with a GeneJET RNA Cleanup Kit ( Thermo Fisher Scientific , Waltham , MA , USA ) . For Genechip analysis , RNA samples were labeled , hybridized to commercially available S . aureus Affymetrix Genechips , and processed as per the manufacturer’s instructions ( Affymetrix , Santa Clara , CA , USA ) . Briefly , 10 μg of each RNA sample was reverse transcribed , resulting cDNA was purified using QIAquick PCR Purification Kits ( Qiagen , Germantown , MD , USA ) , fragmented with DNase I ( Ambion , Carlsbad , CA , USA ) , and 3’ biotinylated using Enzo Bioarray Terminal Labeling Kits ( Enzo Life Sciences , Farmingdale , NY , USA ) . A total of 1 . 5 μg of a labeled cDNA sample was hybridized to a S . aureus GeneChip for 16 hr at 45°C , processed , and scanned in an Affymetrix GeneChip 3000 7G scanner as previously described [53 , 54] . Signal intensity values for each GeneChip qualifier were normalized to the average signal of the microarray to reduce sample labeling and technical variability and the signal for the biological replicates were averaged using GeneSpring GX software ( Agilent Technologies , Redwood City , CA , USA ) [54–57] . Differentially expressed transcripts were identified as RNA species that generated a two-fold increase or decrease in WT cells in comparison to ΔsrrA cells during aerobic and hypoxic conditions ( t-test , p = 0 . 05 ) . All related GeneChip data files were deposited in the NCBI Gene Expression Omnibus repository in the MIAME-compliant format . S . aureus strains were used to inoculate RPMI + 1% CA in glass Erlenmeyer flasks . For aerobic growth , the flask opening was covered lightly with aluminum foil . For hypoxic growth , the flask opening was sealed with a rubber stopper . Cultures were grown for 15 hours . Supernatants were collected after culture centrifugation , and were subsequently filtered through a 0 . 22 μm filter and concentrated with an Amicon Ultra 3 kDa nominal molecular weight limit centrifugal filter unit ( Millipore , Billerica , MA , USA ) per the manufacturer’s instructions . Following concentration , supernatants were filter sterilized again and frozen at -80°C until used . Primary human osteoblasts were obtained from Lonza ( Basel , Switzerland ) and cultured per manufacturer’s recommendations . All cell lines were obtained from the American Type Culture Collection ( ATCC ) and propagated at 37°C and 5% CO2 according to ATCC recommendations . Media was replaced every 2–3 days . All cell culture media was prepared with 1X penicillin/streptomycin and filter sterilized using a 0 . 22 μm filter prior to use . MC3T3 E-1 cells were cultured in α-MEM , supplemented with 10% fetal bovine serum ( FBS ) . The RAW264 . 7 , Saos-2 , and A549 cell lines were grown in Dulbecco’s MEM ( DMEM ) with 10% FBS , McCoy’s 5A medium with 15% FBS , and F-12K medium with 10% FBS , respectively . The Jurkat , U937 , and HL-60 cell lines were propagated using RPMI with 10% FBS . Cytotoxicity assays were performed in 96-well tissue culture grade plates . Cells were seeded one day prior to intoxication with S . aureus concentrated supernatants or sterile RPMI diluted in the recommended cell culture medium . The following cell densities were used for cytotoxicity assays: MC3T3 E1 murine pre-osteoblastic cells at 5 , 000 cells per well , primary human osteoblasts at 3 , 500 cells per well , Saos-2 human osteoblastic cells at 10 , 000 cells per well , RAW264 . 7 murine macrophage cells at 10 , 000 cells per well , A549 lung epithelial cells at 5 , 000 cells per well , U937 monocytic cells at 15 , 000 cells per well , HL-60 premyelocytes at 20 , 000 cells per well , and Jurkat T cells at 50 , 000 cells per well . Concentrated supernatants were added as dilutions , by mixing between 0 . 1 μl to 60 μl in a total volume of 200 μl per well to give a dilution spectrum of 0 . 05%-30% concentrated supernatant ( volume/volume ) . Cell lines in suspension were centrifuged at 3000 x g for 5 minutes prior to intoxication . Cell viability was assessed with CellTiter AQueous One ( Promega , Madison , WI , USA ) per the manufacturer’s instructions at 24 hours post-intoxication . For fluorescence analysis , overnight cultures of WT and ΔsrrA containing the pDB59 reporter plasmid were back-diluted 1:1000 into 10 ml of RPMI + 1% CA with chloramphenicol in 50ml Erlenmeyer flasks and grown either aerobically or hypoxically as above . YFP was measured using an excitation of 485/20 and emission of 528/20 in a BioTek Synergy HT 96-well plate reader at 0 , 6 , 9 , 12 , and 15 hours after back-dilution . Bacteria were grown for 15 hours as for YFP fluorescence measurements , mixed with 1:1 acetone:ethanol , and stored at -80°C until processed for RNA isolation . RNA isolation was performed as for Genechip analysis . Reverse transcription using 2 μg of RNA and M-MLV reverse transcriptase ( Promega , Madison , WI , USA ) was performed following the manufacturer’s instructions . Quantitative RT-PCR ( qRT-PCR ) was performed using iQ SYBR Green Supermix ( Bio-Rad , Hercules , Ca , USA ) and the cDNA generated above for each primer pair , including a no reverse transcriptase negative control for 16S rRNA . PCR was conducted on a CFX96 qPCR cycler ( Bio-Rad , Hercules , Ca , USA ) . The cycling program was carried out as recommended by the manufacturer with an annealing temperature of 56°C . Fold-changes were calculated from Ct values averaged from three technical replicates for at least three biological replicates after normalizing to 16S rRNA . The qRT-PCR primer sequences for agrA , hla , and RNAIII were previously published [58] . The qRT-PCR primer sequence for 16S rRNA was also previously published [36] .
|
Staphylococcus aureus is a leading cause of infectious death , yet is also capable of harmlessly colonizing healthy individuals . These disparate observations imply that S . aureus can modulate its growth and virulence in response to different host environments . To characterize the staphylococcal genetic programs required to sustain invasive infection , we applied a technique known as TnSeq to experimental osteomyelitis . Osteomyelitis is one of the most common invasive manifestations of staphylococcal infection , and a better understanding of the bacterial factors required to colonize and destroy bone will aid in vaccine and antimicrobial development . TnSeq identified more than 200 genes important for invasive staphylococcal infection of bone . Two of these genes encode a bacterial two-component system , SrrAB , which is known to help S . aureus survive in low oxygen . Consistent with this finding , we discovered that oxygen levels in bone decrease during osteomyelitis . Furthermore , we discovered that staphylococcal virulence is augmented by environmental oxygen levels , suggesting one strategy by which S . aureus can respond to different host environments . Collectively , our results define the genetic and metabolic programs required for S . aureus to sustain invasive infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Bacterial Hypoxic Responses Revealed as Critical Determinants of the Host-Pathogen Outcome by TnSeq Analysis of Staphylococcus aureus Invasive Infection
|
The speed of theta brain oscillatory activity is thought to play a key role in determining working memory ( WM ) capacity . Individual differences in the length of a theta cycle ( ranging between 4 and 7 Hz ) might determine how many gamma cycles ( >30 Hz ) can be nested into a theta wave . Gamma cycles are thought to represent single memory items; therefore , this interplay could determine individual memory capacity . We directly tested this hypothesis by means of parietal transcranial alternating current stimulation ( tACS ) set at slower ( 4 Hz ) and faster ( 7 Hz ) theta frequencies during a visuospatial WM paradigm . Accordingly , we found that 4-Hz tACS enhanced WM capacity , while 7-Hz tACS reduced WM capacity . Notably , these effects were found only for items presented to the hemifield contralateral to the stimulation site . This provides causal evidence for a frequency-dependent and spatially specific organization of WM storage , supporting the theta–gamma phase coupling theory of WM capacity .
The theta–gamma cross-frequency coupling theory [1] ( Fig 1A ) proposes that individual fast brain waves ( gamma cycles ) represent individual memory items that are bound together to a multi-item memory by slow brain waves ( theta oscillations ) . Consequently , individual differences in the length of a theta cycle ( 4–7 Hz ) might determine how many gamma cycles ( >30 Hz ) can be nested into a theta cycle and may therefore determine memory capacity . This theory provides a potential neurophysiological mechanism for individual differences in the maximum number of items ( number of gamma cycles ) retained in the memory buffer ( one theta cycle ) . According to this theory , it would be expected that slower theta frequencies will integrate a higher number of gamma cycles per theta cycle , resulting in increased memory capacity . Conversely , faster theta frequencies would bind a comparatively smaller number of nested gamma cycles , resulting in a decreased memory capacity . Correlational studies have provided indirect support for this theory ( e . g . , [2 , 3] ) . For instance , Axmacher and colleagues [2] showed that increasing working memory ( WM ) load leads to a slowing down in the theta frequency . Moreover , recent neurostimulation work has shown that entraining parietal theta oscillations via transcranial alternating current stimulation ( tACS ) [4] or rhythmic Transcranial Magnetic Stimulation ( TMS ) [5] has proven effective in enhancing WM performance , providing causal evidence for the role of theta oscillations in WM performance . These works have so far mainly focused on enhancing theta amplitude by enhancing the theta signal-to-noise ratio , leading to better performance . A recent work has attempted to enhance WM capacity through manipulation of the intrinsic theta cycle length by frontal tACS set at a stimulation frequency slower than the individual theta [6] . However , it is unclear whether the enhanced WM performance obtained in that study is due to slowing of theta frequency or can be attributable to the more general impact of stimulation on theta amplitude , per se . Therefore , previous studies have left unanswered a long-lasting question regarding the exact mechanism by which theta oscillations orchestrate WM capacity: does the cycle length of theta oscillations play a mechanistic role in determining interindividual variability of WM capacity ? Here , we test the prediction based on the theta–gamma phase coupling theory [1] that inducing slower theta cycles will enhance WM capacity , while inducing faster theta cycles will reduce WM capacity . We tested this prediction in a visuospatial WM paradigm based on seminal work by Vogel and Machizawa [7 , 8 , 9] , who showed interindividual differences in visuospatial WM capacity to positively correlate with the amplitude of evoked responses localised over parietal areas contralateral to the hemifield where the stimulus to be kept in memory was presented . Crucially , using the same paradigm , Sauseng and colleagues [3] found a clear lateralisation of theta-locked gamma phase synchronization increase over parietal areas , again predicting individual WM capacity . Therefore , based on recent electroencephalography ( EEG ) [10] , magnetoencephalography ( MEG ) [11] , and behavioural evidence [12] that tACS can drive the intrinsic resonance frequency towards an externally imposed rhythm [13] , we directly tested for the modulation of WM capacity by slow ( 4 Hz ) and fast ( 7 Hz ) theta tACS over parietal areas [3 , 7 , 8 , 9] . In line with our predictions , we found that slow theta tACS enhanced WM capacity while fast theta tACS reduced WM capacity . Importantly , these effects were specifically obtained for the visual hemifield contralateral to the stimulation site .
Two groups of 16 participants were each assigned to two different electrode montages . In both montages , an electrode was placed over the same right parietal region in order to stimulate the parietal area of the frontoparietal WM network , known to be relevant in visuospatial WM tasks [14] . This electrode was paired with either a return electrode over the vertex ( Control Montage ) or over the right supraorbital region ( Experimental Montage ) . A few potential issues were anticipated with the use of the Control Montage for our paradigm to be effective: ( a ) the reduced distance between the two electrodes in the Control Montage , which might result in a significant proportion of current being shunted over the skin [15] , rendering the stimulation less effective; ( b ) the spread of current across both hemispheres , due to the vertex electrode sitting centrally and therefore reducing the expected lateralised impact of stimulation on WM performance [3 , 7 , 8 , 9]; ( c ) the different orientation of current flow relative to the neurons’ orientation across the two montages , specifically due to the differential position of the return electrode , which may play a relevant role in the stimulation efficacy [16 , 17]; and ( d ) the Control Montage being less effective on the target brain area intraparietal sulcus ( IPS ) ( see [18] ) . The Experimental Montage used was designed to overcome these potential confounds . Participants in each montage group underwent active stimulation ( at 4 and 7 Hz ) and sham stimulation while performing a visuospatial delayed match to sample task [3 , 7 , 8 , 9] . The visuospatial WM task involved remembering an array of four to six coloured squares that was briefly presented to either the left or right visual hemifield ( i . e . , contralateral or ipsilateral to the stimulated hemisphere , respectively ) for a short period of time and then assessing whether it was the same or different from a subsequently presented array ( see Fig 1B for details on the task and stimuli example ) . WM capacity across different memory loads was measured using a K-value , which is a standardised measure estimating how many items can be stored in WM ( e . g . , [3 , 7 , 8 , 9] ) . In addition , in order to make the results more comparable with other studies not using K as an estimate of memory capacity , the percentage of correct responses ( accuracy ) was also calculated . Finally , in order to reduce variability induced by the control condition sham ( as the between-group factor ) , data were sham-normalised ( for a non-sham-corrected data analysis , see S1 Fig ) . A mixed factorial ANOVA with the between-factor Montage ( Experimental Montage and Control Montage ) and within-factors Condition ( Sham-corrected 4 Hz , Sham-corrected 7 Hz ) × Load ( 4 , 5 , and 6 items ) × Hemifield ( left and right ) was carried out on the K-values ( see Fig 1C and 1D , and Data analysis ) and accuracy ( see S1 Fig ) . Results showed a main effect of Condition ( K: F ( 1 , 30 ) = 5 . 90 , p = 0 . 021 , η2 = 0 . 16; accuracy: F ( 1 , 30 ) = 6 . 39 , p = 0 . 017 , η2 = 0 . 18 ) , suggesting that stimulating at 4 Hz and 7 Hz relative to Sham had a differential impact on WM capacity . Importantly , a Condition × Hemifield × Montage interaction ( K: F ( 2 , 60 ) = 5 . 79 , p = 0 . 022 , η2 = 0 . 16; accuracy: F ( 2 , 60 ) = 5 . 25 , p = 0 . 029; η2 = 0 . 15 ) showed that the two montages modulated performance differently depending on stimulation Condition and Hemifield . Subsequent ANOVAs were therefore performed separately for each montage . In the Experimental Montage , we found a main effect of Condition ( K: F ( 1 , 15 ) = 5 . 70 , p = 0 . 03 , η2 = 0 . 28; accuracy: F ( 1 , 15 ) = 5 . 75 , p = 0 . 029 , η2 = 0 . 28 ) and a significant interaction of Condition × Hemifield ( K: F ( 1 , 15 ) = 9 . 53; p = 0 . 008; η2 = 0 . 39; accuracy: F ( 1 , 15 ) = 7 . 46; p = 0 . 015; η2 = 0 . 33 ) , suggesting that the different stimulation conditions had a differential impact on left and right hemifields . Given the lateralised application of tACS ( right parietal ) and the contralateral parietal activation during visuospatial WM maintenance observed in previous work [3 , 7 , 8 , 9] , a significant modulation of WM capacity was expected for items presented over the left ( contralateral ) but not the right ( ipsilateral ) hemifield . These trials were analysed separately in two further repeated measures ANOVAs ( Condition × Load ) . As expected , the analysis of the left hemifield trials showed a significant main effect of Condition ( K: 4 Hz: 0 . 36 ± 0 . 016; 7 Hz: −0 . 43 ± 0 . 25; F ( 1 , 15 ) = 23 . 97; p = 0 . 0002; η2 = 0 . 61; accuracy: 4 Hz: 3 . 54% ± 1 . 71%; 7 Hz: −3 . 44% ± 1 . 94%; F ( 1 , 15 ) = 45 . 53 , p = 0 . 000007 , η2 = 0 . 75 ) , while no main effect of Load nor interactions reached significance ( K: all p > 0 . 42; accuracy: all p > 0 . 78 ) . Moreover , one-sample t tests against 0 confirmed that 4-Hz tACS significantly enhanced K-values ( and accuracy ) relative to sham ( K: t ( 15 ) = 2 . 28; p = 0 . 019 , one-tailed; Cohen’s d = 0 . 57; accuracy: t ( 15 ) = 2 . 13; p = 0 . 024 , one-tailed; Cohen’s d = 0 . 53 ) , while 7-Hz tACS significantly reduced K-values ( and accuracy ) relative to sham ( t ( 15 ) = −1 . 78; p = 0 . 047 , one-tailed; Cohen’s d = 0 . 44; t ( 15 ) = −1 . 83; accuracy: p = 0 . 046 , one-tailed; Cohen’s d = 0 . 46 ) ( Fig 1C leftmost graphs for mean and individual data and S1 Fig ) . As expected , analysis of the right hemifield trials showed no significant effect of Condition ( K: F ( 1 , 15 ) = 0 . 12; η2 = 0 . 008; p = 0 . 73; accuracy: F ( 1 , 15 ) = 0 . 002; p = 0 . 97; η2 = 0 . 0001 ) , as well as no significant effect of Load or interactions ( all p > 0 . 19 ) ( Fig 1C , rightmost graphs and S1 Fig ) . Finally , the same analysis performed on the Control Montage showed no main effects or interactions reaching significance ( all p > 0 . 23 ) ( Fig 1D , S1A and S1B Fig ) . In order to clarify the contribution of the electrode configuration on the observed effects , we calculated electric field distribution for both montages based on a realistic head model [21] . Results of this analysis suggest that stimulation in the Control Montage ( P4-Cz ) led to more superior parietal stimulation and more left parietal stimulation , relative to our Experimental Montage ( P4-supraorbital ) , in which participants received stronger stimulation exactly at around the right IPS [18] that then spread throughout the right hemisphere but was confined within it ( Fig 2 ) .
The effects we obtained were in line with the expected empirical results . Theoretically , the difference between cycle length of 4 Hz ( 250 ms ) and 7 Hz ( 143 ms ) would be around 107 ms and would therefore allow for at least two additional gamma cycles to be nested into a theta cycle . Based on this evaluation , a difference of 2 items between stimulation conditions could be expected when stimulating at 4 Hz relative to 7 Hz . However , we note that the modulation of the K-values obtained here at 4 Hz and 7 Hz does not perfectly match the impact that would be expected from theory ( see Fig 1A ) . Specifically , we found a total difference across conditions of about 0 . 8 items , which in terms of accuracy corresponds to about a 7% difference between 4 Hz and 7 Hz stimulation . This difference between the potential maximum impact of stimulation and the observed effect can be explained by the fact that tACS delivers very weak currents that might only partially drive the endogenous ongoing oscillatory activity . There is also recent evidence showing how the impact of tACS on oscillatory activity is strictly dependent on the state of the brain during stimulation [22] . This implies that intervening factors can also induce subtle modulation of tACS effects , rendering the stimulation less effective . For example , in our experimental manipulation , tACS was not delivered in phase with the beginning of the 900-ms retention period shown to be more strictly related to theta synchronization [2] . This might account for some trial-by-trial variability in the impact of theta tACS for maximising cross-frequency coupling in the relevant theta phase ( i . e . , stimulation may have been more effective in those trials in which tACS was in phase with the onset of the retention interval than in other trials that were less in phase or even in counterphase with the onset of the retention period ) . Moreover , one would only expect a full 2-item difference if each participant responded perfectly . Yet , individual differences might significantly interact with the way the method optimally works , and therefore one would also expect overall smaller mean values . However , despite such potential intervening factors , tACS still significantly impacted current results in the predicted direction . WM capacity shows variability both between and within participants . Between-participants variability in WM capacity has been documented . For example , it has been shown that WM variance can be explained by the theta:gamma frequency ratio [23] , and even within participants , trial-by-trial variability can be observed [24] , with the same items being sometimes retained and sometimes not . We speculate that this variability might be reflected in slight variations in oscillatory activity related to this function . Indeed , recent reports support the functional relevance of the between and within trial-by-trial variability in frequency speed in different domains , from visual processing [11 , 25] to multisensory binding [12] and pain perception [26] , thus rather discarding the notion of this being sheer noise in oscillatory fluctuations . Although not directly related to our experimental paradigm or specific oscillatory frequency , these findings provide a more general framework in support and physiological backup to the behavioural results we present here , suggesting that tACS can effectively shift individual trait-like behaviour associated with oscillatory activity in desired directions . In the context of WM capacity , the model of Lisman and Idiart [1] very well matches the theoretical framework of inter- and intraindividual oscillatory variability determining WM capacity , and we have indeed tested this in the current study . Therefore , while WM capacity is a trait-like ability , possibly centred around a person’s individual theta frequency peak , this trait-like dimension may be prone to slight but functionally significant fluctuations around the mean , which could be best explained by the trial-by-trial variability in the theta frequency peak , accounting for the trial-by-trial ability to correctly encode information in WM . Individual theta frequency may vary across participants in the full range between 3 and 8 Hz , thus even beyond the range of 4–7 Hz that we have considered . This might in principle explain why not all of our participants showed a consistent effect of enhanced or reduced WM capacity relative to their sham condition . According to our hypothesis , one would expect that 4-Hz tACS would always improve performance relative to sham and 7-Hz tACS would always reduce performance relative to sham . While at first sight , this might seem to be the case , this conclusion would be based on the misleading assumption that , at the individual level , sham stimulation would necessarily sit in the middle as if it corresponded to a 5 . 5-Hz stimulation . However , as per definition , individual theta frequency may vary within an even wider range than the one we have defined here , conventionally and conveniently , but arbitrarily , of 4 and 7 Hz . Indeed , while literature generally refers to theta as an oscillatory activity in a range between 4 and 7 Hz , it has been reported that both slow theta 3-Hz oscillations and fast theta 8-Hz oscillations can be associated with memory performance [27] . Therefore , 4- and 7-Hz stimulations do not necessarily sit on the lower and upper boundaries of the individual theta frequency . In turn , this may explain in principle why some participants do not show positive Sham-corrected values for 4-Hz stimulation with negative differential scores for 7-Hz stimulation at the same time , or being close to 0 in other cases . Crucially , this perspective would also explain why , in all cases , 4-Hz relative to 7-Hz stimulation always resulted in a better WM performance . At an additional level of analysis , the shape of the individual theta peaks may also vary quite substantially from sharp to broad across individuals , which might in turn determine the net effect size of our data . Indeed , such characteristics may well interact in the way participants respond to the tACS interventions , such that participants with broadband theta might be more susceptible to tACS interference than those with sharp peaks . While we cannot provide a more detailed and conclusive demonstration of this mechanism here , these are all relevant points that future research needs to address . However , the current study already provides a fundamental step forward in the understanding of the mechanisms underlying spatial WM processing , showing the behavioural impact of tACS modulation of WM capacity closely following a theoretical model of WM capacity on the one hand and interventional impact of tACS on the other hand . It might be argued that a way to alternatively demonstrate the impact of tACS on WM capacity would instead be to modulate gamma frequency . So , in principle , stimulating at faster gamma frequency might enhance WM capacity , while stimulating at slower gamma frequency would instead reduce WM capacity . However , the theta–gamma coupling framework of WM capacity does not assume any changes in gamma frequency and , indeed , there is clear evidence against this . For instance , Axmacher and colleagues [2] showed that increasing WM load leads to a reduction in theta frequency , whereas gamma frequency is not significantly slowed down . Also , the theoretical framework assumes that single memory item representations would be reflected by activity in local gamma networks oscillating at exactly this gamma frequency . The single representations themselves would not change , therefore gamma frequency should not either . In support of this notion , there is evidence that locally generated gamma would not even change frequency if the network size were increased [28] . Based on the electric field modelling reconstruction , the Experimental Montage shows its maxima over IPS , exactly underneath the stimulation electrode , as one would expect . The control condition shows instead a maximum more anterior , off the stimulation electrode , with a clearly less lateralised distribution of the electric field , which together could explain why , within our paradigm , there was no significant modulation of WM capacity in the Control Montage . When looking at the Experimental Montage electric field distribution , this is clearly lateralised with maxima over the IPS and spread more widely through the right hemisphere , including frontal areas . According to this picture , it could be argued that the actual significant impact of the Experimental Montage might be due ( i ) to a more effective tACS of the right prefrontal cortex via the supraorbital electrode of the Experimental Montage or ( ii ) to a more effective activation of the frontoparietal network instead and not the P4 stimulation site per se , common to both montages , or even ( iii ) subcortical activations . Although we cannot completely rule out these alternative hypotheses , we argue that these are very unlikely explanations of current results . First , the right supraorbital electrode is not the optimal position for modulating frontal theta oscillatory activity ( see previous tACS work targeting the frontoparietal network , with the frontal sensor sitting more posteriorly , e . g . , over FCz or laterally over F3 and F4 [6 , 29 , 30] ) . Moreover , tACS montages testing the relative impact of frontal and parietal areas have shown a selective modulation of WM by parietal but not frontal stimulation [29 , 30] . Therefore , our Experimental Montage has likely not desynchronised frontal and parietal areas that were actively involved in the WM task but instead optimised a lateralised stimulation of one of the crucial nodes ( right parietal area in this case ) leading to current lateralised effects . Furthermore , if any of the effects observed could be ascribed to activation of frontal or even subcortical activations , it is unlikely to carry lateralised effects , which should instead be driven more specifically by parietal activations . Importantly , the choice of the parietal area was strongly inspired by relevant empirical work [3 , 7 , 8 , 9] . These studies showed that the spatial component is a relevant one in our experimental design that actively calls into play parietal rather than frontal activations . Indeed , both theta oscillations [3] and event related potential ( ERP ) components associated with spatial WM capacity [7 , 8 , 9] , in the very same paradigm we have used here , are systematically localised contralateral to the items to be remembered and are crucially over posterior areas , the very same we have stimulated in our study . Instead , no such modulation of theta oscillations [3] or ERP [7 , 8 , 9] over prefrontal areas has been found whatsoever . Indeed , we are considering here a specific visuospatial cued memory component that strongly relies on the visuospatial components , for which parietal areas are primarily involved . It is very unlikely that such lateralised effects observed here and replicated numerous times ( see , e . g . , [31 , 32] ) in different visuospatial WM experimental paradigms ( see , e . g . , [33 , 34] ) may be driven by frontal activations . Therefore , based on this evidence , we explicitly expected the effects to be mainly driven by parietal stimulation and to be lateralised . One could argue that the effect we observed could be the result of retinal activation [16 , 35] due to supraorbital electrode stimulation . Several arguments , however , discount this alternative hypothesis . First of all , none of the participants saw any phosphenes during the experiment . Indeed , reports of retinal phosphene for stimulation frequency within the theta band are very rare ( see [36] ) . Those few participants who saw phosphenes at the beginning ( n = 3 ) were stimulated during the experiment at a tACS intensity not inducing any phosphene sensation . Yet , if one has to consider the potential impact of some residual retinal phosphenes perception over the WM capacity using the current paradigm , a different pattern of results would be expected than the one we currently observe . Specifically , we would expect any effect to be essentially ipsilateral to the stimulation site rather than contralateral , as observed here , instead . While we do observe a lateralised effect of stimulation on WM capacity , no main effect of hemifield could be detected , but only an interaction between hemifield and the specific frequency of stimulation . Such effects were observed for the contralateral rather than ipsilateral hemifield to the stimulated hemisphere and retina and are thus unlikely to be explained by retinal activation . Finally , if phosphenes were not perceived , it might still be argued that a subthreshold impact of tACS on retinal activity might be induced , which may in turn indirectly induce cortical entrainment in corresponding visual areas . If this were to be the case , we would still argue that the activation of the ipsilateral retina would project onto both hemispheres , thus inducing a bilateral entrainment , which is not compatible with the lateralised effect we observed in the current study . Therefore , we have good reasons to believe that any of the effects we observed were genuinely cerebral in nature . To conclude , we found that , depending on electrode configuration , 4-Hz tACS enhanced visuospatial WM capacity , while 7-Hz tACS reduced visuospatial WM capacity compared to sham stimulation . As a result of the hemifield specificity effect , each participant served as their own internal control , depending on the hemifield being tested relative to the hemisphere being stimulated , with the effects being found for items presented to the hemifield contralateral to the stimulation site only . These results are also supported by recent reports in monkeys [37] performing a similar change detection task and showing an independence of the two hemispheres for visual WM function . The findings of this study are in line with the theta–gamma phase coupling theory of WM capacity [1] . While direct electrophysiological evidence supporting our conclusions is still lacking due to the technical challenge of combining online tACS and EEG ( see [6 , 10] ) , here we provide relevant behavioural evidence that slow theta tACS enhances visuospatial WM capacity and fast theta tACS reduces visuospatial WM capacity , in line with the idea that slow and fast theta tACS might induce slower and faster theta oscillations , respectively ( see [10 , 11 , 12 , 13] ) . These findings may provide the basis for potential therapeutic interventions aimed at enhancing poor memory capacity in the ageing population and to ameliorate memory-related neuropathologies in clinical settings .
The study has been approved by the University of Essex Ethics Committee ( VR1301 ) and conducted according to the principles expressed in the Declaration of Helsinki . Written informed consent has been obtained from the participants before taking part in the study . Based on power analysis ( conducted on [6] , actual effect size , f = 0 . 47 ) , with a conservative estimated effect size f = 0 . 25 , alpha = 0 . 05 , and 80% power , a total sample size of 28 participants ( 14 per group ) is suggested . Therefore , in a between-subjects design , we assigned 16 participants ( 7 female ) to the Experimental Montage and 16 ( 9 female ) to the Control Montage . All participants were adults ( 18 years or older ) with mean ages of 28 . 3 years ( ±7 . 6 ) and 22 . 8 years ( ±5 . 2 ) , respectively . Participants completed a safety screening questionnaire before taking part in each session of the study . A repeated measures design was used for stimulation conditions and memory load . There were two active stimulation conditions ( 4- and 7-Hz tACS ) and one control condition ( sham ) . These stimulation conditions took place during three separate sessions that occurred on separate days , and the order was counterbalanced across participants . At least 24 hours passed between different sessions . A single-blind design was used , with participants unaware that different stimulation protocols were being used for each session ( and unaware that one session consisted of sham stimulation ) . A variation of the visual delayed match to sample task based on Vogel and Machizawa [8] was used . Two arrays of coloured squares were situated on either side of a white fixation cross in the centre of a black screen . The number of squares in each array ( the memory load ) was 4 , 5 , or 6 . The left and right arrays were always different from each other in any given trial; however , the number of squares in each trial was always the same for the left and right arrays .
|
Our ability to temporarily retain sensory information is limited to a handful of items and is referred to as working memory capacity . Such memory capacity has been shown to vary across the general population , with some people retaining a higher number of items than others . An influential theory suggests that this individual capacity might be determined by the speed of slow brain waves ( so-called theta waves ) that range in frequency between four and seven cycles per second . It is hypothesized that these theta waves act as glue for items to be memorised such that the slower the theta waves , the higher the number of items that can be clustered and retained in memory . We tested this hypothesis by applying to human participants noninvasive current stimulation at slower or faster theta frequencies over a part of the brain that is involved in visuospatial working memory during a visuospatial task . In line with this influential theory , we found that stimulation at slower theta frequencies enhanced working memory capacity relative to stimulation at faster theta frequencies , which instead reduced working memory capacity . These effects were limited to visual items processed by the stimulated brain areas , confirming the importance of theta waves for the organization of visuospatial working memory .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"electricity",
"brain",
"electrophysiology",
"social",
"sciences",
"electrophysiology",
"neuroscience",
"learning",
"and",
"memory",
"surgical",
"and",
"invasive",
"medical",
"procedures",
"number",
"theory",
"cognitive",
"neuroscience",
"transcranial",
"alternating",
"current",
"stimulation",
"mathematics",
"functional",
"electrical",
"stimulation",
"cognition",
"brain",
"mapping",
"memory",
"electric",
"field",
"bioassays",
"and",
"physiological",
"analysis",
"vision",
"research",
"and",
"analysis",
"methods",
"chemistry",
"transcranial",
"stimulation",
"electrophysiological",
"techniques",
"short",
"reports",
"electrode",
"potentials",
"physics",
"working",
"memory",
"psychology",
"electrochemistry",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"sensory",
"perception",
"cognitive",
"science",
"neurophysiology"
] |
2018
|
The speed of parietal theta frequency drives visuospatial working memory capacity
|
While it is known that musculotendon units adapt to their load environments , there is only a limited understanding of tendon adaptation in vivo . Here we develop a computational model of tendon remodeling based on the premise that mechanical damage and tenocyte-mediated tendon damage and repair processes modify the distribution of its collagen fiber lengths . We explain how these processes enable the tendon to geometrically adapt to its load conditions . Based on known biological processes , mechanical and strain-dependent proteolytic fiber damage are incorporated into our tendon model . Using a stochastic model of fiber repair , it is assumed that mechanically damaged fibers are repaired longer , whereas proteolytically damaged fibers are repaired shorter , relative to their pre-damage length . To study adaptation of tendon properties to applied load , our model musculotendon unit is a simplified three-component Hill-type model of the human Achilles-soleus unit . Our model results demonstrate that the geometric equilibrium state of the Achilles tendon can coincide with minimization of the total metabolic cost of muscle activation . The proposed tendon model independently predicts rates of collagen fiber turnover that are in general agreement with in vivo experimental measurements . While the computational model here only represents a first step in a new approach to understanding the complex process of tendon remodeling in vivo , given these findings , it appears likely that the proposed framework may itself provide a useful theoretical foundation for developing valuable qualitative and quantitative insights into tendon physiology and pathology .
Tendons are dense fibrous tissues that transfer tensile forces from muscles to bones . During normal daily activity , human Achilles tendon experiences high intensity cyclic loads , up to 4–8 times the body weight [1–3] . Achilles tendon stores potential strain energy as it is stretched , which is then recovered later in the gait cycle [4 , 5] . This strain energy cycling reduces muscular work and improves the economy of locomotion . Furthermore , the uncoupling of tendon and muscle lengths , due to the elastic deformation of Achilles tendon , enable the muscle fibers to operate at more favorable lengths and velocities , thus improving locomotion economy even further [6] . It is apparent that there are variable tendon geometries and properties , and that tendon tissue has the capacity to adapt to its mechanical environment [7–10] , but how ? Here we develop a biologically plausible computational model of tendon adaptation . The basis of our model is a collagen fiber damage and repair models , which in turn are based on known biological processes . When an Achilles tendon model is incorporated into a simplified model of the Achilles-soleus unit and allowed to adapt over time with usage , we observe the capacity of model tendon to remodel towards a stable equilibrium tendon geometry , which can coincide with minimum metabolic cost of the model musculotendon unit operation . We begin by introducing the key biological processes incorporated into the model . Tendon extracellular matrix ( ECM ) is primarily composed of Type I collagen ( up to 86% dry mass ) . Collagen fibers , run mainly along the axial length of the Achilles tendon , arranged in a hierarchical structure [11 , 12] . At the smallest scale , tropocollagen molecules self-assemble into microscopically visible strands of collagen fibrils [13] . Despite the small diameter of collagen fibrils , typically 100–150 nm in human adult Achilles tendon [11 , 14 , 15] , their total lengths are believed to be much longer , and potentially extend continuously from muscle fibers to bone [16] . A bundle of collagen fibrils form primary collagen fibers [11 , 17] that are then hierarchically aggregated to form primary , secondary and tertiary fascicles across the whole tendon [11 , 18–20] . Tendon ECM is maintained by resident cells known as tenocytes , which mediate the synthesis and degradation of the ECM components [21] . There is compelling evidence for continuous tendon remodeling by the tenocytes [22–24] . Normally , primary collagen fibers are enclosed by a confluent sheet of tenocytes [25] . This implies that collagen fibril adaptation must normally occur via processes acting at a distance from the tenocyte itself . The tenocytes synthesize proteases and new collagen molecules that then self-assemble to repair fibrillar damage . We envisage that these physiologic processes are consistent with homeostatic and adaptive processes within tendon [26] , and it is these processes that are the focus of our proposed tendon adaptation model . Typically , the stress-strain response of a whole tendon is composed of four regions: ( i ) extension without significant force up to the tendon’s slack length , ( ii ) a toe region , ( iii ) a linear elastic region , and ( iv ) a failure region [27] . The slack length is defined as the tendon length at which the tendon first experiences load [28] . Most of the tendon crimp is removed during extension up to the slack length [29] . The toe region corresponds with further sequential straightening of crimped collagen fibers [29] . Experimental observations report a unique crimp pattern to each individual collagen fascicle [29–31] indicating that each fascicle's stretched length is also unique . Based on the premise of distribution of collagen fiber lengths , it is possible to reproduce tendon’s non-linear force-extension behavior [32–34] , Fig 1 . Cyclic loading of Achilles tendon during habitual activities , such as walking or running , damages the tendon and initiates repair processes to maintain tendon homeostasis . It is observed that cyclic loading of tendons gradually induces ‘micro-damage‘ or ‘sub-failure injuries’ at the collagen molecular and fibrillar levels , which are collectively referred to as ‘mechanical fatigue damage’ [35–41] . The fatigue response of human Achilles tendon to cyclic mechanical loading is clearly evidenced in the experimental studies of Wren et al . [38] . At the microscale , mechanical fatigue damage to collagen fibrils may present itself in ‘focal’ or ‘generalized’ modes . Generalized damage is evidenced by repeating patterns of kinks and distortions along a number of fibrils [42] , whereas focal damage is evidenced by fracture of collagen fibrils [43] . One mode of damage may dominate the other depending on the prevalence and type of cross-links within collagen fibrils [44] . More cross-linking between tropocollagen molecules results in stiffer tendons , such as Achilles and patellar tendons , and so favors the focal fatigue damage mode [44] . For simplicity , in this paper we have focused on the focal mode of damage only . Nevertheless , the damage model employed here may be modified to include other modes of collagen fiber damage as required . In addition to mechanical damage , tenocyte-mediated proteolytic collagen degradation also occurs , and is an essential component of tendon homeostatic processes [14] as it facilitates tissue remodeling [23 , 42] . In normal tendon , proteolytic damage is usually meditated primarily by members of the matrix metalloproteinases ( MMPs ) family of proteases [45 , 46] . Intriguingly , mechanical tensile strain of collagen fibrils has been shown to reduce , and even completely prevent , proteolytic damage of collagen fibrils by collagenase MMPs at physiologically relevant strains [14 , 22 , 47 , 48] . From the above it is clear that while the basic mechanical and physiological aspects of tendon tissue and adaptation have been studied , how these processes are integrated to produce physiologically relevant outcomes during habitual loading is not completely understood . Here we hypothesize that through utilization of the abovementioned biological processes , Achilles tendon is capable of adaptation by remodeling its geometry , to reach a dynamic stable equilibrium that represents tendon homeostasis . In the following , we demonstrate that for a musculotendon unit with constant muscle fiber length , our proposed tendon model is able to remodel its geometry to reach a stable homeostatic equilibrium , and that this equilibrium state can coincide with minimizing the total metabolic cost of the model musculotendon unit .
In the following we first develop and then test a discretized model of Achilles tendon adaptation . The model tendon is based upon damage and repair processes that take place at the level of primary collagen fibers . The mechanical load experienced by the tendon is based upon a simplified Achilles-soleus model and experimental measurements made during normal human gait mechanics . Over time , repeated cycles of damage and repair of the collagen fibers gradually remodel the whole Achilles tendon . We note here that the timescale for the simulated sequential damage and repair processes are not critical to our proposed adaptation model , but for convenience and definiteness we have assumed a daily cycle of damage and repair . This timescale is likely to accord with at least some of the important cyclic homeostatic processes taking place within tendon [49 , 50] . To frame our model of tendon adaptation in accordance with the known abovementioned tendon physiological process , the following computational sub-models are developed: ( i ) a discretized fiber model of tendon mechanics , Fig 2 ( a ) , from which the force-extension behavior of the whole tendon can be estimated , Fig 2 ( b ) . ( ii ) a simplified model of the musculotendon ( Achilles-soleus ) unit , Fig 2 ( c ) , from which the metabolic cost of the ( effective ) soleus muscle mechanical work at the ankle joint is estimated , Fig 2 ( d ) . ( iii ) a model for calculating the intensity of daily load as a function of metabolic cost , Fig 2 ( e ) , ( iv ) models for mechanical damage and repair to individual collagen fibers , Fig 2 ( f ) , and ( v ) models for proteolytic damage and repair to individual collagen fibers , Fig 2 ( g ) . A single sequential pass through these sub-models , results in a remodeled tendon geometry , with a modified force-extension response , Fig 2 ( h ) . In other words , the model tendon adapts by repeatedly cycling through this algorithm . We now describe each of these sub-models in detail . In principle , either the collagen fibrils or the primary collagen fibers may be the anatomical units regarded as the load carrying ‘string elements’ in our model . However to render the computations more tractable , here we have chosen to represent the primary collagen fibers as the discretized string elements . For the rest of this paper , collagen fibers in our model refer to primary collagen fibers . Human Achilles tendon cross-sectional area ( CSA ) at the mid-section is reported to range from under 50 mm2 [51 , 52] to values higher than 80 mm2 [38 , 53] . Assuming an Achilles tendon to have a CSA of 60 mm2 and assuming a circular cross-section for collagen fibers with an average fiber diameter of 28 μm [54] , the total number of primary collagen fibers ( Ntotal ) in human Achilles tendon is estimated to be: Ntotal=ATAF≈ 100 , 000 ( fibers ) ( 1 ) where AT and AF denote whole tendon cross-sectional area and average collagen fiber cross-sectional area In the present study , we explicitly model 100 , 000 collagen fibers that are of non-uniform length , but otherwise similar . We note here that it may be appropriate to have fibers with variable stiffness along their length , as it has been reported that the ‘free’ Achilles tendon is more compliant than the aponeurosis [21 , 55 , 56] . Nevertheless , for simplicity we adopt a uniform stiffness along the fiber length . And for simplicity we assume a Gaussian ( normal ) distribution for the initial profile of fiber lengths , and show that it can approximate the non-linear stress-strain curve of a tendon , Fig 1 . Starting mean fiber length is chosen to be in the range 250 to 280 mm to correspond with reported total anatomic length of human Achilles tendon , including the aponeurosis [57–60] . Choosing LT as the mean length of the tendon and Li the slack length of the ith fiber , then the linear extension of the ith fiber ( ΔLi ) as the whole tendon undergoes linear extension ΔLT is calculated from: ΔLi= { Li− ( LT+ΔLT ) if LT+ΔLT>Li0if LT+ΔLT≤Li ( 2 ) For convenience , we take mean of the slack lengths of all fibers to represent the mean tendon length ( LT≈L¯i ) . While the measured estimates vary , the commonly reported Young’s modulus of the Achilles tendon ( ET ) is about 1 GPa [61 , 62] . Assuming homogenous material properties , and uniform cross-sectional area ( AF ) for all collagen fibers , the stiffness of the ith fiber ( ki ) is then estimated by: ki=AF⋅ETLi ( 3 ) Using Eqs ( 2 ) and ( 3 ) , fiber force ( Fi ) and whole tendon force ( FT ) at a given tendon extension can then be calculated by: Fi=ki⋅ΔLi ( 4 ) FT=Σi=1NtotalFi ( 5 ) Repeated cyclic loading of tendon during daily activity damages collagen fibers [38 , 63] . Our estimates for the likelihood of mechanical fatigue damage of primary collagen fibers is based on the empirical fatigue damage data for the whole human Achilles tendon obtained by Wren et al [38] . The implicit assumption we employ to use this data is that the whole Achilles tendon fatigue behavior is also representative of the individual primary collagen fibers fatigue behavior making up the human Achilles tendon . The assumption that each part of the Achilles tendon is similar to all others is likely to be a reasonable assumption for healthy tendon , but we note this is less likely to be a reasonable assumption for diseased tendon . The average age of the 25 human subjects from whom the Achilles tendon samples were obtained by Wren et al ( 2003 ) was 75 ( ±12 ) years [38] . Therefore , in order to represent the in vivo damage in young adults , we chose to rescale the ultimate tensile stress value of 70 MPa reported by Wren et al ( 2003 ) [38] to 100 MPa [64–66] , while leaving the slope of the fatigue curve unchanged . Fig 3 shows the normalized fatigue curve for collagen fibers employed in our model . This rescaled human Achilles tendon fatigue curve is probably more representative of younger adults , though other scalings may be deemed appropriate depending on data and the intended purpose of the model . However provided that reasonable values are chosen , the actual values for scaling are not critical , and do not substantially change the findings reported here ( see later sensitivity analysis ) . From Fig 3 , the number of loading cycles to tendon failure ( nfail ) at a given peak fiber stress ( σmax ) are calculated using: σmax=a−b⋅log10 ( nfail ) ( 6 ) ∴nfail=10a−σmaxb ( 7 ) where a corresponds to the ultimate tensile stress at one cycle , in this case 100 ( MPa ) , and b is the slope of the logarithmic fatigue curve in Fig 3 , in this case 8 . 25 ( MPa/log ( n ) ) . Peak tensile stress of the ith fiber ( σmaxi ) in a tendon undergo ) ing linear extension ΔLT is calculated by: σmaxi=ET⋅ΔLiLi ( 8 ) where fiber extension ΔLi is calculated from Eq 2 . It is clear that typical daily activities lead to peak stress levels that rarely ( if ever ) result in complete failure of a normal tendon . Consequently we need to devise a ‘cumulative damage function’ to estimate the amount of damage arising from daily activity . For our ‘string’ tendon model , cumulative tendon damage is assumed to be directly proportional to the fraction of broken fibers . The fraction of broken fibers as a result of daily activity can be estimated from a failure ( or reliability ) function for individual collagen fibers . However due to the lack of experimental data on failure functions ( Pfailmech ) or reliability functions ( R=1−Pfailmech ) for tendon , we employ a commonly adopted ‘exponential failure function’ [67] to describe focal damage failure of individual collagen fibers within the Achilles tendon . Therefore the probability of mechanical failure of an individual fiber experiencing peak fiber stress σmax and n load cycles is estimated by: Pfailmech=−κ+κ⋅eλ nnfail ( 9 ) The fitting constants κ and λ in ( Eq 9 ) are defined such that Pfailmech=0 at = 0 , Pfailmech=0 . 1 at n = nfail/2 and Pfailmech=1 at n = nfail . These fitting constants are chosen based on reported typical cyclic fatigue test on human Achilles tendon reported in Fig 2 ( b ) and 2 ( c ) of Wren et al ( 2003 ) [38] . Fitting Eq 9 to this figure suggests reasonable parameter values are , κ = 0 . 0125 and λ = 4 . 395 . A typical cumulative damage probability curve is shown in Fig 3 . However clearly these fitting constants can be adjusted to fit experimental results as required , while the influence of these parameters on our model outputs are quantitated in a later sensitivity analysis . In our model for an Achilles tendon with normal physiology , if a fiber mechanically fails , it is always repaired ( which may not happen in a diseased tendon ) . A repaired fiber may ( probabilistically ) be repaired either shorter or longer , however , we bias the repair of mechanically damaged fibers towards lengthening ( Fig 4 ) . A probabilistic interpretation of fiber repair as used in our model is depicted in Fig 4 ( b ) , which shows the probability distribution of relative length changes to a fiber following its repair . We suggest that this repaired length change , depicted Fig 4 ( a ) , is consistent with the following conceptual model of the repair process following mechanical damage . First , the two ends of the broken fiber are enzymatically debrided by proteases to obtain a suitable undamaged surface from which a new portion of collagen fiber can be constructed . A new portion of collagen fiber is then created by polymerization of tropocollagen molecules [68 , 69] . While the section of fiber debrided may be longer than the newly formed portion of collagen fiber , leading to further fiber shortening , on average the broken ends are more likely to lie somewhat apart , so the new portion of fiber bridges both this gap and any fiber debridement , and so the repaired mechanically damaged fiber is on average longer . The gap between the broken ends arises at least partly because mechanically damaged fibers are on average shorter than remaining nearby fibers , but it seems plausible that the gap between broken fiber ends may also be partly promoted by other events , such as the elastic recoil of the fractured ends of a failed fiber or subsequent cyclic friction forces between fibers . The gap between the broken ends is observable in SEM images of damaged collagenous matrices [43 , 70 , 71] . A schematic depiction of this is shown in Fig 4 ( a ) . It has been noted previously by Provenzano et al [43] , that if the gaps between the fractured ends are filled by newly polymerized collagen fiber , then the repaired fibers lengths are increased . Proteases remove damaged or unwanted ECM as part of normal tissue turnover and collagen fiber homeostasis [14 , 23 , 41] , but the rate of collagen degradation is modified by collagen strain . For example , Wyatt el al . [72] reports an almost complete cessation in collagenase degradation rate when rat tail fascicles were strained 4–5% [72] . However , Flynn et al . performed similar tests on single collagen fibrils , thereby avoiding rotational deformations of collagen fibers during extension ( rotational deformations are often observed in collagen networks made up of fibrils with a variety of fibril orientations ) [73] . For the experimental tests reported by Flynn et al ( 2013 ) , which probably most closely approximate the ( linear ) fibril structure observed in Achilles tendon , collagenase degradation of fibrils is prevented at strains larger than about 1 . 5% . These test results suggest that in terms of our ‘string’ model of tendon , for a given tendon strain , relatively long collagen fibers are less stretched along their length , which render them more susceptible to being degraded by active proteases [22 , 41 , 48 , 74–76] . Based on the experimental results of Flynn et al [47] , we employ an exponentially decreasing probability of fiber cleavage with increasing strain ( Fig 5 ) , viz: Pfailproteo=e−ϕ⋅εmax ( 10 ) where εmax is the fiber peak strain during a gait cycle and ϕ is a fitting constant . To accord with the observations of Flynn et al . [73] , ϕ is calculated to be 300 . This selection of ϕ results in almost no proteolytic damage in fibers experiencing peak strains εmax ≥ 1 . 5% . Clearly the sensitivity to proteolytic damage can be altered by varying the constant . It is equally clear that Eq 10 is a very crude modeling representation of an actual proteolytic process occurring within tendon , as binding of MMPs to collagen fibers , the movement of MMPs along collagen fibers , the surface state of the collagen fibers , and the history of cyclic strain experienced by the collagen fibers , are all time dependent , and so the level of protection afforded to collagen fibers must also be time dependent . A more sophisticated proteolytic damage model would include such time dependencies , but unfortunately to date there is little data available in the literature to suggest more precise functional relationships . When proteolytic degradation of the collagen fiber is complete , once again new collagen molecules polymerize to create newly formed collagen fiber , which bridges the gap between the degraded ends ( see schematic Fig 4 ( c ) ) . For each fiber that is proteolytically degraded , the fiber repair model determines a probabilistic repair length . As in the case of mechanical damage , the repair length is found by sampling a triangular probability distribution . However , this time the triangular probability distribution has a greater tendency to shorten the original fiber , see Fig 4 ( d ) . In other words , the fiber section removed proteolytically is on average longer than that filled by newly formed fiber section , and so the repaired fiber is shortened . We note that the mechanism for fiber shortening is not known with certainty . However , it is possible that passive mechanical forces may contribute to fiber shortening ( e . g . cyclic frictional forces , or compressive residual stresses in long fibers may relax ) , or active forces generated by cells may contribute to fiber shortening ( e . g . a number of experiments have reported that tenocyte generated contractile forces can actively shorten collagen fibers [41 , 74 , 75 , 77] . To provide an appropriate context to test adaptation of Achilles tendon to its loading environment , we set our new tendon model within a standard three-component Hill-type model . Representing the musculotendon unit , the Hill-type model is composed of a contractile element and two elastic elements , Fig 2 ( c ) . The contractile element and the parallel elastic element simulate the integrative behavior of the human soleus through gait cycles [78 , 79] . The soleus is the muscle of choice for examining our tendon model for the following reasons: ( i ) among the plantarflexor muscles it is the largest muscle , and from several modeling studies it has become clear that the soleus is the primary muscle responsible for producing ankle power and work in both walking and running [80–82] . Furthermore , modeling results has found the soleus to be among the most important producers of mechanical work during walking and running across all lower limb muscles [83 , 84] . ( ii ) The soleus only crosses the ankle , unlike the gastrocnemius muscles that cross both the ankle and knee joints ( knee flexor ) . This simplifies the modeling of muscle force using our hill-type model , eliminating potentially complicating factors . Joint torque sharing between the soleus and the other synergist muscles is simplified by initially attributing the torque produced by the soleus to the relative physiological cross sectional area of the soleus and the other ankle plantarflexors combined . Soleus force is subsequently computed from a joint angle-specific soleus moment arm [59] . The activation required to produce this force is modelled incorporating muscle force-length-velocity constraints . Muscle fiber lengths and velocities are influenced in our calculation by ankle joint angle , muscle pennation angle ( we assumed a constant volume muscle model ) as well as tendon stretch . The series elastic element in the musculotendon unit shown in Fig 2 ( c ) represents the Achilles tendon , with its mechanical properties obtained from the discretized tendon model described above . As we allow tendon remodeling over time , mechanical properties of this series elastic element , representing the Achilles tendon , also changes over time . For simplicity the contractile element and its parallel elastic element are taken to have constant properties in all our simulations , though in reality these may also adapt over time [85 , 86] . This simplification is to focus our attention on the adaptation process of the Achilles tendon alone , and to exclude muscle adaptation ( which has been investigated elsewhere [85 , 86] ) . Allowing the muscle to adapt simultaneously with the tendon could potentially complicate our understanding of tendon adaptation , and possibly obscure tendon responses that are of interest here . But clearly muscle does adapt too , and inclusion of such mechanisms is an obvious extension for developing a more realistic future model . Fig 6 illustrates the algorithm for updating the tendon properties in the musculotendon model as a result of activity . The musculotendon model uses an inverse dynamics approach with ankle torque and kinematics as inputs to calculate the required muscle force . The ankle torque and kinematics were experimentally obtained from motion tracking and force measurements from an adult subject during walking [87] . The algorithmic steps in Fig 6 include: ankle torque , Fig 6 ( a ) , and musculoskeletal dimensions ( moment arms ) and ankle angles , Fig 6 ( b ) are employed to first calculate tendon force . The tendon model response to force is then used to calculate tendon extension , Fig 6 ( c ) . Tendon extension and ankle angles , Fig 6 ( b ) and 6 ( c ) , are then used to calculate muscle fiber pennation angle , muscle fiber operating length and muscle force ( FM ) , Fig 6 ( d ) , which balances the calculated tendon force . Muscle activation ( Mact ) to generate the required muscle force is estimated based on muscle length , Fig 6 ( e ) [79] , and muscle velocity , Fig 6 ( f ) [79 , 88] . The expression used to calculate muscle activation , based on formulations by Buchanan et al . [79] , is: Mact=FMFL⋅FV⋅FM , max ( 11 ) Mact ( muscle activation ) is a dimensionless number between zero and one that represents the fraction of maximal muscle activation . FM is the muscle force , Fig 6 ( d ) , while FL is the muscle force at current length as a fraction of the maximum isometric force ( FM , max ) , Fig 6 ( e ) , and FV is the muscle force at the current contraction velocity as a fraction of the maximum isometric force ( FM , max ) , Fig 6 ( f ) . The muscle activation Mact from Eq ( 10 ) is then used to calculate the metabolic cost rate ( power ) , and integration of this quantity with respect to time gives the total metabolic cost during a gait cycle , as described in [78] . The metabolic cost rate ( Q˙ ) during a gait cycle , Fig 6 ( g ) , is expressed as the sum of four terms: activation heat rate ( h˙act ) , maintenance heat rate ( h˙m ) , shortening/lengthening heat rate ( h˙sl ) and the mechanical work rate performed by the muscle ( w˙M ) [78] , viz: Q˙=h˙act+h˙m+h˙sl+w˙M ( 12 ) Now that we have defined our musculotendon unit model , we need to recognize that the musculotendon unit does not operate in isolation from the whole organism , but is in fact part of the whole organism . This relationship between the musculotendon unit and whole organism puts constraints on the muscultotendon units operation , which help guide the tendon to an in vivo equilibrium state . Potential factors at the whole organism level affecting the musculotendon operation in the adult include sensory feedback signaling ( including pain ) , neural muscle activation patterning , higher order cognitive inputs ( e . g . willpower ) and oxygen and metabolic energy availability from the organism to make sure there is matching of supply-demand functions over the whole musculotendon unit . The equilibrium state achieved by tendon operating within the musculotendon unit , operating within the whole organism , depends on these interactions/constraints , which may be formulated mathematically as optimization of a multi-objective function . Clearly this multi-objective function can vary over time with changes in environmental and sensory inputs , nutritional status and determination of the individual , and it is these changes that usually drive tendon adaptation in vivo . But for our modeling purposes , how can we simply and reasonably take into account this substantial in vivo complexity ? At equilibrium it is likely that for everyday repetitive activities such as walking , an important contributor to the multi-objective function is musculotendon unit economy . For habitual repetitive activities such as walking , energy minimization is regarded by some as a key optimization criterion dictating locomotor behavior [89 , 90] . There is much evidence pointing to movement patterns that minimize energy expenditure , from the selection of preferred walking speeds in humans and other species [91 , 92] to preferred stride frequencies [93 , 94] and preferred gaits [95 , 96] . Consequently , it is likely that Achilles-soleus unit economy is close to being maximized when the tendon geometry has reached its geometrical equilibrium state , and that Achilles-soleus unit economy will ‘fall’ on either side of the tendon’s geometrical equilibrium state . Making these assumptions , the simplest way for us to approximate the change in walking economy with tendon geometry in our model is to reflect the effect of current metabolic cost on the musculotendon unit load intensity . This can be implemented most simply by scaling the ankle torque while keeping both gait pattern and number of load cycles per day constant . Taking this approach effectively acts as a constraint on the musculotendon unit operation , preventing the total metabolic cost from becoming physiologically unrealistic at some tendon geometries . Therefore ankle torque ( τa ) can be calculated via: τa=β⋅τa , m ( 13 ) where τa , m is the lab-measured ankle torque and β is an activity scale factor calculated by: β=QminQ ( 14 ) where Q refers to the total metabolic cost of muscle activation and Qmin refers to the minimum total metabolic cost of muscle activation . Fig 6 ( h ) , shows the relationship between the activity scale factor ( 0 < β ≤ 1 ) and the metabolic cost for a range of tendon lengths and constant fiber length dispersion . The calculated activity scale factor β is used in the next cycle to determine subsequent ankle torques , from which flows tendon forces , muscle forces and metabolic cost , which are calculated via cycling through the algorithm for muscultotendon unit operation depicted in Fig 6 . Clearly a more realistic model would take into account changes in the multi-objective function governing musculotendon unit interactions/constraints with the whole body in a much more sophisticated way , and also involve changes in both gait patterns and number of load cycles . Relaxing the assumptions made here represent an interesting direction for future research . We chose a single cycle of remodeling , shown in Fig 2 , to represent a 24-hour period . To correspond the tendon activity level with this time-frame , we subjected all our tendon models to a total of n = 5 , 000 load cycles per day , which approximates the number of gait cycles of active adults [97–99] . To reduce possible time discretization errors , the tendon model is subjected to loading cycles in three equally spaced blocks of simulated activity cycles during a day . At the end of an activity block , the tendon’s fatigue damage is assessed as outlined above , resulting in loss of some intact fibers . With an updated metabolic cost , tendon peak force for the next loading block is updated wherein tendon continues undergoing mechanical loading . At the end of the third loading block , proteolytic damage and finally repair of the mechanically and proteolytically damaged fibers take place , as previously explained , resulting in a new fiber length distribution ( and restored fiber number ) . This daily cycle is repeated to simulate tendon remodeling over weeks or months , and adaptation of tendon properties can be tracked over time . A summary of parameter symbols and values used in the model are shown in Table 1 below .
We first consider fiber length distribution of the model tendon , as fiber dispersion strongly influences its mechanical properties . A model tendon with reduced fiber length dispersion ( i . e . reduced standard deviation ) experiences more rapid fiber recruitment and force development with strain , while a tendon with a larger fiber length dispersion ( i . e . larger standard deviation ) exhibits slower fiber recruitment and force development with strain ( see Fig 7 ) . Strain energy is the potential energy stored by elastic materials as they undergo deformation . For elastic materials such as tendon , it is quantified as the dot product of force and displacement . The elastic strain energy is equal to the area under the force-extension curves shown in Fig 7 . For a given tendon extension of two otherwise identical tendons , the tendon with smaller fiber dispersion has a higher force and more strain energy is stored . However for constant force , the tendon with the lower stiffness will store more strain energy [7] . The effects of damage and repair models on tendon fiber length distributions are next demonstrated in Figs 8 and 9 . All tendon simulation results shown in these two figures are for cyclic loading n = 5 , 000 cycles/day and a peak stress of 55 MPa ( i . e . 5 . 5% strain ) . The model tendons initially have a normal distribution of fiber lengths , with mean fiber length 275mm and fiber length standard deviation of 2mm . Fig 8 ( a ) represents remodeling of tendon by mechanical damage only ( i . e . there is no repair operating ) . The shortest fibers , where fiber strain is the highest , are damaged rapidly and so break first ( they are not repaired , as no repair is operating ) . The remaining shortest fiber are now more abundant ( as the probability density function of fiber lengths progressively increases up to the mean ) , which slows the rate of advance of the broken fiber . Note that without including subsequent repair , the total number of intact fibers gradually decreases over time . If mechanical fatigue damage continues for a long time without repair , eventually all the fibers would fail , and the tendon would rupture [38] . In Fig 8 ( b ) both mechanical damage and repair are operating , so the total number of fibers remains constant . However , repaired lengths are on average longer than the original fiber slack lengths , so the mean fiber length increases , and the whole fiber population ‘marches’ towards longer tendon lengths ( see Fig 8 ( b ) ) . We note that standard deviation of fiber lengths initially changes , but becomes relatively constant over time , even as the population of fibers marches towards longer tendon lengths . Fig 9 ( a ) represents remodeling of a tendon by only proteolytic damage ( i . e . there is no repair operating ) . The longest fibers , where strain is the low , undergo rapid proteolytic damage . The remaining shorter fibers of the population experience progressively higher strains , which helps preserve them . As fibers are removed proteolytically there are fewer fibers and so they experience higher the average strain , which slows their removal . In Fig 9 ( b ) , both proteolytic damage and repair processes operate . The longest fibers are first proteolytically damaged , but when repaired , on average they are shorter . Progressive damage of the longer fibers in the population combined with repair with fiber shortening results in the entire fiber distribution ‘marching’ to the left . Results shown in Figs 8 and 9 are obtained using only the tendon model . Hereafter , activity levels are calculated by incorporating the tendon model within the Hill-type musculotendon model , as described above . The calculated total metabolic cost for a single gait cycle for a range of tendon slack lengths and fiber length standard deviations are shown in the color map plot of Fig 10 . To illustrate tendon remodeling behavior with the musculotendon unit , we chose four arbitrary initial tendon geometries ( i . e . each tendon is given a different initial tendon length and fiber dispersion ) . All tendons are then allowed to remodel for a period of 720 days , subjected to 5 , 000 loading cycles every day , and all other model parameters are held constant . Fig 10 shows remodeling paths for each of the four tendons ( see paths A-D in Fig 10 ) . Because the tendons are initially in disequilibrium states for the musculotendon unit conditions , they remodel towards their normal tendon length , which is an equilibrium state . Importantly , we note that with appropriate selection of damage and repair parameters ( as detailed in the model development described above ) , for each of the initially different tendon geometries illustrated in Fig 10 , their remodeling paths all converge towards final equilibrium states within a region of tendon geometries . Within the assumptions of the computational model as described above , this clearly illustrates that the chosen parameters in the proposed remodeling processes are capable of directing tendon adaptation so that each tendon approaches an equilibrium geometrical state , and that this state can coincide with a region of minimum metabolic cost per gait cycle . During the remodeling from initial states A to D shown in Fig 10 , collagen fibrils are mechanically and proteolytically degraded and repaired as they remodel over time . Fig 11 shows collagen fiber degradation and synthesis turnover times during the remodeling as each tendon moves along its remodeling path ( i . e . along each of the paths A to D shown in Fig 10 ) . Collagen fiber degradation turnover time refers to the ratio of total initial fibrillar collagen content to the rate of collagen fiber removal , as a result of both mechanical and proteolytic damage processes . Collagen fiber synthesis turnover time refers to the ratio of total initial fibrillar collagen content to the rate of new fibrillar collagen formation , as a result of repair processes . At equilibrium , when there is no apparent change in tendon length or fiber dispersion , the degradation and synthesis turnover times are equal . Fig 12 explores the tendon geometries after 720 days of remodeling for each of 55 equally spaced initial tendon geometries placed over the whole tendon geometry domain of tendon length and fiber dispersion while all other model parameters are held constant . Based on the tendon geometries after 720 days remodeling , these 55 initial tendon geometries can be categorized into two distinct groups . We observe that some initial geometries remodel to an equilibrium state ( dark shaded points ) , coinciding with minimum metabolic cost , while the remaining initial geometries remodel towards non-physiological states ( light shaded points ) , where tendons are short ( and growing shorter ) and metabolic costs are very high ( and growing higher ) . Finally , we report the sensitivity of equilibrium tendon length and dispersion to variations in model parameters . For our purpose , we define normalized sensitivity to be: SYP= 1/Ye∂Y1/Pe∂P ( 15 ) where Y is the length of the tendon ( or tendon fiber dispersion ) , and P is a model parameter . Model sensitivities are estimated by incrementing model parameters by 5% . Normalized sensitivities for fourteen model parameters are shown in Table 2 .
The tendon adaptation model presented here primarily revolves around the interplay between mechanical and proteolytic damage and repair processes operating in physiologically normal Achilles tendon . It is now well established that mechanical fatigue can damage collagen fibrils , that MMPs are present in tendon and they proteolytically degrade collagen fibrils , that mechanical strain reduces and can even prevents proteolytic degradation of collagen fibrils , and that collagen fibrils can be repaired in vivo . But what is not yet known is how these individual processes are functionally integrated in vivo to facilitate tendon homeostasis and adaptation [100 , 101] . The key achievement in the present work is that we have demonstrated how these processes can be logically combined to facilitate tendon length adaptation in a robust fashion . In the content of our model assumptions , and for suitable parameter selection ( which based on a sensitivity analysis appear to be robust ) , we show that the tendon can autonomously remodel until it reaches a stable , equilibrium length state . Assuming that the multi-objective function representing external influences on the musculotendon unit is dominated by musculotendon unit economy , we find the perturbations of tendon geometry result in remodeling towards a stable tendon geometry , which coincides with a region minimizing the metabolic cost of muscle activity . As with the initial development of any model , we have invoked many important model simplifications in the interests of building a parsimonious model to highlight fundamental theoretical concepts . But it is clear that upon relaxing these model simplifications , there is great scope for further subtleties of tendon adaptation to emerge . Building upon the foundation established here , it seems likely that more sophisticated and complex extensions of the model will reveal important new interactions that may help explain experimental observations or suggest new experiments on tendon biology .
|
It is now widely acknowledged that tendon plays a vital role in locomotion , while experiments have revealed that tendon is much more metabolically active than previously believed . There have been increasing numbers of papers describing the responses of tenocytes to mechanical loading and speculation about the origins of tendinopathy , but to date there is currently no basic theoretical framework describing how tendon maintains tissue homeostasis consistent with the experimental data , or indeed how tendon adapts to its environmental load conditions . Based on established biological principles of tendon damage and repair , for the first time we develop a dynamic model of tendon homeostasis that is capable of adaptation . We show that for a model soleus musculotendon unit with muscle fiber length kept constant , our model tendon is ‘capable’ of dynamically adjusting itself to find a stable equilibrium tendon geometry , which coincides with minimum metabolic cost of muscle activation . This new theoretical framework for tendon homeostasis and adaptation offers the possibility of refocusing research in basic and clinical science .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"material",
"fatigue",
"medicine",
"and",
"health",
"sciences",
"legs",
"classical",
"mechanics",
"limbs",
"(anatomy)",
"biomechanics",
"collagens",
"physiological",
"processes",
"homeostasis",
"materials",
"science",
"damage",
"mechanics",
"materials",
"physics",
"musculoskeletal",
"mechanics",
"muscle",
"physiology",
"musculoskeletal",
"system",
"proteins",
"connective",
"tissue",
"biological",
"tissue",
"torque",
"physics",
"biochemistry",
"anatomy",
"tendons",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"ankles",
"motion"
] |
2016
|
Adaptive Remodeling of Achilles Tendon: A Multi-scale Computational Model
|
Rickettsia typhi is an intracellular bacterium that causes endemic typhus , a febrile disease that can be fatal due to complications including pneumonia , hepatitis and meningoencephalitis , the latter being a regular outcome in T and B cell-deficient C57BL/6 RAG1-/- mice upon Rickettsia typhi infection . Here , we show that CD4+ TH1 cells that are generated in C57BL/6 mice upon R . typhi infection are as protective as cytotoxic CD8+ T cells . CD4+- as well as CD8+-deficient C57BL/6 survived the infection without showing symptoms of disease at any point in time . Moreover , adoptively transferred CD8+ and CD4+ immune T cells entered the CNS of C57BL/6 RAG1-/- mice with advanced infection and both eradicated the bacteria . However , immune CD4+ T cells protected only approximately 60% of the animals from death . They induced the expression of iNOS in infiltrating macrophages as well as in resident microglia in the CNS which can contribute to bacterial killing but also accelerate pathology . In vitro immune CD4+ T cells inhibited bacterial growth in infected macrophages which was in part mediated by the release of IFNγ . Collectively , our data demonstrate that CD4+ T cells are as protective as CD8+ T cells against R . typhi , provided that CD4+ TH1 effector cells are present in time to support bactericidal activity of phagocytes via the release of IFNγ and other factors . With regard to vaccination against TG Rickettsiae , our findings suggest that the induction of CD4+ TH1 effector cells is sufficient for protection .
Rickettsiae ( R . ) are small obligate intracellular bacteria that are transmitted to humans by arthropod vectors . R . prowazekii and R . typhi represent the two members of the typhus group ( TG ) of Rickettsiae [1 , 2] and are the causative agents of epidemic and endemic typhus , respectively . Both diseases appear with similar symptoms including high fever , headache , myalgia and joint pain , nausea and vomiting . Furthermore , neurological symptoms such as confusion and stupor are common [3] . Many patients develop a characteristic rash which is due to local blood vessel damage and inflammation as endothelial cells belong to the main target cells of these bacteria [4] . Fatal complications include pneumonia , myocarditis , nephritis and encephalitis/meningitis [3 , 5] and are more common in epidemic typhus ( 20–30% lethality ) [5–7] . The course of disease of endemic typhus caused by R . typhi is usually milder and the lethality is estimated to be <5% [7 , 8] if untreated with antibiotics such as tetracyclins or chloramphenicol . As clinical presentations are often non-specific , endemic typhus , however , is clearly underdiagnosed and , thus , often unrecognized [3 , 9] . Epidemic and endemic typhus generally occur worldwide . Epidemic typhus that is transmitted from human-to-human by the body louse sporadically appears in low-income countries of South America and Africa but also in upper-middle economies such as Peru [10] and Algeria [11] and industrial countries such as Russia [12] . The most recent larger outbreak of epidemic typhus was in the context of civil war in Burundi in 1995 [13] . Endemic typhus is much more prevalent and actually one of the most abundant rickettsial infections [14] . Rats and mice serve as natural reservoirs of R . typhi and the bacteria are transmitted to humans by fleas , predominantly the rat flea Xenopsylla cheopis , from these animals [14 , 15] . Endemic areas are mainly found in low-income countries in Africa [16–22] and Asia [23–26] where the disease primarily occurs in warm coastal areas and ports where populations of rats and mice are numerous and hygienic standards are low . Within the past twenty years the incidence of endemic typhus has been increasing also in western countries . Rising case numbers are reported from the US ( Hawaii [27] and Texas [28 , 29] ) as well as from Southern Europe ( e . g . Cyprus [30 , 31] , Greece [32] , Spain [33–36] including Canary Islands [37 , 38] ) and Portugal [39 , 40] . Especially homeless people are at enhanced risk to acquire the infection . For example , 9 . 6% of the homeless were seropositive in Houston , Texas , in 2008 [41] . Another study indicates increasing occurrence of R . typhi in France . While 0 . 54% of the homeless in Marseille were seropositive in the years 2000–2003 , seropositivity increased to 22% in the years 2010–2013 [42] . A vaccine against rickettsial infections is not available but clearly desired for several reasons . It is known that some rickettsial species persist and can re-appear . This is true for R . prowazekii , the closest relative of R . typhi . R . prowazekii can cause the so-called Brill-Zinsser disease years to decades after primary infection which appears with similar symptoms as the primary infection and is usually accompanied by meningitis and neurological symptoms [43–46] . Stress or waning immunity is suggested to re-activate R . prowazekii [47] . Similar may be true for R . typhi because we recently showed that R . typhi persists in mice [48] . Moreover , in mice it has been shown that R . prowazekii persists irrespective of antibiotic treatment [49] . In addition , there is the risk of the development of antibiotic resistances . Finally , TG Rickettsiae are considered potential bioweapons . Vaccine development requires understanding of protective immune responses as well as of a possible contribution of immune reactions to pathology . To date little is known about immune response against Rickettsiae although animal models of rickettsial infections have been established . Current studies mainly focused on immunity against spotted fever group ( SFG ) Rickettsiae that represent the vast majority of Rickettsiae but phylogenetically differ from TG Rickettsiae . Especially immunity against R . conorii and R . rickettsii of this group has been studied in mice . Generally , BALB/c and C57BL/6 mice have been reported to be resistant against rickettsial infections [50–54] while C3H/HeN mice were found to be susceptible [50 , 54] . In C3H/HeN mice it has been shown that CD8+ T cells that can directly kill infected cells play an important role in defense against Rickettsiae . CD8+ T cells showed enhanced cytotoxic activity in C3H/HeN mice upon infection with R . conorii and R . australis [55] . Furthermore , depletion of CD8+ T cells led to enhanced susceptibility of C3H/HeN mice to R . conorii and R . australis [55 , 56] as well as to R . typhi [57] as reflected by enhanced bacterial burden and pathology while adoptive transfer of CD8+ immune T cells protected C3H/HeN mice against a lethal dose of R . conorii [56] . This was also true for adoptively transferred CD4+ T cells [56] . An important effector molecule produced by T cells as well as by NK cells is IFNγ . This cytokine has been shown to contribute to rickettsial control . C57BL/6 IFNγ-/- mice succumbed to the infection with a normally sublethal dose of R . conorii . Furthermore , neutralization of IFNγ led to severe disease in C3H/HeN mice upon R . conorii infection [58] . Similar was also true for the neutralization of TNFα [58] which can be produced by various cell types including CD4+ T cells and macrophages ( MΦ ) [59 , 60] . Nevertheless , C57BL/6 RAG1-/- mice that lack adaptive immunity survive the infection with R . conorii at least for 20 days [61] which is also true for R . typhi [48] , suggesting that innate immune mechanisms can control the bacteria at least for a certain period of time . In C57BL/6 RAG1-/- mice , however , R . typhi , re-appears several months after infection and then grows predominantly in the brain [48] . Animals develop massive CNS inflammation accompanied by neuronal cell loss and succumb to neurological disorders [48] . These findings clearly demonstrate the need of adaptive immunity for the control of persisting R . typhi . In the present study we show that CD4+ T cells are sufficient to protect against R . typhi in the C57BL/6 RAG1-/- infection model by activating MΦ via IFNγ and other factors .
All experimentations and procedures were approved by the Public Health Authorities ( Amt für Gesundheit und Verbraucherschutz , Hamburg; No 61/12 and No 88/13 ) and performed according to the German Animal Welfare Act . C57BL/6 , C57BL/6 RAG1-/- [62] , C57BL/6 MHCI-/- lacking CD8+ cytolytic cells [63 , 64] and C57BL/6 MHCII-/- lacking CD4+ T cells [65 , 66] mice were bred in the animal facilities of the Bernhard Nocht Institute for Tropical Medicine and housed in a biosafety level 3 facility for experimentation . The facilities are registered by the Public Health Authorities ( Amt für Gesundheit und Verbraucherschutz , Hamburg ) . R . typhi ( strain Wilmington , accession no . AE017197 ) was cultured in L929 mouse fibroblasts ( ATCC CCL-1 ) in RPMI1640 ( PAA , Cölbe , Germany ) supplemented with 10% FCS ( PAA , Cölbe , Germany ) , 2 mM L-glutamine ( PAA , Cölbe , Germany ) and 10 mM HEPES ( PAA , Cölbe , Germany ) without antibiotics ( standard culture medium ) . 1×107 γ-irradiated ( 1966 rad ) L929 cells were seeded in 175 cm2 culture flasks ( Greiner Bio-One , Frickenhausen , Germany ) . One day later cells were infected with R . typhi and incubated for 5 to 7 days . For the preparation of bacterial stocks , infected L929 cells were resuspended in 1 . 5 ml PBS . 200 μl silicium particles ( 60/90 grit silicon carbide; Lortone inc . , Mukilteo , USA ) were added and cells were vortexed thoroughly for 1 min . The crude lysate was strained through a 2 μm cell strainer ( Puradisc 25 syringe filter 2 μm; GE Healthcare Life Sciences , Freiburg , Germany ) . Bacteria were centrifuged at 4300×g for 5 min at room temperature and frozen in FCS with 7 . 5% DMSO in liquid nitrogen in Cryo . S tubes ( Greiner Bio-One , Frickenhausen , Germany ) . Thawed bacterial stocks were centrifuged at 6200×g for 5 min at room temperature , washed twice with PBS and analyzed for bacterial content by quantitative real-time PCR ( qPCR ) and immunofocus assay as described previously [48] to determine spot forming units ( sfu ) . R . typhi stocks were thawed and washed in PBS as described above . 2×106 sfu were administered in 50 μl PBS subcutaneously ( s . c . ) into the base of the tail . Blood samples were obtained by submandibular bleeding . The state of health was evaluated by a clinical score . The following parameters were assessed: body mechanics/motion ( 0: normal; 1: tremor/swaying motion; 2: paresis/ataxia; 3: paralysis ) and weight loss ( 0: normal ( <10% ) ; 1: mild ( >10% ) , 2: severe ( >15% ) ) giving a maximum score of 5 . Mice were considered healthy with a score ≤1 , moderately ill with a score of 2–3 and severely ill with a score of 4–5 . Mice were sacrificed reaching a total score of ≥4 or showing weight loss of >20% . DNA was prepared from purified bacteria , cell cultures and organs employing the QIAamp DNA Mini Kit ( QIAGEN , Hilden , Germany ) . 10 mg tissue were homogenized in 500 μl PBS in Precellys ceramic Kit tubes ( Peqlab , Erlangen , Germany ) in a Precellys 24 homogenizer ( Peqlab , Erlangen , Germany; two times 6000 rpm for 45 sec with a 60 sec break ) . 80 μl tissue homogenizate or up to 1×106 cells were used for DNA preparation according to the manufacturer´s instructions . Quantification of purified R . typhi , bacteria in cell cultures and organs from infected mice was performed by amplification of a 137 bp fragment of the PrsA gene as previously described [48] . Cytokines were detected in cell culture supernatants with LegendPLEX assay ( Biolegend , London , UK ) according to the manufacturer´s instructions . Cell culture supernatants were diluted 1:2–1:10 . 50 μl Griess 1 reagent ( 0 . 5 g sulfonamide in 50 ml 1M HCl ) and 50 μl Griess 2 reagent ( 0 . 15 g naphtylethylendiamine-dihydrochloride in 50 ml H2O ) were added to 100 μl cell culture supernatant in microtiter plates ( Greiner Bio-One , Frickenhausen , Germany ) . A serial dilution of sodium nitrite ( NaNO2 ) in cell culture medium was used as a standard ( cmax 125 μM ) . The absorbance was measured at 560 nm with a Dynex MRXII spectrophotometric microplate reader ( Dynex Technologies , Chantilly , USA ) . Brain from naïve and infected mice were homogenized . In addition , blood samples were analyzed . Erythrocyte lysis was performed for blood by incubating the cells in erythrocyte lysis buffer ( 10 mM Tris , 144 mM NH4Cl , pH7 . 5 ) for 5 min at RT . Cells were afterwards washed two times in PBS . Brain and spinal cord cells were strained through a 30 μm CellTrics cell strainer ( Partec , Görlitz , Germany ) and directly used for stainings . Cells were fixed and permeabilized with Cytofix/Cytoperm and Perm/Wash solutions ( BD Biosciences , Heidelberg , Germany ) according to the manufacturer´s protocol . Fc receptors were blocked with 5% CohnII human IgG fraction ( Sigma-Aldrich , Deisenhofen , Germany ) in Perm/Wash for 15 min at 4°C followed by the addition of either mouse anti-R . typhi ( BNI52 ) [48] or mouse IgG3 isotype antibody ( clone B10; SouthernBiotech , Birmingham , USA ) , each at a concentration of 1 μg/ml in Perm/Wash . Cells were washed in Perm/Wash after 20 min of incubation at 4°C followed by incubation for 20 min at 4°C with rat anti-mouse IgG3-FITC ( 1:200 in Perm/Wash; #1100–02 , SouthernBiotech , Birmingham , USA ) . After washing cells were further stained with rat anti-mouse iNOS-PE ( clone CXNFT; eBioscience , Frankfurt , Germany ) and rat anti-mouse CD11b-PerCPCy5 . 5 . ( clone M1/70; BD Biosciences , Heidelberg , Germany ) . To discriminate immune cells in brain , cells were additionally stained with rat anti-mouse CD45-AF647 ( clone 30-F11; Biolegend , London , UK ) . Rat anti-iNOS-PE , rat anti-CD11b-PerCPCy5 . 5 . and rat anti-mouse CD45-AF647 were used at 1:200 dilutions in Perm/Wash solution . After 20 min at 4°C cells were finally washed and resuspended in PBS/1% paraformaldehyde . CD4+ and CD8+ T cells were stained extracellularly in blood with anti-mouse CD4-PE ( clone GK1 . 5; BD Biosciences , Heidelberg , Germany; 1:200 ) and anti-mouse CD8-PerCP-Cy5 . 5 ( clone 53–6 . 7; BD Biosciences , Heidelberg , Germany; 1:200 ) and in brain with anti-mouse CD4 PerCPCy5 . 5 ( clone RM4-5; eBioscience , Frankfurt , Germany; 1:200 ) and anti-mouse CD8-Alexa488 ( clone 53–6 . 7; Biolegend , London , UK; 1:200 ) . In the brain , cells were additionally stained with rat anti-mouse CD45-AF647 ( clone 30-F11; Biolegend , London , UK ) to identify CD45high infiltrating cells . KLRG1 and CD11a were detected by extracellular staining employing anti-mouse KLRG1-PE ( clone 2F1/KLRG1; Biolegend , London , UK; 1:800 ) and anti-mouse CD11a-eFluor450 ( clone M17/4; eBioscience , Frankfurt , Germany; 1:200 ) . Intracellular IFNγ and Granzyme B were detected with anti-mouse IFNγ-PE/Dazzle ( clone XMG1 . 2; Biolegend , London , UK; 1:333 ) and anti-mouse Granzyme B-PacificBlue ( clone GB11 , Biolegend , London , UK; 1:200 ) in spleen cells restimulated with 10 ng/ml PMA and 500 ng/ml Ionomycin in 200 μl in 96well plates in the presence of 1 μl GolgiStop ( BD Biosciences , Heidelberg , Germany ) for 4h and permeabilized with cytofix/cytoperm ( BD Biosciences , Heidelberg , Germany ) according to the manufacturer´s protocol . For the analysis of surface expression of MHC molecules and CD80 bmMΦ were stained with anti-mouse CD11b-BV421 ( clone M1/70 , Biolegend , London , UK; 1:50 ) . R . typhi was detected by intracellular staining with anti-R . typhi BNI52 ( 1 μg/ml ) , anti-mouse IgG3-FITC ( #1100–02 , SouthernBiotech , Birmingham , USA; 1:200 ) and anti-FITC-Alexa488 ( Thermo Fisher Scientific , Braunschweig , Germany; 1:1000 ) . MHC molecules and CD80 were stained extracellularly with anti-mouse MHCI ( H2-Kb ) -PE ( AF6-88 . 5 . 5 . 3 , eBioscience , Frankfurt , Germany; 1:50 ) , anti-mouse MHCII ( I-A/I-E ) -APC ( M5/114 . 15 . 2 , eBioscience , Frankfurt , Germany; 1:50 ) and anti-mouse CD80-PE/Dazzle ( 16-10A1 , Biolegend , London , UK; 1:50 ) . Analysis was performed employing a Accuri C6 ( BD Biosciences , Heidelberg , Germany ) or LSRII flow cytometer ( BD Biosciences , Heidelberg , Germany ) and FlowJo software ( FlowJo LLC , Ashland , USA ) . bmMΦ were generated from tibia and femur of the hind legs . 2×106 bone marrow cells were cultured in petri dishes ( Sarstedt , Nuembrecht , Germany ) in IMDM medium ( PAA Laboratories , Cölbe , Germany ) supplemented with 2 mM L-glutamine , 5% horse serum and 10% M-CSF-containing cell culture supernatant from L929 fibroblasts . Medium was exchanged every 3 days . Cells were harvested for experimentation after 10–12 days of culture . Virtually 100% of the cells were CD11b+F4/80+ MΦ . For the analysis of cytokine and NO production and the expression of MHCI , MHCII and CD80 on the cell surface 2×105 bmMΦ were infected with 10 , 25 or 50 copies of R . typhi per cell in 24well plates . Control cells were left untreated or stimulated with LPS ( 500 ng/ml ) . Supernatants and cells were harvested at 24h and 48h post infection . Bacterial content and the expression of MHCI , MHCII and CD80 on the cell surface was analyzed by flow cytometry at 24h and 48h . Cytokines and NO were detected in the supernatant of cultures that were infected with 50 copies R . typhi/cell at 48h by LegendPLEX assay and Griess reaction . For the analysis of bacterial growth 2×105 bmMΦ were infected with 5 copies R . typhi per cell . Free bacteria were washed out 3h afterwards and cells were further incubated for 96h . At indicated points in time qPCR was performed to quantify R . typhi . Immune CD4+ and CD8+ T cells were isolated from C57BL/6 mice on day 21 post R . typhi infection employing the MagniSort Mouse CD4 and CD8 T cell enrichment kits from eBioscience ( Frankfurt , Germany ) using an EasySep Magnet from Stemcell Technologies ( Köln , Germany ) . Procedures were performed according to the manufacturer´s instructions . Purity was analyzed by flow cytometry and found to be >98% for each cell population in all preparations . CD4+ T cells were generally absent in CD8+ T cell preparations and vice versa . 1×106 CD4+ or CD8+ T cells were injected i . v . in 100 μl PBS into C57BL/6 RAG1-/- mice 63 days after R . typhi infection , which is the point in time when bacteria become detectable by qPCR in the brain and approximately 20–40 days prior to the usual onset of disease . R . typhi-infected control animals received PBS instead of T cells . Additional C57BL/6 RAG1-/- control mice were not infected but received either CD4+ or CD8+ immune T cells . 1×106 bmMΦ were infected with 5 copies R . typhi per cell in 24well plates . Free bacteria were washed out after 3h and cells were further incubated for 24h at 37°C to allow bacterial entry and replication . Immune CD4+ T cells were isolated as described above from R . typhi-infected C57BL/6 wild-type mice 7 days post infection . Control CD4+ T cells were purified from the spleen of mice that received PBS instead . 1 . 5×106 purified CD4+ T cells were added per well and cells were further incubated for 72h . IFNγ was neutralized by simultaneous addition of anti-IFNγ ( 1 μg/ml; BioXCell , West Lebanon , USA ) . In another experimental setup 1×106 bmMΦ were infected with 5 copies R . typhi per cell . After 3h hours the medium was exchanged by cell culture medium with or without recombinant IFNγ ( 10 U/ml; Merck Millipore , Eschborn , Germany ) or a combination of recombinant IFNγ that was pre-incubated with anti-IFNγ ( 1 μg/ml ) for 15 min at RT for neutralization . Supernatants and cells were harvested 96h after infection . Cytokines and NO were quantified in the supernatants by LegendPLEX and Griess assay . Bacterial content was determined in the cell pellet by PrsA qPCR . For immunohistochemistry ( IHC ) tissues from infected mice were fixed in 4% formalin in PBS and embedded in paraffin . Deparaffinization of the sections was performed using standard methods . Sections were first heated at 63°C for 30 min in a heating cabinet followed by treatment with Xylol for 30 min and EtOH ( 3x 100% EtOH , 3x 96% EtOH , 80% EtOH , 70% EtOH ) . Each step was performed for 3–5 min . Slides were finally washed in H2O . Deparaffinized sections were boiled for 30 min in 10 mM citrate buffer ( 10 mM sodium citrate , 0 . 05% Tween20 , pH6 . 0 ) for antigen retrieval . Staining was performed using a Ventana Benchmark XT apparatus ( Ventana , Tuscon , USA ) . Antibodies were diluted in 5% goat serum ( Dianova , Hamburg , Germany ) in Tris-buffered saline pH7 . 6 ( TBS ) and 0 . 1% Triton X100 in antibody diluent solution ( Zytomed , Berlin , Germany ) . Rabbit anti-mouse CD3 ( 1:100; clone SP7; Abcam , Cambridge , USA ) , rabbit anti-mouse IBA1 ( 1:500; #019–19741; WAKO , Neuss , Germany ) and rabbit anti-mouse iNOS ( 1:75; ABIN373696 , Abcam , Cambridge , USA ) were used . Anti-rabbit or anti-rat Histofine Simple Stain Mouse MAX peroxidase-coupled antibodies ( Nichirei Biosciences , Tokyo , Japan ) were used as secondary antibodies . Detection was performed with ultraview universal DAB detection kit ( Ventana , Tuscon , USA ) . Sections were covered with Tissue-Tek embedding medium ( Sakura Finetek , Staufen , Germany ) . Images were taken with a BZ9000 Keyence microscope ( Keyence , Neu-Isenburg , Germany ) . Statistical analysis was performed with GraphPad Prism 5 software ( GraphPad Software Inc . , La Jolla , USA ) . Student´s T test , Mann-Whitney U test , or One-way ANOVA test followed by Kruskal-Wallis and Dunn´s post test were performed as indicated in the figure legends .
T cells , especially cytotoxic CD8+ T cells , are important in the elimination of intracellular pathogens . Therefore , we first analyzed the CD8+ T cell response in R . typhi-infected C57BL/6 mice during the course of infection . For these analyses spleen cells were re-stimulated in vitro with PMA/Ionomycin followed by flow cytometric analysis of IFNγ expression and the expression of Granzyme B as a cytotoxic effector molecule . In addition , CD8+ T cells from the spleen of the mice were analyzed for the expression of KLRG1 as a marker for terminally differentiated non-replicative T cells [67 , 68] and CD11a as a marker for antigen-experienced cells [69] . Analyses were performed on days 0 ( naïve ) , 3 , 7 , 15 and 35 post infection . CD8+CD11a+ T cells began to rise on day 3 , peaked on day 7 and declined until day 15 ( Fig 1A , left ) . Differentiated CD8+KLRG1+ T cells peaked on day 7 and were not yet detectable on day 3 post infection ( Fig 1 , middle ) . CD8+KLRG1+ T cells also declined until day 15 . However , enhanced numbers as well as enhanced frequencies of KLRG1+ CD8+ T cells were still observed on day 35 post infection although these differences were not significant . Cell numbers as well as the percentage of KLRG1+ CD8+ T cells were approximately doubled at this late point in time ( day 35: 2 . 43x105±3 . 70x104 , 2 . 27±0 . 27%; naïve: 1 . 16x105±2 . 77x104 , 1 . 11±0 . 18%; Fig 1A , middle and right ) . In line with these findings a significantly enhanced proportion of CD8+ T cells expressed IFNγ as well as Granzyme B on day 7 post infection . These populations declined until day 15 . The frequency of IFNγ producers , however , did not reach basal levels again ( Fig 1B ) . Similar was also true for CD4+ T cells that were analyzed in parallel . A peak of enhanced frequencies of IFNγ-producing CD4+ T cells was detectable on day 7 post infection . The population of IFNγ-expressing CD4+ T cells declined until day 15 but did not return to basal level until day 35 ( Fig 1C , left ) . These data show that C57BL/6 mice mount an efficient cytotoxic CD8+ T cell response and a CD4+ TH1 response that is characterized by the expression of IFNγ . We next asked whether CD8+ and/or CD4+ T cells exert protective functions in vivo . In a first approach we used C57BL/6 MHCII-/- and C57BL/6 MHCI-/- mice that either lack CD4+ T cells or CD8+ T cells . All of these mice were protected against R . typhi-induced disease and survived the infection more than 150 days ( S1A Fig ) , demonstrating that CD8+ as well as CD4+ T cells are sufficient for protection . Second , we performed adoptive T cell transfer into R . typhi-infected C57BL/6 RAG1-/- mice to show whether CD8+ and CD4+ T cells would still be protective in an established R . typhi-infection . Because these animals are capable to control the bacteria for a long period of time , antigen availability may be low and not sufficient to elicit efficient T cell responses in time . Therefore , we decided to transfer isolated immune CD8+ and CD4+ T cells from R . typhi-infected C57BL/6 wild-type mice instead of T cells from naïve animals as these get activated much faster upon antigen recognition . CD8+ and CD4+ T immune cells were obtained from C57BL/6 wild-type mice on day 21 post R . typhi infection and first transferred into C57BL/6 RAG1-/- mice either on day 45 or on day 55 post infection . None of these animals developed disease whether receiving CD8+ or CD4+ T cells and survived the infection . S1B Fig shows the results for the transfer experiment performed on day 55 . We next adoptively transferred isolated immune CD4+ and CD8+ T cells into C57BL/6 RAG1-/- mice later in infection on day 63 , which is approximately 20–40 days prior to the usual onset of neurological symptoms . At this point in time the bacteria become detectable in the brain by qPCR . Fig 2A shows a schematic overview of the procedure . Control animals received PBS instead of T cells . Additional control groups of mice received either CD4+ or CD8+ T cells but were non-infected . First , we analyzed if the transferred T cells were detectable in the blood of recipient mice and if the cells would enter the brain . These analyses were performed on day 7 post T cell transfer . In the T cell recipient C57BL/6 RAG1-/- mice CD4+ as well as CD8+ T cells were clearly detectable in the blood of all animals that received the respective T cell population ( Fig 2B ) . In the brain infiltrating cells can generally be distinguished from resident immune cells ( mainly microglia ) by the expression of high levels of CD45 . CD45high cells including T cells are virtually absent in the brain of naïve C57BL/6 wild-type mice ( Fig 2C ) and were not detectable in the brain of non-infected C57BL/6 RAG1-/- mice that received either immune CD4+ or CD8+ T cells . In contrast , all R . typhi-infected C57BL/6 RAG1-/- mice showed high numbers of CD45high cellular infiltrates in the brain . Among these , CD4+ T cells were detectable at high frequencies in all CD4+ recipient mice ( 31 . 46±6 . 50% ) while CD8+ T cells were present in the brain of only 2 out of 5 recipient mice at this point in time ( 36 . 20±2 . 20% ) ( Fig 2C ) although all animals of this group showed low frequencies of CD8+ T cells in the blood ( Fig 2B ) . These data demonstrate that immune T cells enter the CNS of R . typhi-infected C57BL/6 RAG1-/- mice . 90% of the control C57BL/6 RAG1-/- mice that were infected with R . typhi but received PBS instead of T cells succumbed to the infection between day 70 and 120 , showing neurological symptoms such as tremor , ataxia and paralysis followed by body weight loss . These were evaluated by a neurological score ( Fig 3A ) . Despite the late transfer on day 63 , C57BL/6 RAG1-/- mice that obtained CD8+ immune T cells did not show any signs of disease at any point in time . Moreover , all of these animals survived the infection . In contrast , approximately 40% of the C57BL/6 RAG1-/- mice that obtained CD4+ immune T cells developed a neurological score and the same symptoms as R . typhi-infected C57BL/6 RAG1-/- control mice including tremor and/or ataxia . Furthermore , disease progressed in these animals with similar kinetics as in control mice . The affected CD4+ T cell recipients lost weight and died within the same time frame as control mice ( Fig 3A ) . The remaining CD4+ T cell recipients did not show symptoms of disease at any point in time and survived ( Fig 3A ) . These data demonstrate that CD4+ T cells are less protective against R . typhi-induced disease than CD8+ T cells when applied late in advanced infection . Next , we performed R . typhi-specific qPCR from DNA of different organs from R . typhi-infected CD4+ and CD8+ T cell recipients and control mice as well as from non-infected CD4+ and CD8+ T cell recipients to quantify the bacterial load . Analysis was performed for all groups on day 7 post T cell transfer or treatment with PBS , respectively . Furthermore , bacterial load was determined at the time of death in control animals and CD4+ T cell recipients that did not survive the infection . Organs from CD8+ and surviving CD4+ T cell recipients were taken on day 210 when the experiment was terminated . As described previously , the bacteria were predominantly found in the brain of R . typhi-infected C57BL/6 RAG1-/- mice but were also present at lower amounts in the spinal cord , spleen and lung ( Fig 3B and 3C upper panel ) and virtually absent in the liver . The average copy numbers actually measured in each organ of each group are shown in Table 1 . In mice that had received CD8+ immune T cells the bacteria were already significantly reduced and almost eliminated in all analyzed tissues as early as at day 7 post T cell transfer ( Fig 3B and Table 1 ) . Furthermore , on day 7 post transfer mice that had received CD4+ T cells also showed reduced bacterial loads in the brain , spinal cord and lung but not in the spleen where the bacterial content was generally low ( Fig 3B ) . On day 210 when the experiment was terminated R . typhi was not detectable anymore in six out of ten CD8+ T cell recipient mice . However , the remaining four animals showed low amounts of bacteria predominantly in the spinal cord . Three of these mice also had few bacteria in the brain while R . typhi was not detectable in the spleen and lung ( Fig 3C lower panel and Table 1 ) . Surprisingly , the bacteria were not present anymore in the five CD4+ recipients that succumbed to the infection . However , six mice of the seven surviving animals of this group had bacteria in the spinal cord at the end of the experiment on day 210 . Three of these animals also showed low copy numbers in the brain and one mouse in the lung ( Fig 3C lower panel and Table 1 ) . The bacteria were generally not detectable in the organs from non-infected animals that received either immune CD4+ or CD8+ T cells ( S2 Fig ) , excluding that a possible co-transfer of low amounts of contaminating R . typhi might have contributed to infection . These data clearly demonstrate that immune CD4+ T cells are capable to eradicate R . typhi although bacterial elimination by CD8+ T cells is much more efficient and faster . They further show that both CD8+ and CD4+ T cells are capable to control persisting R . typhi and prevent recurrence of disease . We have previously shown that R . typhi-induced CNS inflammation is characterized by the expansion of microglia as well as by the infiltration of MΦ from the periphery and that the latter harbor the bacteria [48] . Therefore , we further performed flow cytometric analyses of the brain to quantify microglia and MΦ . As observed previously a significant increase in the numbers of microglia ( CD45lowCD11b+ ) as well as of infiltrating MΦ ( CD45highCD11b+ ) was detectable in the brain of R . typhi-infected control C57BL/6 RAG1-/- mice ( microglia: 3913±386 . 7; MΦ: 3160±433 . 3 among 1×106 events ) compared to naïve mice ( microglia: 625 . 6±162 . 8 ) where infiltrating MΦ were virtually absent ( MΦ: 68 . 19±10 . 42 ) as expected . Counts of microglia were not signicantly altered in CD4+ T cell recipients ( 4285±774 . 4 ) while an enhanced infiltration of MΦ , although not significant , was observed ( 5082±995 . 3 ) compared to infected control mice . In contrast , mice that had received CD8+ T cells showed significantly reduced numbers of microglia ( 1663±397 . 4 ) and numbers of infiltrating MΦ were unchanged compared to infected control mice ( 2272±767 . 4 ) ( Fig 4A ) . We further analyzed the activation status of microglia and infiltrating MΦ . For this purpose we performed intracellular staining of inducible nitric oxide synthase ( iNOS ) . This enzyme is usually expressed by activated MΦ and important for the killing of intracellular bacteria by catalyzing the generation of nitric oxide ( NO ) [70] . We previously described that a significantly enhanced proportion of infiltrating MΦ but not microglia express iNOS in R . typhi-infected C57BL/6 RAG1-/- mice at the time of death [48] . Here , analyses were performed much earlier 7 days post T cell transfer ( day 70 post infection ) which is before the usual onset of symptoms . Although enhanced frequencies of iNOS-expressing infiltrating MΦ were observed in the brain of C57BL/6 RAG1-/- control mice ( 8 . 375±1 . 118% ) compared to naïve mice ( 0 . 6533±0 . 6533% ) , these differences were not significant at this point in time ( Fig 4B ) . The frequency of iNOS-expressing infiltrating MΦ was not significantly altered in mice that had received CD8+ T cells ( 11 . 1±3 . 156% ) . In contrast , the transfer of CD4+ T cells led to a significantly increased proportion of iNOS-expressing MΦ ( 28 . 10±6 . 713% ) . Furthermore , the presence of CD4+ T cells led to a significantly increased frequency of iNOS-expressing microglia ( 3 . 132±0 . 717% ) . These cells normally do not express iNOS in R . typhi-infection of C57BL/6 RAG1-/- mice [48] . In line with these previous findings , iNOS-expressing microglia were not detectable in R . typhi-infected control mice ( 0 . 8054±0 . 2214% ) compared to 0 . 5350±0 . 2074% in naïve mice . Furthermore , enhanced frequencies of iNOS-expressing microglia were not present in mice that had received CD8+ T cells ( 1 . 254±0 . 2734% ) ( Fig 4B ) . Enhanced numbers of iNOS-expressing cells were also detectable in the CNS of CD4+ T cell recipients in histological stainings . Fig 5 shows representative overview stainings of iNOS in the brain from a R . typhi-infected control mouse , a CD4+ and a CD8+ T cell recipient 7 days post transfer . iNOS-expressing cells were hardly detectable in control mice and animals that received CD8+ T cells . In CD4+ T cell recipients locally accumulating iNOS-expressing cells were found at the ventricle borders as well as in the parenchyma and the pia mater of the cerebellum ( Fig 5 ) . Furthermore , infiltrating CD4+ and CD8+ T cells were observed in R . typhi-infected recipients mainly at the ventricle borders but also present in the parenchyma ( Fig 6C and 6D ) while T cells were not detectable in the brain of non-infected mice that received CD8+ or CD4+ T cells ( Fig 6A ) . Infiltrating T cells were accompanied by accumulating IBA1+ cells , most likely infiltrating MΦ . Only the brains of mice that had received CD4+ T cells showed high numbers of iNOS-expressing IBA1+ cells that colocalized with infiltrating T cells ( Fig 6C ) . Similar observations were made in histological stainings of the spinal cord 7 days post transfer . In R . typhi-infected control mice that had received PBS instead of T cells , accumulating IBA1+ cells were observed in the gray and white matter ( Fig 7B ) . These cells most likely represent microglia . In addition , accumulating IBA1+ cells , most probably infiltrating MΦ , were detectable in peripheral areas , presumably the pia mater and subarachnoid space where the bacteria reside [48] . CD4+ and CD8+ T cell infiltrates were also predominantly detectable in these areas while only few T cells were observed in the gray and white matter in recipient mice ( Fig 7C and 7D ) . In R . typhi-infected control mice iNOS-expressing cells were still rare at this point in time of infection ( Fig 7B ) and also hardly detectable in mice that received CD8+ T cells ( Fig 7D ) . In contrast , CD4+ T cell infiltrates were associated with iNOS-expressing infiltrating IBA1+ MΦ . Exclusively in the spinal cord of animals that had received CD4+ T cells few iNOS-expressing cells were also detectable in the gray and white matter ( Fig 7C ) . Neither T cell infiltration nor accumulation of IBA1+ cells and iNOS expression occurred in the spinal cord of non-infected mice that had received CD8+ or CD4+ T cells ( Fig 7A ) . Collectively , these results demonstrate that CD4+ T cells significantly induce iNOS-expression not only in infiltrating MΦ but also in microglia , indicating enhanced bactericidal function of these cells and enhanced inflammatory response in the CNS of these animals . MΦ represent target cells for R . typhi that infiltrate the CNS of infected C57BL/6 RAG1-/- mice [48] . Because these cells were found to express iNOS in vivo , we further asked if R . typhi would infect MΦ in vitro and how these cells would react to the bacteria . To this end we incubated bmMΦ with titrated amounts of R . typhi or stimulated the cells with LPS while control cells were left untreated . First , we assessed bacterial uptake . After 48h of incubation approximately 15% of the bmMΦ that were infected with 50 bacterial particles per cell were positive for R . typhi as detected by flow cytometry ( Fig 8A ) . In fact , at least one bacterium was detectable in the cytosol of every cell 48h after infection in immunofluorescent stainings ( Fig 8A , insertion ) demonstrating that all cells had been in contact with R . typhi . We further analyzed the expression of MHCI , MHCII and costimulatory molecules on the cell surface 24h and 48h after inoculation . LPS-stimulated and untreated MΦ were used as a control . LPS stimulation led to a significant up-regulation of CD80 and MHCI with maximum expression at 48h while MHCII expression was temporarily enhanced at 24h ( Fig 8A ) . R . typhi-infected MΦ up-regulated the expression of CD80 , MHCI and MHCII in a dose-dependent manner with similar kinetics ( Fig 8A ) demonstrating MΦ activation . We further assessed cytokine and NO production in cultures that were infected with the highest dose of R . typhi ( 50 bacterial copies per cell ) . As expected , LPS significantly induced the release of several cytokines as well as the production of NO ( Fig 8B ) . Surprisingly , R . typhi-infected bmMΦ hardly released any cytokines . Only very low amounts of IL-6 ( 28 . 4±15 . 4 pg/ml ) , TNFα ( 11 . 1±4 . 6 pg/ml ) and IL-10 ( 10 . 9±3 . 5 pg/ml ) were detectable . These were negligible compared to LPS-induced levels of these cytokines ( IL-6: 27039±1896 pg/ml , TNFα: 1865±101 , 7 pg/ml , IL-10: 184 . 6±12 . 5 pg/ml ) . Furthermore , the release of bactericidal NO was not induced in R . typhi-infected bmMΦ cultures ( Fig 8B ) . These data show that MΦ do not react to R . typhi in a classical manner . The observation that R . typhi does not induce the production of bactericidal NO further led to the question whether bmMΦ are capable to kill R . typhi in vitro . Therefore , we analyzed bacterial growth in bmMΦ cultures . Bacterial content was quantified by qPCR at indicated points in time . Significantly enhanced copy numbers were detected already at 48h and bacteria further increased until 96h post inoculation ( Fig 8C ) , demonstrating that MΦ are incapable to eliminate the bacteria in vitro and that R . typhi grows within these cells . Having shown that CD4+ T cells produce IFNγ in R . typhi infection and enhance MΦ activation in vivo , we finally asked whether immune CD4+ T cells or IFNγ can activate bmMΦ for bacterial elimination in vitro . To this end , bmMΦ were infected with R . typhi in vitro as described in the previous section . 24h after inoculation either immune CD4+ T cells or CD4+ T cells from non-infected C57BL/6 control animals were added . In addition , IFNγ was neutralized by anti-IFNγ . The release of cytokines and NO and bacterial growth was assessed 96h post bmMΦ infection . Control CD4+ T cells did not react to infected bmMΦ with the release of detectable amounts of cytokines and did not induce the release of NO whereas cultures containing immune CD4+ T cells produced very high amounts of IFNγ ( 23456±4758 pg/ml ) and IL-2 ( 1219 . 7±130 . 8 pg/ml ) . In addition , low levels of TNFα ( 85 . 2±26 . 7 pg/ml ) , IL-6 ( 11 . 8±6 . 8 pg/ml ) , IL-10 ( 288 . 6±28 . 4 pg/ml ) and bactericidal NO ( 3 . 9±2 μM ) were detectable in cultures containing immune CD4+ T cells ( Fig 9A ) . Despite the high levels of IFNγ the cytokine could be neutralized to a certain extent ( 5203±1606 pg/ml ) . Neutralization of IFNγ , however , had no significant effect on the production of the other cytokines or NO ( Fig 9A ) . Furthermore , immune CD4+ T cells nearly completely inhibited bacterial growth ( 43±26 R . typhi PrsA copies ) compared to cultures with control CD4+ T cells ( 17320±13458 R . typhi PrsA copies ) ( Fig 9B ) . This effect was in part , although not significantly , inhibited by the addition of neutralizing anti-IFNγ antibody ( 451±305 copies ) ( Fig 9A ) . To further elucidate the role of IFNγ , we incubated infected and non-infected bmMΦ with recombinant IFNγ . The cytokine alone did not induce the release of detectable amounts of cytokines or NO in bmMΦ , whether infected or not . However , IFNγ partially inhibited bacterial growth ( 2685±317 copies ) compared to untreated infected bmMΦ cultures ( 7457±1261 copies ) . This effect was abolished by pre-incubation of IFNγ with neutralizing anti-IFNγ antibody ( 8200±1190 copies ) , demonstrating that it relies on the biological activity of the cytokine ( Fig 9B ) . Collectively , these results demonstrate that immune CD4+ T cells enhance bacterial elimination by activating bactericidal functions of MΦ which is at least in part mediated by IFNγ .
We recently described that R . typhi shows a neurotropism in C57BL/6 RAG1-/- mice that lack adaptive immunity . In these mice , R . typhi re-appears months after infection predominantly in the CNS and causes severe CNS inflammation and lethal paralysis . Employing this model and adoptive transfer we describe here that CD8+ and CD4+ T cells enter the CNS and that both T cell populations are protective against this infection and R . typhi-induced disease . C57BL/6 wild-type mice mount a classical CD4+ TH1 T cell response that is characterized by IFNγ expression in response to R . typhi infection . In addition , functional cytotoxic CD8+ T cells expressing IFNγ and Granzyme B , an effector molecule that is crucial for target cell killing by rapid induction of apoptosis [71] , were generated . Both IFNγ expression in CD4+ and CD8+ T cells as well as Granzyme B expression in CD8+ T cells peaked on day seven post infection which was consistent with the expression of CD11a and KLRG1 on CD8+ T cells , demonstrating that these cells were antigen-experienced effector cells . In line with these findings KLRG1 was expressed on CD8+ T cells with similar kinetics in C3H/HeN mice upon R . typhi infection [72] . Furthermore , enhanced cytotoxic CD8+ T cell responses were observed in R . conorii- and R . australis-infected C3H/HeN mice peaking at day 10 post infection [55] . T cell response to R . typhi in C57BL/6 mice declined until day 15 but did not reach basal levels again until day 35 . This is consistent with the observation that R . typhi persists in these mice [48] and suggests that a certain level of activated T cells is needed for durable control of the bacteria . Here we demonstrate that CD4+ T cells are sufficient for protection against R . typhi infection . Neither CD4+ T cell deficient C57BL/6 MHCII-/- nor CD8+ T cell-deficient C57BL/6 MHCI-/- mice developed disease and survived the infection . We further performed adoptive transfer of immune T cells isolated from C57BL/6 wild-type mice into R . typhi-infected C57BL/6 RAG1-/- . Transfers were performed at a point in time when the bacteria already start to dramatically increase in the brain , the organ with the highest bacterial load in these animals [48] . T cells for transfer were isolated from R . typhi-infected C57BL/6 wild-type mice on day 21 post infection . These mice are highly resistant to the infection with R . typhi and other rickettsiae . Pure T cells from these mice were used for transfer . Because T cells are not target cells for the bacteria it is highly unlikely that significant amounts of bacteria were co-transferred and might have established a superinfection . In line with that , the bacteria were not detectable in the organs of non-infected animals that received either immune CD4+ or CD8+ T cells . Moreover , both immune CD8+ and CD4+ T cells entered the brain in R . typhi-infected C57BL/ RAG1-/- mice but not in non-infected animals , efficiently eliminated the bacteria and were protective . The survival rate of CD4+ T cell recipients , however , was lower than that of CD8+ T cell recipients , indicating that CD4+ T cells are less efficient in protecting animals with already advanced infection . Generally , CD8+ T cells were absent in CD4+ recipients and vice versa ( Fig 2B and 2C ) so that a contribution of the other party to the effects discussed in the following can be excluded . C57BL/6 RAG1-/- CD8+ T cell recipients eliminated R . typhi very quickly from the brain as well as from other organs in established R . typhi-infection in C57BL/6 RAG1-/- mice . Already 7 days after the transfer of CD8+ T cells , R . typhi was not detectable anymore in CD8+ T cell recipients . Moreover , in some mice CD8+ T cells were not detectable anymore in the brain already on day 7 after transfer although the cells were still present in the blood . This may indicate bacterial elimination and an already declining immune response in these animals while bacterial load might have been higher in those mice that still showed CD8+ T cell infiltrates and , thus , ongoing immune response . In the end , adoptive transfer of CD8+ T cells completely prevented disease in all animals , demonstrating that CD8+ T cells alone are sufficient for protection . So far , a protective function of CD8+ T cells against R . typhi has been only demonstrated in the C3H/HeN infection model where the depletion of CD8+ T cells resulted in enhanced bacterial load and pathology [57] while the role of cytotoxic CD8+ T cells has been studied in more detail in mouse models of the infection with SFG Rickettsiae . C3H/HeN mice depleted of CD8+ T cells and challenged with a normally sublethal dose of R . conorii died or remained persistently infected . Moreover , adoptive transfer of immune CD8+ T cells protected C3H/HeN mice against a lethal challenge with R . conorii [56] . In addition , C57BL/6 MHCI-/- mice that lack CD8+ T cells were highly susceptible to a lethal outcome of R . australis infection [55] . Defense against R . australis is largely mediated by the cytotoxic activity of CD8+ T cells rather than IFNγ as C57BL/6 Perforin-/- mice showed enhanced susceptibility to this infection compared to wild-type mice . Furthermore , adoptive transfer of CD8+ T cells from C57BL/6 IFNγ-/- mice into R . australis-infected C57BL/6 MHCI-/- still reduced bacterial load [55] . These observations indicate a critical function of cytotoxic CD8+ T cells in defense against rickettsial infections and our results further show that CD8+ T cells alone can provide long-term control of persisting R . typhi and prevent recurrence of disease . However , C57BL/6 MHCI-/- mice that lack CD8+ T cells were not susceptible to R . typhi infection and did not develop symptomatic disease even until 150 days post infection . Moreover , adoptive transfer of CD4+ T cells from immune mice into C57BL/6 RAG1-/- with advanced R . typhi infection still protected at least 60% of the mice from disease and death . In fact , the bacteria were efficiently eliminated from the CNS in those CD4+ T cell recipients that succumbed to the infection . Collectively , these results demonstrate for the first time that CD4+ T cells alone are sufficient to protect against R . typhi-induced disease as long as they are present in time . Furthermore , CD4+ T cells are as efficient as CD8+ T cells in providing long-term control of persisting R . typhi and are also capable to prevent recurrence of disease . So far , a protective function of CD4+ T cells has only been demonstrated in the infection of mice with SFG Rickettsiae . Here , adoptive transfer of immune CD4+ T lymphocytes protected C3H/HeN mice against challenge with a lethal dose of R . conorii [56] . C3H/HeN mice depleted of CD4+ T cells underwent a comparable course of disease in a sublethal infection with R . conorii as control mice . The mice cleared the infection and recovered [56] . This situation is similar to the infection of CD4+ T cell-deficient C57BL/6 MHCII-/- mice with R . typhi where CD8+ T cells are present and can compensate for the absence of CD4+ T cells . Surprisingly , although both adoptively transferred immune CD8+ and CD4+ T cells managed to almost completely eliminate the bacteria from the organs of R . typhi-infected C57BL/6 RAG1-/- mice in the first days to weeks after transfer , the bacteria were again detectable by qPCR predominantly in the brain and spinal cord of some animals when the experiments were terminated . We interprete these findings as follows: Transferred T cells strongly react to R . typhi immediately after transfer into infected C57BL/6 RAG1-/- mice which is reflected by the strong induction of iNOS in CD4+ T cell recipients . This immediate response is obviously strong enough to reduce the bacterial amounts below qPCR detection limit in the initial phase after transfer . Later in the chronic phase of infection ( until day 210 ) the T cell response calms down . In this situation a certain threshold of either activated CD4+ or CD8+ T cells as observed in R . typhi-infected C57BL/6 mice ( Fig 1 ) is capable to prevent fatal bacterial outgrowth and to keep bacterial amounts at a low level that is detectable in some animals by qPCR . The co-action of both cell populations , however , is obviously more efficient in bacterial suppression as numbers of persisting R . typhi in infected C57BL/6 wild-type mice are much lower and not detectable by qPCR [48] . The induction of IFNγ-producing CD4+ TH1 cells is usually associated with the release of IL-12 which is the main IFNγ-inducing cytokine for T cells and NK cells [73] and it has been shown that efficient immune response against R . typhi in C3H/HeN mice was associated with enhanced serum levels of IL-12 on day 5 post infection [57] . IL-12 , however , was neither detectable in the sera of C57BL/6 RAG1-/- nor C57BL/6 wild-type mice at any point in time upon R . typhi infection [48] . IL-12 is predominantly derived from antigen-presenting cells ( APC ) such as MΦ and dendritic cells ( DCs ) and is usually induced by the recognition of bacteria via pattern recognition receptors such as toll-like receptors ( TLR ) [74] . The analyses of MΦ responses , however , revealed that these cells generally do not react to R . typhi in a classical manner . Although MΦ up-regulated MHCI and CD80 and showed temporary upregulation of MHCII after infection with R . typhi in vitro , neither proinflammatory cytokines including IL-12 nor bactericidal NO were released . In contrast to the infection of MΦ in vitro , however , infiltrating MΦ in the CNS of R . typhi-infected C57BL/6 RAG1-/- mice expressed iNOS and , thus , released NO . Whether the activation of APC such as MΦ in vivo is directly induced by R . typhi or supported by other mechanisms such as locally expressed mediators in the affected tissue and how efficient induction of CD4+ TH1 cells is achieved is unknown . IFNγ and TNFα produced by CD4+ TH1 cells [75] are important mediators of protection against intracellular pathogens . IFNγ induces the expression of iNOS and subsequent release of NO in MΦ [70 , 76] and endothelial cells [77] both of which are target cells of Rickettsiae [4 , 78 , 79] . Also TNFα can induce iNOS expression in MΦ and synergizes with IFNγ in this effect [80] . In this way , these cytokines support the bactericidal activity of these cells and contribute to bacterial elimination . In concordance , we observed that the addition of IFNγ to R . typhi-infected bmMΦ in vitro leads to reduced bacterial growth and , thus , enhanced bacterial killing although the amount of IFNγ used for these experiments did not induce the release of detectable amounts of NO . Immune CD4+ T cells from R . typhi-infected C57BL/6 mice , however , produced very high amounts of IFNγ in the presence of infected bmMΦ and induced NO release in vitro . Furthermore , immune CD4+ T cells dramatically reduced bacterial growth in MΦ cultures in vitro compared to cultures with control CD4+ T cells . This effect was partially inhibited by the neutralization of IFNγ although neutralization of IFNγ was clearly not complete . This observation indicates a dominant role of IFNγ as a mediator of bacterial killing . In addition other factors such as TNFα that was also detectable in cultures of immune CD4+ T cells and infected bmMΦ at low amounts may contribute to bacterial elimination . In vivo infiltrating MΦ in the CNS of C57BL/6 RAG1-/- mice colocalized with infiltrating IFNγ-producing CD4+ T cells and expressed iNOS at enhanced frequencies . As these infiltrating MΦ were found to harbor the bacteria , this observation suggests that IFNγ-mediated MΦ activation contributes to bacterial elimination and defense against R . typhi in vivo . Both IFNγ and TNFα have been demonstrated to play an important role in protection against Rickettsiae . For example , C3H/HeN mice depleted of either IFNγ or TNFα showed enhanced disease upon infection with a dose of R . conorii that normally does not result in symptomatic disease [58] and similar observations were made for R . typhi infection of C3H/HeN mice [57] . Furthermore , a strongly enhanced susceptibility of C57BL/6 IFNγ-/- mice for the infection with R . australis was observed [55] . Interestingly , both neutralization of IFNγ as well as of TNFα resulted in impaired NO production and overwhelming disease in R . conorii-infected C3H/HeN mice [58] . Therefore , both cytokines obviously play a role in CD4+ T cell-mediated protection by activating MΦ . Apart from MΦ , immune CD4+ T cells also induced iNOS expression in microglia . These cells normally do not express the enzyme in R . typhi infection and also do not take up the bacteria [48] . Microglial accumulation and activation is associated with neurodegenerative diseases such as multiple sclerosis [81] and Alzheimer´s disease [82] . Activated microglia can directly mediate neuronal damage which is usually associated with iNOS expression [83 , 84] . Thus , additional activation of microglia , especially the induction of iNOS in these cells by CD4+ T cells , may not directly participate in the elimination of R . typhi but may have rather non-beneficial immunopathological effects . The observation that the bacteria were not detectable anymore in CD4+ T cell recipients that succumbed to the infection , suggests that neurological disease caused by neuronal damage in the CNS of these animals [48] is at least in part an immunopathological effect rather than a result of cellular destruction by R . typhi itself . Thinking about vaccination , the induction of an efficient CD8+ T cell response is considered the most promising means . Such responses , however , are difficult to induce as antigen must be introduced into the MHCI presentation pathway . Our results , however , show that CD4+ T cells can be as protective as CD8+ T cells in R . typhi-infected mice , provided that these cells are present in time . Moreover , CD4+ T cells prevented recurrence of disease . These findings suggest that CD4+ T cell-inducing vaccination might be as effective as the induction of CD8+ T cells . Vaccination with conserved CD4+ T cell antigens of TG Rickettsiae might even protect against both R . typhi and R . prowazekii because it has been shown that animals experimentally infected with R . typhi are immune to R . prowazekii infection and vice versa and that similar solid cross-immunity exists for humans [85] .
|
Endemic typhus caused by Rickettsia typhi usually is a relatively mild disease . However , CNS inflammation and neurological symptoms are complications that can occur in severe cases . This outcome of disease is regularly observed in T and B cell-deficient C57BL/6 RAG1-/- mice upon infection with R . typhi . We show here that CD4+ T cells are as protective as cytotoxic CD8+ T cells against R . typhi as long as they are present in time . This is evidenced by the fact that neither CD8+ nor CD4+ T cell-deficient C57BL/6 mice develop disease which is also true for R . typhi-infected C57BL/6 RAG1-/- mice that receive immune CD8+ or CD4+ at an early point in time . Moreover , adoptive transfer of immune CD4+ T cells still protects approximately 60% of C57BL/6 RAG1-/- mice when applied later in advanced infection when the bacteria start to rise . Although CD8+ T cells are faster and more efficient in bacterial elimination , R . typhi is not detectable in CD4+ T cell recipients anymore . We further show that immune CD4+ T cells activate bactericidal functions of microglia and macrophages in the CNS in vivo and inhibit bacterial growth in infected macrophages in vitro which is in part mediated by the release of IFNγ . Collectively , we demonstrate for the first time that CD4+ T cells alone are sufficient to protect against R . typhi infection . With regard to vaccination our findings suggest that the induction of R . typhi-specific CD4+ TH1 effector T cells may be as effective as the much more difficult targeting of cytotoxic CD8+ T cells .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"cytokines",
"nervous",
"system",
"immunology",
"neuroscience",
"microglial",
"cells",
"animal",
"models",
"developmental",
"biology",
"model",
"organisms",
"molecular",
"development",
"cytotoxic",
"t",
"cells",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"spinal",
"cord",
"staining",
"white",
"blood",
"cells",
"animal",
"cells",
"t",
"cells",
"mouse",
"models",
"glial",
"cells",
"immune",
"system",
"cell",
"staining",
"neuroanatomy",
"cell",
"biology",
"anatomy",
"central",
"nervous",
"system",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types"
] |
2016
|
CD4+ T Cells Are as Protective as CD8+ T Cells against Rickettsia typhi Infection by Activating Macrophage Bactericidal Activity
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.